Unix

(Cyanide/186064) Is filesystem or file system, or FileSystem, or ?

Cyanide@rss.iscabbs.com — Wed, 22 Feb 2012 09:48:00 -0600

Is filesystem or file system, or FileSystem, or ?
But I tend to use them all interchangably depending on the target audience..

(John Public/186063) gvfs-open! Thanks!! I sort of thought you were trolling at first...

John Public@rss.iscabbs.com — Tue, 21 Feb 2012 04:51:00 -0600

gvfs-open! Thanks!! I sort of thought you were trolling at first.

(The Rapture/186062) gvfs-open?

The Rapture@rss.iscabbs.com — Tue, 21 Feb 2012 04:49:00 -0600

gvfs-open?
xdg-open (maybe)?

(John Public/186061) Nope.

John Public@rss.iscabbs.com — Tue, 21 Feb 2012 04:39:00 -0600

Nope.

(The Rapture/186060) Try gnome-open

The Rapture@rss.iscabbs.com — Tue, 21 Feb 2012 04:34:00 -0600

Try gnome-open

(John Public/186059) gnome3: Is there a shell command that "opens" a file just like t...

John Public@rss.iscabbs.com — Tue, 21 Feb 2012 04:28:00 -0600

gnome3: Is there a shell command that "opens" a file just like the gui would
open it? For example,

    $ open-this  foo.pdf  ## Launch the preview thing.
    $ open-this  foo.7z   ## Launch the archiver gui thing.
    $ open-this  foo.doc  ## Launch the office document thing.

    etc.?

(Kena/186058) Well, yeah. 23 steps involving Unix interaction for non-Unix-sa...

Kena@rss.iscabbs.com — Mon, 20 Feb 2012 21:08:00 -0600

Well, yeah.  23 steps involving Unix interaction for non-Unix-savvy end-users
was definitely one big part.  The other was creating the hierarchy without
first creating a local tar, then copying it, then untarring it remotely -- did
it all in one with the whole tar-over-ssh trick.  Also saved a fair bit of time
by using zlib compression instead of bzip2 compression -- bzip2 will give you a
solid 10% smaller file, but it takes something like double the CPU time.  I
guess the bottom line was that someone who "knew enough to be dangerous" wrote
the procedure.

(Cyanide/186057) probably not using verbose cut the time down.. :) It is a lesson...

Cyanide@rss.iscabbs.com — Mon, 20 Feb 2012 11:24:00 -0600

probably not using verbose cut the time down.. :) It is a lesson I learned on a
shell account trying to compile a mud client, over a 300 baud modem line.
It saves oodles of time... Although there i just redirected stdout/stderr to
a file..but it is applicable to tar..

(Kena/186056) Every now and then, I feel a bit cool. My wife, who works for C...

Kena@rss.iscabbs.com — Mon, 20 Feb 2012 11:21:00 -0600

Every now and then, I feel a bit cool.  My wife, who works for Cisco, was
stepping through this 23-step procedure that, basically:
- Got a build version
- CD'd to the approrpiate directory
- Created a tar
- Opened another SSH window to a different machine
- Created a recipient directory
- SCP'd the tar
- Untarred it

I wrote a one-liner that got it down to *three* steps, and, because the build
tree was fairly large, cut computer time (a/k/a "waiting around watching the
screen" time) from about 45 minutes to about 15.  Ain't often that computers
actually *help* one's marital standing.

(Kena/186054) Orrrr... "The setuid permission set on a directory is ignored on...

Kena@rss.iscabbs.com — Sat, 18 Feb 2012 20:03:00 -0600

Orrrr... "The setuid permission set on a directory is ignored on UNIX and Linux
systems.[4] FreeBSD can be configured to interpret it analogously to setgid,
namely, to force all files and sub-directories to be owned by the top directory
owner."

(Cyanide/186053) it sounds like the permissions of the web server are set up diff...

Cyanide@rss.iscabbs.com — Sat, 18 Feb 2012 18:20:00 -0600

it sounds like the permissions of the web server are set up differently.
ie is the staging server running as www, instead of say root, where it
essentially sudo to the actual username?

(Fongaboo/186052) Running two almost identical Intel servers with FreeBSD and Apac...

Fongaboo@rss.iscabbs.com — Sat, 18 Feb 2012 18:09:00 -0600

Running two almost identical Intel servers with FreeBSD and Apache. Have
Wordpress installed on both machines, but the 2nd one is a staging server for
development.

When someone logs into Wordpress through the web dashboard and installs a
plug-in on the live server, it creates new files with ownership <user>:www. But
on the staging server it creates them as www:www. This is a problem because I
have non-root users trying to do editing of such files, so they can do so on
the live server but not the staging server.

Any setting in Wordpress/PHP/Apache that could change this?

(LOGAN/186051) You can create an additional MX record for mail-test.monetra.com...

LOGAN@rss.iscabbs.com — Fri, 17 Feb 2012 09:47:00 -0600

You can create an additional MX record for mail-test.monetra.com without
affecting the mail routing for monetra.com.  As far as the auth goes, I'd need
more information.  Are you authenticating as a local user, or a virtual user?
Also, the conversation with the mail server should start with HELO, and should
also contain AUTH LOGIN prior to sending credentials.

(Raubvogel/186050) LOGAN, that is what we are are doing right now; that is what I

Raubvogel@rss.iscabbs.com — Fri, 17 Feb 2012 07:10:00 -0600

        LOGAN, that is what we are are doing right now; that is what I
meant when I said we have a mail.domain.com server, which is the MX
record for domain.com. But, mail-test is my test box; I want to have
mail being sent to me@mail-test.monetra.com not go to the machine in the
MX record (and then have it forward it to mail-test) but to mail-test
directly.

        While we are talking about mail, shouldn't I be able to do tls
smtp auth testing by feeding the output i(base64-output-from-perl) of

perl -MMIME::Base64 -e   'print encode_base64("\000test-user\@mail-test.domain.
com\000passwd")'

to

openssl s_client -starttls smtp -crlf -connect mail-test.domain.com:25

I know my username and pw are right but when I tried it, I got:

250 DSN
AUTH PLAIN base64-output-from-perl
535 5.7.8 Error: authentication failed:
Q
DONE

(LOGAN/186048) Raubvogel> You might want to take a quick look at postfixadmin....

LOGAN@rss.iscabbs.com — Thu, 16 Feb 2012 13:10:00 -0600

Raubvogel>  You might want to take a quick look at postfixadmin.  It is
currently my preferred simple way to do true virtual hosting with postfix
without the hassle of some sort of crazy ISP panel.  I've used it for a long
time, and I am happy with it.

In order to get mail addressed to your subdomain to route directly to the
responsible server, you'll either need to create an MX record for the
subdomain, or adjust the config on the MTA responsible for the parent domain to
forward the mail to you.

(Copper Lethe/186047) You want postfix.

Copper Lethe@rss.iscabbs.com — Thu, 16 Feb 2012 12:52:00 -0600


You want postfix.
Unless you want sendmail.

(Raubvogel/186046) Feeling a bit blonde today regarding postfix. So I have one test

Raubvogel@rss.iscabbs.com — Thu, 16 Feb 2012 07:56:00 -0600

        Feeling a bit blonde today regarding postfix. So I have one test
machine, mail-test.domain.com, which I want to only handle emails sent to its
hostname but treat it as a virtual host setup. What is th ebest way to set that
up? Also, since there is a mail.domain.com server, which is the mx record for
domain.com, how can I make emails addressed to me@mail-test reach that machine?

(Feoh/186045) Yup, Postfix is the way and the light. My company sends literal...

Feoh@rss.iscabbs.com — Thu, 16 Feb 2012 06:07:00 -0600

Yup, Postfix is the way and the light.  My company sends literally millions of
E-mails a day using a fleet of Linux boxes running Postfix and our custom PHP
stack.

(Kena/186044) A lot more painful. Postifx is the way to fly these days, IMH...

Kena@rss.iscabbs.com — Wed, 15 Feb 2012 20:24:00 -0600

A lot more painful.  Postifx is *the* way to fly these days, IMHO.  If you're
still having trouble, feel free to ask -- we can get you going.  And trust me
-- Sendmail uses the m4 macro language for a reason.  Nothing against Eric
Allman, but there's a reason it's been supplanted in popularity by MTAs like
Exim, Postfix, and, yes, even qmail.  And Postfix, with a remarkably short
configuration file, can be made to sit up and beg.

(Kael/186043) Keogk: you want:

Kael@rss.iscabbs.com — Wed, 15 Feb 2012 16:15:00 -0600

Keogk: you want:
1) postfix
2) in main.conf, relayhost=blah
3) on the actual relayhost, mynetworks= whatever subnet

Maybe one or two other items if necessary. You can of course do with sendmail,
it's just more painful.

(Snuffy/186042) Egregious:

Snuffy@rss.iscabbs.com — Wed, 15 Feb 2012 13:05:00 -0600

Egregious:

and there's always 'tr'
http://en.wikipedia.org/wiki/Tr_%28Unix%29

though you've solved your immediate problem already

(Keogk/186040) Asked about this a few months ago but finally have some time to ...

Keogk@rss.iscabbs.com — Wed, 15 Feb 2012 10:01:00 -0600

Asked about this a few months ago but finally have some time to play with it
I have a test network at work with about 25 boxes. The network guys do not want
to create a rule so that every one of my 25 boxes can send an email
So I want 1 box that will recieve mail from the other 25 boxes then relay to my
 main mail server
So I am guessing I need to setup a mail relay?
I started playing with CentOS and postfix folling some guides but I have been
unable to get anything to happen
I switched my MTa to sendmail and I am seeing some logs about rejecting
Just want to make sure I am using the correct terms. What I wnt to do is relay
mail correct?

(Feoh/186039) Kael Ruby is only Perl-esque when you want it to be. The Perl f...

Feoh@rss.iscabbs.com — Wed, 15 Feb 2012 09:35:00 -0600

Kael Ruby is only Perl-esque when you want it to be.  The Perl features are
'syntactic sugar' and you can use them or not.

(Kael/186038) python feels like perl 8 years ago.

Kael@rss.iscabbs.com — Wed, 15 Feb 2012 08:17:00 -0600

python feels like perl 8 years ago.

It might be installed, but features are very version/installation specific.

I'm starting to learn ruby, which is sometimes perl-esque and sometimes
certainly not. Meh.

(Kena/186037) Yeah. I mean, Perl is -- kind of by definition -- a bit on the ...

Kena@rss.iscabbs.com — Wed, 15 Feb 2012 07:42:00 -0600

Yeah.  I mean, Perl is -- kind of by definition -- a bit on the bloated side,
but a) it really *is* installed most everywhere, and b) damn, but it makes
things easier.
I find Python to be more pleasing to the eye; I feel like I'm actually, y'know,
*coding* when I'm done with a Python project.  But golly, is Perl handy fixit
glue.

(Cyanide/186036) It is probably best anyway. sed was great when perl wasnt instal...

Cyanide@rss.iscabbs.com — Wed, 15 Feb 2012 07:25:00 -0600

It is probably best anyway. sed was great when perl wasnt installed by default,
It was more portable, and lighter weight but now that isn't that big of
an issue.

(Egregious/186035) JP> I broke down and used perl.

Egregious@rss.iscabbs.com — Wed, 15 Feb 2012 05:59:00 -0600

JP> I broke down and used perl.

perl -pe "s:><tr:>\n<tr:g"

(Kael/186034) lvm snapshots take up space depending on how much the source cha...

Kael@rss.iscabbs.com — Wed, 15 Feb 2012 05:44:00 -0600

lvm snapshots take up space depending on how much the source changes, snapshots
are essentially a bitmap of what has changed. One can grow a snapshot as needed
so start small and adjust based on use, monitor free space via lvs.

Now, if the volume group has no free space to allocate, you're doing it wrong.
:)

(Feoh/186033) AfterParty> I dunno your comfort level with Vim, but Vimpager wi...

Feoh@rss.iscabbs.com — Tue, 14 Feb 2012 15:35:00 -0600

AfterParty> I dunno your comfort level with Vim, but Vimpager will do all this
and more.

https://github.com/rkitover/vimpager

(John Public/186032) Egregious> This comes up a lot. sed can't be used if you need to...

John Public@rss.iscabbs.com — Tue, 14 Feb 2012 14:52:00 -0600

Egregious> This comes up a lot. sed can't be used if you need to process
newlines. Are you trying to insert newlines, or removed them?

Tap> Yes, &!foo|bar   will hide all lines matching foo|bar.
Remove the "!" and it will show lines matching foo|bar.

(The After Party/186031) "fold"

The After Party@rss.iscabbs.com — Tue, 14 Feb 2012 13:28:00 -0600

"fold"

  Does "less" or some other util allow you to 'fold'/'ignore'/'grep -v'
  while viewing a file? When looking thru log files, i went to 'grep -v' a
  large portion of the file so I can skip over the useless stuff and
  find what I'm looking for. Yeah, I could
  grep -v 'phrase' file | less
  then when I found another piece i wanted to ignore:
  grep -v 'phrase' file | grep -v 'phrase2' | less
  repeat...
  I'd rather avoid all the pipes and avoid exiting/entering the file.

(Kena/186030) Disallowing/disabling doesn't give enough context: X is simply b...

Kena@rss.iscabbs.com — Tue, 14 Feb 2012 13:15:00 -0600

Disallowing/disabling doesn't give enough context: X is simply being
*supplanted*.  Theoretically, it'll be able to act as a client to Wayland, but
I'll believe how well that does (or does not) work when I see it.
http://www.h-online.com/open/features/Wayland-Beyond-X-1432046.html
(As part of the commentary on LWN, someone pointed out Xpra, which is "screen
for X" -- apparently allows you to fire up an X client to one remote host, then
re-direct it to a *different* remote host at will.  Apparently (gad-zooks) this
even includes 'doze.  http://xpra.org/ )

(Egregious/186029) Is there a control-code for sed that it will interpret as a CRLF...

Egregious@rss.iscabbs.com — Tue, 14 Feb 2012 10:40:00 -0600

Is there a control-code for sed that it will interpret as a CRLF?

I can put a literal ^M in there, but I want what I do to be plain ASCII,
if possible.

sed -e 's/\>\<tr/\>\\n\<tr/' is putting in a literal \n
ADditional crazy layers of \\\\ or using "" instead of '' don't seem to fix it.

(Feoh/186028) Kena> But surely they're not completely disabling/disallowing X,...

Feoh@rss.iscabbs.com — Tue, 14 Feb 2012 09:20:00 -0600

Kena> But surely they're not completely disabling/disallowing X, right?

(Raubvogel/186027) I was going to suggest netcat but it does not seem to be install...

Raubvogel@rss.iscabbs.com — Tue, 14 Feb 2012 08:15:00 -0600

        I was going to suggest netcat but it does not seem to be installed by
default either.

(Kena/186026) Ahhh, Solaris. That explains it. Hmmm... surprising it doesn't...

Kena@rss.iscabbs.com — Tue, 14 Feb 2012 07:33:00 -0600

Ahhh, Solaris.  That explains it.  Hmmm... surprising it doesn't exist for
Linux (or, at least, isn't in the Ubuntu repo); go figger.  Next time I'm on a
Solaris box with issues, though, that could well come in handy.  (Doesn't
happen often at my current employer, but it *does* happen.)

(The After Party/186025) Kena

The After Party@rss.iscabbs.com — Tue, 14 Feb 2012 06:45:00 -0600

Kena

  Solaris 5.10
  nmap isn't installed, so I can't compare.
  You can google pping just like I could google nmap.

  My foggy memory of nmap, probably that that different in terms of
  making a connection to a specific port.

  The script I wrote checks the return code ($?) so pping might be better
  in this situation.

(Kena/186024) Wayland is coming, I guess. cries I do not want to lose ne...

Kena@rss.iscabbs.com — Tue, 14 Feb 2012 05:47:00 -0600

Wayland is coming, I guess.  *cries*  I do *not* want to lose network
transparency.  Even a little.  Yes, X is a 25-year-old piece of bloat, but I
really, really like being able to run remote apps seamlessly.

(Kena/186023) Yeah. We had a stack o' servers. One was being realllllly unpl...

Kena@rss.iscabbs.com — Tue, 14 Feb 2012 03:48:00 -0600

Yeah.  We had a stack o' servers.  One was being realllllly unpleasant, so I
finally decided it was in need of re-imaging.

I realized roughly, oh... 5 seconds in that I was imaging the *wrong* host.

It was about 10 minutes later that I realized that, for various reasons, it had
been pulled from the backup rotation.  That was A Bad Day.

(Danix/186022) Kena> Maybe I got lucky. I realized my mistake 1 second after pr...

Danix@rss.iscabbs.com — Tue, 14 Feb 2012 02:38:00 -0600

Kena> Maybe I got lucky. I realized my mistake 1 second after pressing enter,
then I yanked the power cord. The first superblock did get overweitten, but the
backup ones were still there. I'm pretty sure I lost some of the first files,
but didn't really check that out.

Even more lucky was chkdsk's ability to resoore the same thing happening on an
NTFS filesystem. It took a lot of time, and the ACLs were all b0rked, but the
files did reappear.

Oh, I no longer do repartitioning/formatting while sleep deprived as a security
measure as well.

(Kena/186021) Errr, okay. But, thus-far, you've neglected to say which platfo...

Kena@rss.iscabbs.com — Tue, 14 Feb 2012 02:39:00 -0600

Errr, okay.  But, thus-far, you've neglected to say which platform you are/it
is on, and, if it's Linux, as I've never heard of it, I'll assume it's not a
stock install.  And telling me that nmap isn't installed doesn't really say
anything about how it's similar or different from pping.  Etc.

(The After Party/186020) Kena

The After Party@rss.iscabbs.com — Mon, 13 Feb 2012 13:22:00 -0600

Kena

  One big difference:

-bash: nmap: command not found

  hard to get shit installed around here.

(John Public/186019) Kena> $_ is not touched, so it keeps the same value it had befor...

John Public@rss.iscabbs.com — Mon, 13 Feb 2012 12:41:00 -0600

Kena> $_ is not touched, so it keeps the same value it had before the loop.
You can do the same thing with other places where $_ is used.

    for my $x (@arr)    ## $_ is not assigned. $x is.

    -f $x               ## $_ is not stat'd. $x is.

    $x =~ m/blah/gc     ## pos($_) is not updated. pos($x) is.

....etc.

(Kena/186018) It's not in the Ubuntu repo; how does it differ from nmap?

Kena@rss.iscabbs.com — Mon, 13 Feb 2012 12:36:00 -0600

It's not in the Ubuntu repo; how does it differ from nmap?

(The After Party/186017) firewall testing

The After Party@rss.iscabbs.com — Mon, 13 Feb 2012 10:31:00 -0600

firewall testing

  Thought I posted here, maybe it scrolled
  Anyways, co-worker told me about pping (port ping).
  This seems to be a good general tool to test connectivity so I can
  script whether a firewall request was completed successfully.
  Thought I'd share.

(Lochner/186016) If anybody has any experience with a Novell license audit, speci...

Lochner@rss.iscabbs.com — Mon, 13 Feb 2012 10:27:00 -0600

If anybody has any experience with a Novell license audit, specifically
Novell's "LMS Assist" product, please Mail> me.

(Kena/186015) Errrrrr...

Kena@rss.iscabbs.com — Mon, 13 Feb 2012 10:23:00 -0600

Errrrrr...
.... $p=<STDIN> ...
I didn't know you could *do* that. What happens to $_ ?

(John Public/186014) sed '/^Reply-To:/q; /^From:/h; /./d;g;q'

John Public@rss.iscabbs.com — Mon, 13 Feb 2012 09:07:00 -0600

sed '/^Reply-To:/q; /^From:/h; /./d;g;q'

Direct translation to Perl:

    my $p, $hold;
    while ($p = <STDIN>){
        chomp($p);
        if ($p =~ /^Reply-To:/){
            quit();
        }

        if ($p =~ /^From/){
            $hold = $p;
        }

        if (length($p) == 0)
            next;

        $p = $hold;
        quit();
    }

    sub quit { print "$p\n"; exit; }

(Kena/186013) Heh. If I spoke sed, I'd give it a go. It's close enough to Pe...

Kena@rss.iscabbs.com — Mon, 13 Feb 2012 08:35:00 -0600

Heh.  If I spoke sed, I'd give it a go.  It's close enough to Perl that I
*almost* know what it's trying to do, but I'd be guessing, and the last bit
just plain loses me.  This, honestly, is why I don't bother with sed: because
to make it truly powerful, it needs to meld with awk (or something like it),
and then you might as well just have both, and more, with Perl.
$.02

(Raubvogel/186012) Ok, I admit I suck at regexp. Can anyone explain this example fr...

Raubvogel@rss.iscabbs.com — Mon, 13 Feb 2012 08:18:00 -0600

        Ok, I admit I suck at regexp. Can anyone explain this example from
http://www.eng.cam.ac.uk/help/tpl/unix/sed.html:

 # get return address header
 sed '/^Reply-To:/q; /^From:/h; /./d;g;q'

(Kena/186011) You recovered from mkfs on the wrong partition?? Wow. I'd be c...

Kena@rss.iscabbs.com — Sun, 12 Feb 2012 21:17:00 -0600

You recovered from mkfs on the wrong partition??  Wow.  I'd be curious to know
how you managed *that*.

(Danix/186010) fsck has fixed truly b0rked filesystems in the past, so I'd be w...

Danix@rss.iscabbs.com — Sun, 12 Feb 2012 12:46:00 -0600

fsck has fixed truly b0rked filesystems in the past, so I'd be wary to use
fsck-less btrfs. I've had lots of weird stuff happen (mostly on laptops) and a
coule of bloopers like mkfs on the wrong partition. All of there, recovered by
fsck.

Not havingg gsck makes me antsy...
            ^fsck

(Kena/186009) "It depends." Largely, on the filesystem, and what can go wrong...

Kena@rss.iscabbs.com — Sun, 12 Feb 2012 06:54:00 -0600

"It depends."  Largely, on the filesystem, and what can go wrong.  Really, with
a "conventional" filesystem, you've got two things that can go wrong: the
allocation table can get out of sync with reality (e.g., unclean mounts), or
the allocation table can get corrupted -- which is where the real fun comes in.
Usually, duplicate tables are kept, or, worst-case, you find the files and
reverse-create the allocation table... which, especially with non-contiguous
files, is fraught with issues.

Now, btrfs, for example, lets you do do your reverse-creation of metadata much
more easily, as there are backreferences in the file's data (or something like
that; it's a bit hazy).  But its btree nature also means that there's lots more
to how the filesystem is put together, and lots more to go wrong; something
botches somewhere in the tree, and suddenly, your dependencies got issues. 
fsck -- especially with redundant metadata -- should be able to fix that easily
(indeed, maybe the redundant metadata part is what gets fixed by kernel 3.2 and
higher, since that's pretty braindead: if the checksum's valid, the data's
valid; pull it from the redundant copy, update the invalid metada, call it a
day).

(John Public/186008) I guess it was never really clear to me what fsck does. I thoug...

John Public@rss.iscabbs.com — Sun, 12 Feb 2012 04:35:00 -0600

I guess it was never really clear to me what fsck does.  I thought it just
dealt with the partially sync'd data from unclean unmounts, such as due to
power failures.

(Kena/186007) Hmmm. I realize I didn't state correclty the full extent of how...

Kena@rss.iscabbs.com — Sun, 12 Feb 2012 02:57:00 -0600

Hmmm.  I realize I didn't state correclty the full extent of how badly fsck is
needed.  Because, when something goes medium-to-badly wrong with btrfs, it's
time for triage -- fix what you can, give up on what you can't, and make the
fs valid one way or another.  But... without fsck, there's the more-or-less
weekly e-mail from someone whose FS is probably more-or-less okay, but
unmountable.  (Indeed, I have an image around somewhere that meets this exact
description.)  And without fsck, unmountable it will remain.  While you can
sometimes get help from the list, there *needsneedsneeds* to be a generic tool
to deal with these situations, especially before production rollout can even be
seriously considered.

As for what taking snapshots has to do with being low on space and needing a
new drive, I honestly don't know.  But one cool thing that btrfs does is that,
with small enough files (e.g., a huge number of your "4K" files), it stores the
data and the metadata in the same inode, which is really handy both performance
and space-wise.

(River Rat/186006) kill uses -15 by default.

River Rat@rss.iscabbs.com — Sat, 11 Feb 2012 16:33:00 -0600

kill uses -15 by default.
kill -9 force kills a process.  But it -9 wont work on zombied processes.

(Cyanide/186005) Has anyone successfully installed ros on Fedora? Im having issue...

Cyanide@rss.iscabbs.com — Sat, 11 Feb 2012 16:26:00 -0600

Has anyone successfully installed ros on Fedora? Im having issues, the
installer is broken..

(Cyanide/186004) true. but they don't take up that much space. :)

Cyanide@rss.iscabbs.com — Sat, 11 Feb 2012 16:26:00 -0600

true. but they don't take up that much space. :)

(River Rat/186003) But doesn't snapshots take up room? If you are short on space a...

River Rat@rss.iscabbs.com — Sat, 11 Feb 2012 16:22:00 -0600

But doesn't snapshots take up room?  If you are short on space and can't afford
another drive, then you can't do snapshots.

(Kael/186002) One should always try kill before kill -9.

Kael@rss.iscabbs.com — Sat, 11 Feb 2012 06:52:00 -0600

One should always try kill before kill -9.
Difference (man kill for more):
kill by default sends a TERM signal or "please cleanup and stop process".
kill -9 sends a KILL signal, meaning "stop this process NOW".

Sometimes a kill -9 is needed, but using it by default is a bad habit (kind of
like rm -f). Hmm, maybe I'll add this to my interview questions. :)


For users of various Linux distros, take a look at pgrep and pkill for use
instead of killall. Take note though, I found at least one case where pgrep
failed to find a process where ps aux | grep did for the same string.

(Kena/186001) Me? I'm fine with chrome, itself. Don't know why it would beco...

Kena@rss.iscabbs.com — Fri, 10 Feb 2012 09:53:00 -0600

Me?  I'm fine with chrome, itself.  Don't know why it would become
unresponsive; there were some issues there a couple years ago with browsers,
fsync(), and delayed writes, but all that stuff was polished off shortly after
ext-4 was released.  If you want to start learning The Ways of the Force, well,
you could try to *see* what it's doing when it's unresponsive.
ps auxww | grep -i iron    [To get the Program ID ("pid") of the process]
strace -s 1024 -f -o /tmp/irontrace.log -p <pid>  [Where <pid> is your PID]
Let that run for 10 or 15 seconds.  You'll now have a file, /tmp/irontrace.log,
which will have ALL SORTS of completely irrelevant information, but, possibly,
some nuggets of what's going on.  There will also very likely be several red
herrings -- e.g., "connection timed out", "file not found", which aren't
(usually) errors, but, rather, the normal state of affairs as the system
wanders around, figuring stuff out.  And it may tell you absolutely nothing of
interest, which is probably what happens with me 50% of the time.  But if
you're wondering what something's doing, "strace" is the easiest way for a
non-developer to get their feet wet.

(Mild Seven/186000) Thanks Kena. killall -9 iron worked well.

Mild Seven@rss.iscabbs.com — Fri, 10 Feb 2012 08:53:00 -0600


Thanks Kena.  killall -9 iron worked well.
For some reason Iron browser (after being installed for a few weeks) likes to
become nonresponsive.  Do you recommend a different browser? :-)

(Kena/185999) Or... you can cheat, by asking here.

Kena@rss.iscabbs.com — Fri, 10 Feb 2012 08:48:00 -0600

Or... you can cheat, by asking here.
killall chrome
if that somehow, magically, doesn't do the trick
killall -9 chrome
(Note that the only reason I know the process name is "chrome" is because I
looked on my system.  Further note that, if you do this as root, it'll kill
chrome processes owned by any/everyone, not just you, in case you were
magically on a multi-user machine.  LASTLY, note that "killall" does something
VERY DIFFERENT on SysV machines -- it's part of the final shutdown procedure,
and used to kill all processes.)

(Feoh/185998) Try 'grep -i' maybe you're getting the case wrong.

Feoh@rss.iscabbs.com — Fri, 10 Feb 2012 08:43:00 -0600

Try 'grep -i' maybe you're getting the case wrong.

(Mild Seven/185997) I tried that (I use the Iron browser which is basically Chrome)....

Mild Seven@rss.iscabbs.com — Fri, 10 Feb 2012 08:41:00 -0600


I tried that (I use the Iron browser which is basically Chrome).  Here's what
it returns on the terminal:

linux@linux-desktop:~$ ps auxw | grep Chrome
linux     2867  0.0  0.0   4440   780 pts/2    S+   02:40   0:00 grep
--color=auto Chrome
linux@linux-desktop:~$ ps auxw | grep Iron
linux     3073  0.0  0.0   4440   780 pts/2    S+   02:41   0:00 grep
--color=auto Iron
linux@linux-desktop:~$

(Feoh/185996) ps auxw | grep Chrome

Feoh@rss.iscabbs.com — Fri, 10 Feb 2012 08:39:00 -0600

ps auxw | grep Chrome
Look for the process ID (first column) of what it spits back.  Then 'kill -9
<pid>'

(Mild Seven/185995) My Chrome browser quit responding in Ubuntu. How can I kill it u...

Mild Seven@rss.iscabbs.com — Fri, 10 Feb 2012 08:36:00 -0600


My Chrome browser quit responding in Ubuntu. How can I kill it using the
terminal?  Sorry, I'm bit of a noob.

TIA.

(Egregious/185994) The answer is "py-setuptools."

Egregious@rss.iscabbs.com — Fri, 10 Feb 2012 06:18:00 -0600


The answer is "py-setuptools."

(Kena/185993) That's why I want a search engine that takes regular expressions...

Kena@rss.iscabbs.com — Fri, 10 Feb 2012 05:38:00 -0600

That's why I want a search engine that takes regular expressions.  Boy, that'd
be cool.  And also why I miss AltaVista's "near" parameter, which was really
handy when trying to weed out false positives.

(Egregious/185991) Where is python's "ez_install" or "easy_install" package under O...

Egregious@rss.iscabbs.com — Fri, 10 Feb 2012 05:19:00 -0600


Where is python's "ez_install" or "easy_install" package under OpenBSD?

Search engines really hate looking for things with underscores in them.

(Kena/185990) /boot works fine with btrfs; that's been addressed since Grub 2....

Kena@rss.iscabbs.com — Fri, 10 Feb 2012 04:05:00 -0600

/boot works fine with btrfs; that's been addressed since Grub 2.0, so a good
year.  As for EFI, I'm not sure what you mean; I imagine if it works for Linux
at all, it works for the primary FS's, no?  (I could well be missing
something.)  As for *detecting* corruption, it does that extraordinarily well,
as both data and metadata are checksummed.  So, great: you know that something
is broke.  But... how do you fix it?  Answer, right now, is "Try to mount it
with kernel 3.2 or later, as there are now inherent smarts in there."  But
there's no tool to sit there and sift through, figuring out how to actually
*fix* everything that could reasonably be expected to get fixed.

(Cyanide/185989) I think you can detect corruption, because it does wonky things,...

Cyanide@rss.iscabbs.com — Fri, 10 Feb 2012 03:17:00 -0600

I think you can detect corruption, because it does wonky things, but there is
nothing to -fix- the corruption. IE some id10t pulls the plug on your server,
removes a disk from the bay, or something else goes flaky.. There is nothing to
help get your system back online, short of rebuilding it, or -maybe- if you are
lucky, you can get a snapshot reverted, but if the filesystem is corrupt, then
reverting it might not work either.

Which is kind of why I use a portable for testing, since it can get shut off
unsafely, ie the kid hits the button,  go into sleep mode, the disk can get
jarred, etc. Since it does get backed up, the lost data usually isn't the
issue. But it is a lot more "hostile" environment then a server sitting in a
rack with UPS backup on a raid system.

(John Public/185988) I don't get the `fsck` problem. Is it that there is no way for t...

John Public@rss.iscabbs.com — Thu, 09 Feb 2012 17:59:00 -0600

I don't get the `fsck` problem. Is it that there is no way for the filesystem
to detect corruption, or is it that there is no tool called `fsck` for admins
to run?

(Cyanide/185987) Did they mention anything about /boot working with btrfs? or the...

Cyanide@rss.iscabbs.com — Thu, 09 Feb 2012 15:32:00 -0600

Did they mention anything about /boot working with btrfs? or the EFI partition
working with btrfs?   Those are the biggest issues beyond fsck.

However I am guessing most of it is done. just put on hold.. I just
hope it doesnt become a mess like grub2 and UEFI, etc... However, they
are usually decent at fixing the messes..

(Kena/185986) For my data, btrfs ain't too shiny. For production data? Wel...

Kena@rss.iscabbs.com — Thu, 09 Feb 2012 14:05:00 -0600

For *my* data, btrfs ain't too shiny.  For production data?  Well, f*** -- it
*still* doesn't have an fsck, though one is rumored Any Day Now.  (Oracle wants
btrfs to be the default FS for their next Oracle-branded Linux release; they
don't want to do that without an fsck, and it's scheduled to ship within the
next week or two.  I think we'll either see an fsck, or a release of the code
so's others can hack it; right now, Chris hasn't made the fsck code public, for
fear (he says, and I believe him) that it's incomplete, and the only thing
worse than a dead-but-probably-recoverable filesystem is a
dead-and-now-utterly-destroyed filesystem.

(Kael/185985) One could create/manage lvm snapshots on a regular basis for use...

Kael@rss.iscabbs.com — Thu, 09 Feb 2012 10:57:00 -0600

One could create/manage lvm snapshots on a regular basis for use strictly to
recover files. Can even add extents as needed. That said, I'm a fan of amanda
for this purpose.

btrfs: Sorry, too shiny for my data right now.

(Feoh/185984) Mandel> Heh! You're a better man than I. I just installed it, h...

Feoh@rss.iscabbs.com — Thu, 09 Feb 2012 09:04:00 -0600

Mandel> Heh! You're a better man than I.  I just installed it, have been using
and loving it, and am a trusting soul that it all works as advertised :)  The
Git support alone is awesome (It will telegraph whether my repo is clean or
dirty on the command line, as well as time since last commit and all manner of
other grooviness).

(John Public/185983) This creates a filesystem tree. From there, you can use various ...

John Public@rss.iscabbs.com — Thu, 09 Feb 2012 07:58:00 -0600

This creates a filesystem tree. From there, you can use various existing tools
to browse the tree.

$ cat /tmp/treedata.txt | \
    while read p x c; do
        d=`find -name $p`; [ -d "$d" ] || d=$p; mkdir -p $d/$c
    done


$ tree
+-- 6047214154f3403019a500
    +-- 13022515214f340301c6414
        |-- 19712802644f340301dea83
        |-- 21285559034f340301cb246
        |   +-- 13214261184f340301d01ee
        |       +-- 916861584f340301d4e5d
        +-- 9901684734f340301d9c90

(Egregious/185982) Answer for the archives: I dumped it into a database and used a ...

Egregious@rss.iscabbs.com — Thu, 09 Feb 2012 07:12:00 -0600

Answer for the archives: I dumped it into a database and used a bunch of
self-joins to build it out.  A bit ugly but it got the job done.

(Egregious/185981) I've got some 50,000 lines of the form:

Egregious@rss.iscabbs.com — Thu, 09 Feb 2012 07:00:00 -0600

I've got some 50,000 lines of the form:

6047214154f3403019a500 : 13022515214f340301c6414
13022515214f340301c6414 : 21285559034f340301cb246
21285559034f340301cb246 : 13214261184f340301d01ee
13214261184f340301d01ee : 916861584f340301d4e5d
13022515214f340301c6414 : 9901684734f340301d9c90
13022515214f340301c6414 : 19712802644f340301dea83

This is a debug of a tree structure.  Before I spend an hour re-inventing a
wheel, is there some unix-like tool that will make this into a tree for me?

(Kena/185980) Good article (PDF): https://db.usenix.org/publications/login/201...

Kena@rss.iscabbs.com — Thu, 09 Feb 2012 06:41:00 -0600

Good article (PDF): https://db.usenix.org/publications/login/2012-02/openpdfs/B
acik.pdf errr
https://db.usenix.org/publications/login/2012-02/openpdfs/Bacik.pdf
Also makes a point I forgot to mention: with btrfs, you don't need separate
storage for your snapshots; they all reside in the same partition.

(Kena/185979) sigh Sadly, no, it doesn't do individual files, like, say, WA...

Kena@rss.iscabbs.com — Thu, 09 Feb 2012 05:12:00 -0600

*sigh*  Sadly, no, it doesn't do individual files, like, say, WAFL on... on...
whoever that company is.  But what I do is I run a snapshot on Mon., Wed., Fri.
on my work computer, and weekly on my relatively static home computers.  Since
it's all CoW, the only big hit is if I delete something really big, and then
create something else really big, before my snapshots expire (I keep four of
them, so just over a week, at work).  (NetApp!  That's the company.  Anyway...)
That being said, I have written wrappers for some commonly edited files, such
that timestampped clones are created, which use essentially only delta extra
space.  *ponders*  Hmmm... I wonder if there's a way one could use some of that
kernel foo to create per-file snapshots for *every* write.  That would be
intesting.

(John Public/185978) Kena> It does snapshotting of individual files? And you don't ha...

John Public@rss.iscabbs.com — Thu, 09 Feb 2012 04:07:00 -0600

Kena> It does snapshotting of individual files? And you don't have to tell it
to do that in advance?

The lvm2 snapshotting works like this, "Oops, I forgot to take a whole system
backup before I edited a config file!", which is not so handy.

(Knightshade/185977) depends on what a casual user needs.

Knightshade@rss.iscabbs.com — Thu, 09 Feb 2012 03:22:00 -0600

depends on what a casual user needs.

My house isn't an enterprise environment, but if I ever move my media server
over to *nix it's gonna run ZFS so the media will live on a RAID Z2 setup.

(Kena/185976) There's one feature in btrfs that's hard to overstate the cool...

Kena@rss.iscabbs.com — Wed, 08 Feb 2012 23:29:00 -0600

There's *one* feature in btrfs that's hard to overstate the coolness of -- but
it does take some "getting to know you" time.  Simply put, snapshots rock. 
"Wups!  I didn't mean to delete that file!" Well, not a problem.  "Wow, that
upgrade sure didn't work."  Not a problem.  "Whoah, I really messed up that
config file, and didn't make a copy."  Not a problem.  The only time I've had
issues dealing with a wholesale reversion is when the boot stuff gets tweaked;
single point of failure, there, I'm afraid.

(MrPiolo/185975) In my experience, ext4 feels more responsive on the desktop. Tri...

MrPiolo@rss.iscabbs.com — Wed, 08 Feb 2012 19:25:00 -0600


In my experience, ext4 feels more responsive on the desktop. Tried it with dif.
kernels and distros, and ext4 just felt faster launching apps, redrawing
windows, bootings, etc.

(John Public/185974) Mild Seven> No, you're not missing anything. Even the difference...

John Public@rss.iscabbs.com — Wed, 08 Feb 2012 14:33:00 -0600

Mild Seven> No, you're not missing anything. Even the difference between ext2
and ext3 only matters if you reboot frequently, and don't care to watch an
fsck spin on a few hundred GB. :)

Do any of the new file systems offer features that would matter to a casual
user? They're all enterprise features, right?

(Mandel/185973) Oh My Zsh>> Oh... That's a lot of a really weird plugins that ...

Mandel@rss.iscabbs.com — Wed, 08 Feb 2012 13:18:00 -0600

Oh My Zsh>>  Oh...  That's a lot of a really weird plugins that I don't know.
This is going to take some time.

(Mild Seven/185972) When I installed Lubuntu 11.10 I chose ext3. Am I missing out on...

Mild Seven@rss.iscabbs.com — Wed, 08 Feb 2012 12:58:00 -0600


 When I installed Lubuntu 11.10 I chose ext3. Am I missing out on anything as a
causal home user by not using ext4?

(River Rat/185971) The Fedora / Red Hat developers working on the Beefy Miracle are...

River Rat@rss.iscabbs.com — Wed, 08 Feb 2012 11:36:00 -0600

The Fedora / Red Hat developers working on the Beefy Miracle are tentatively
moving ahead with their plan to use Btrfs as the default Linux file-system for
Fedora 17 and beyond.

Yesterday was another Fedora Engineering Steering Committee (FESCo) where the
committee discussed the latest round of F17 features, after approving an
impressive amount of features for this next Red Hat-sponsored Linux
distribution release. With the feature freeze being today (7 February),
yesterday was the last FESCo meeting where new Fedora 17 features were
approved.

The unified /usr file-system still has the go-ahead, English Typing Booster,
firewall-d as the default firewall solution, and Network Zones were all of the
Fedora 17 features discussed and approved yesterday. But perhaps most
interesting is that Btrfs as the default file-system was brought up again.

The summary/minutes for yesterday's FESCo meeting states "AGREED: We'll try
btrfs by default as long as it lands in alpha - if not, push to F18." So Fedora
developers have until the Fedora 17 Alpha release to switch from EXT4 to Btrfs
by default otherwise it will happen with Fedora 18. Fedora 17 Alpha is set to
be released at the end of the month (28 February).

Those not familiar with the Fedora 17 Btrfs file-system proposal, the
information is available from the Fedora Project Wiki. The scope is to enable
Btrfs as the default file-system on new Fedora Linux installations rather than
EXT4. Btrfs has long been available as an install-time option in Fedora (and in
Ubuntu, openSUSE, and other Linux distributions) but would be the first
tier-one Linux distribution switching to the file-system by default.

Btrfs support in Fedora has been interesting going all the way back to the
Fedora 13 release where it gained support for Linux system roll-backs by taking
advantage of the Btrfs copy-on-write snapshots within Yum.

Btrfs was supposed to become the default file-system in Fedora 16 but in the
end the switch was postponed since a proper fsck utility for Btrfs was not
ready, among other blocking bugs.

A proper Btrfs fsck utility is imminent (by next week) as Oracle is making
Btrfs production-ready within their next Oracle Enterprise Linux release and
SUSE Linux Enterprise Server is also needing btrfs.fsck for their next
enterprise release in order to correct any file-system errors.

On the Red Hat BugZilla is the Btrfs tracking bug for Fedora. There's still
seven open bugs while three have been closed. Among the open Btrfs bugs
remaining is poor performance using Btrfs for VM storage, the lack of a working
fsck (what should be fixed soon), support for creating Btrfs images with
livecd-tools package, support for "various special things" Btrfs does in the
installer, verifying Btrfs support works properly, a buffer overflow in Btrfs
if the device name is too long, and a particular Btrfs file-system
corruption/crash situation.

If these Btrfs bugs are worked out by the alpha release in about two weeks --
along with the expected premiere of the new Btrfs fsck -- and Btrfs flipped on
within the Anaconda installer, we will have this next-generation open-source
file-system as the default in Fedora 17. If it doesn't happen for the "Beefy
Miracle" release, it should definitely be a good candidate for Fedora 18
considering the remaining open issues have been whittling away.

Fedora 17 final is due for an official release in early May with lots of new
features. Below is the relevant FESCo IRC discussion from yesterday regarding
Btrfs.

(Feoh/185970) Oh My Zsh has been the biggest productivity boost I've seen in f...

Feoh@rss.iscabbs.com — Wed, 08 Feb 2012 06:00:00 -0600

Oh My Zsh has been the biggest productivity boost I've seen in forever.
Especially if you do development with OSX, Git, or Ruby, but even if you don't.
It's amazing.
https://github.com/robbyrussell/oh-my-zsh

(Cyanide/185969) ACPI has been supporting for throttling for a number of years. I...

Cyanide@rss.iscabbs.com — Tue, 07 Feb 2012 11:08:00 -0600

ACPI has been supporting for throttling for a number of years. I almost
always disable it because it has a habit of being flaky with various
unix, depending on the implementation. It throttles but it is kind of
like throwing a bandaid on a ruptured artery.

(Danix/185968) IIRC, most motherboards now support ACPI and can throttle the sa...

Danix@rss.iscabbs.com — Tue, 07 Feb 2012 06:22:00 -0600

IIRC, most motherboards now support ACPI and can throttle the same way laptops
do. Hell, ACPI will even detect your USB-connected UPS and show that as your
desktop's "battery"!

(Wang Master/185967) Jupiter is just an applet to interface with the exisitng power m...

Wang Master@rss.iscabbs.com — Mon, 06 Feb 2012 15:16:00 -0600

Jupiter is just an applet to interface with the exisitng power management
capabilities built into the kernel.
 '
in most cases, modern cpus can pretty much all do voltage and clock speed
management as part of the cpufreq kernel interface and it's usually turned on
by default already.  Jupiter just let's you manipulate it manually.

i have a core i7-2760qm in my laptop and my cores are usually at 800mhz.

(Mild Seven/185966) I guess Jupiter really does work. With "power savings" on or "m...

Mild Seven@rss.iscabbs.com — Mon, 06 Feb 2012 10:49:00 -0600


I guess Jupiter really does work.  With "power savings" on or "maximum
performance" I get the following values respectively:

linux@linux-desktop:~$ more /proc/cpuinfo | grep MHz
cpu MHz         : 1120.500
linux@linux-desktop:~$ more /proc/cpuinfo | grep MHz
cpu MHz         : 1795.500

It's a neat program. When I'm on the bbs or IRC, no need to have it set
to maximum performance.. maybe save 10 or 20 watts.

(John Public/185965) yes. gnome 3 fedora 16.

John Public@rss.iscabbs.com — Mon, 06 Feb 2012 10:35:00 -0600

yes. gnome 3 fedora 16.

(River Rat/185964) Gnome 3?

River Rat@rss.iscabbs.com — Mon, 06 Feb 2012 10:21:00 -0600

Gnome 3?
You didn't specify.

(John Public/185963) Gnome app icons: How do I see the properties? I want to modify t...

John Public@rss.iscabbs.com — Mon, 06 Feb 2012 10:10:00 -0600

Gnome app icons: How do I see the properties? I want to modify the command
line for one of the apps. I've tried right/left/middle/shift/alt/ctrl clicking
the friggin thing.

(Mild Seven/185962) I have an old desktop PC (AMD Duron cpu) and installed Jupiter (...

Mild Seven@rss.iscabbs.com — Mon, 06 Feb 2012 09:50:00 -0600


I have an old desktop PC (AMD Duron cpu) and installed Jupiter (power
management software) in Lubuntu.

I set it to "power savings" and the temperature of the cpu dropped from 45C to
41C.

I thought it is only for laptops that can throttle down the processor when on
battery?

(Dr Doctor/185961) What you see in Finder is exactly this, minus those things that ...

Dr Doctor@rss.iscabbs.com — Sun, 05 Feb 2012 13:02:00 -0600


What you see in Finder is exactly this, minus those things that have the "don't
show in Finder" attribute (which can be set with setFile(1).

[cbell@athena ~]$ ls -lF /
total 30453
drwxrwxr-x+ 99 root  admin      3366 Feb  2 18:48 Applications/
drwxrwxr-x  15 root  admin       510 Jul 24  2011 Developer/
drwxr-xr-x+ 69 root  wheel      2346 Feb  4 21:43 Library/
drwxr-xr-x@  2 root  wheel        68 Jun 18  2011 Network/
drwxr-xr-x+  4 root  wheel       136 Feb  4 21:57 System/
lrwxr-xr-x   1 root  admin        60 Oct  6  2010 User Guides And Information@
-> /Library/Documentation/User Guides and Information.localized
drwxr-xr-x   5 root  admin       170 Jan 17 10:04 Users/
drwxrwxrwt@  5 root  admin       170 Feb  5 17:10 Volumes/
drwxr-xr-x@ 39 root  wheel      1326 Feb  4 21:43 bin/
drwxrwxr-t@  2 root  admin        68 Jun 18  2011 cores/
dr-xr-xr-x   3 root  wheel      4274 Feb  5 17:09 dev/
lrwxr-xr-x@  1 root  wheel        11 Jul 20  2011 etc@ -> private/etc
dr-xr-xr-x   2 root  wheel         1 Feb  5 17:26 home/
-rw-r--r--@  1 root  wheel  15566052 Jan 12 20:49 mach_kernel
dr-xr-xr-x   2 root  wheel         1 Feb  5 17:26 net/
drwxr-xr-x@  6 root  wheel       204 Jul 20  2011 private/
drwxr-xr-x@ 62 root  wheel      2108 Feb  4 21:44 sbin/
lrwxr-xr-x@  1 root  wheel        11 Jul 20  2011 tmp@ -> private/tmp
drwxr-xr-x@ 15 root  wheel       510 Oct  8 17:16 usr/
lrwxr-xr-x@  1 root  wheel        11 Jul 20  2011 var@ -> private/var
[cbell@athena ~]$

(Mac OS X 10.7.3)

(River Rat/185960) I am reading this article.

River Rat@rss.iscabbs.com — Sun, 05 Feb 2012 12:55:00 -0600

I am reading this article.
http://www.osnews.com/story/25556/Understanding_the_bin_sbin_usr_bin_usr_sbin_S
plit

Anyone have a mac so I can see the root directory structure?

Solutions so far tend to be just layers upon layers upon layers to hide the
mess of the directory structure. Mac OS X is especially egregious in this
regard - open a terminal and check the directory listing at root - it's an even
bigger mess than regular UNIX. Load up a finder window, and you're looking at
an entirely different directory structure.

Can some one post this directory structure in here please?

(River Rat/185959) I was asking about the x.509 thing, because I wanted to generate...

River Rat@rss.iscabbs.com — Sun, 05 Feb 2012 10:33:00 -0600

I was asking about the x.509 thing, because I wanted to generate my own keu for
the amazon EC2.
Just wondered how it was done.
Now come to think about it. I have seen it done in openssl.
It has just been a while and it floated my mind.

(River Rat/185958) I use epic irc client.

River Rat@rss.iscabbs.com — Sun, 05 Feb 2012 10:32:00 -0600

I use epic irc client.
It is term based.

(LOGAN/185957) irssi provides backscroll funtionality in each of it's "windows"...

LOGAN@rss.iscabbs.com — Sun, 05 Feb 2012 08:46:00 -0600

irssi provides backscroll funtionality in each of it's "windows" much like
screen.  It also allows for display of multiple "windows" within a single
terminal.  I used xchat for a long time, but eventually settled on irsii since
it is quite nice, and doesn't require X.

(Mild Seven/185956) I installed XChat instead. Problem solved. :-) Although I wish...

Mild Seven@rss.iscabbs.com — Sun, 05 Feb 2012 03:06:00 -0600


I installed XChat instead.  Problem solved. :-)  Although I wish I could still
use IRC the way I did in the "good old days" of the early '90s.

(Mild Seven/185955) I'm trying to use IRC for the first time in a terminal since '97...

Mild Seven@rss.iscabbs.com — Sat, 04 Feb 2012 13:04:00 -0600


I'm trying to use IRC for the first time in a terminal since '97.

When typing "/list" to get a list of channels, the screen scrolls through the
list without stopping.

I thought that typing /list |more would make it pause.  What is the command to
make it pause so I can actually stop the screen from scrolling through the
hundreds of IRC channels?

thanks in advance..

(Danix/185954) Yup, the first instruction I gave will generate a self-sign cert...

Danix@rss.iscabbs.com — Sat, 04 Feb 2012 07:48:00 -0600

Yup, the first instruction I gave will generate a self-sign cert, but will add
the CA:TRUE X509 extension required if that's what he wants. :)

Some of the places where I've been working as of lately have their own internal
CA for intranet stuff, so even production servers will have certs signed by an
internal CA, instead of Verisign. Of course, internet-facing stuff will have a
Verisign-signed one. :)

(The After Party/185953) Danix

The After Party@rss.iscabbs.com — Sat, 04 Feb 2012 07:39:00 -0600

Danix

  CA=Certificate authority

  If you want this to be a test cert, you can sign it yourself.
  Your browser/etc will not automatically trust a cert you've self-signed,
  but its free and easy to get around.

  If this is a real cert, you'll want verisign/etc to sign it.
  Er, be the CA

(Danix/185952) Want a CA?

Danix@rss.iscabbs.com — Sat, 04 Feb 2012 06:09:00 -0600

Want a CA?

openssl req -extensions v3_ca -sha1 -keyout private/cakey.pem -newkey rsa:2048
-new -x509 -days 3650 -out cacert.pem

Want to generate a CSR to get signed?

openssl genrsa -out mykey.pem 2048
openssl req -key mykey.pem -new -out mykey.req

send that (mykey.req) for signing. Or, if you want to sign it with your
newly-generated CA, after setting up that one:

openssl ca -md sha1 -in mykey.req -out mykey.cer

just remember to shave off the descriptive text in the resulting mykey.cer

(Justin Case/185951) OpenSSL should do that for you. Check this here link out:

Justin Case@rss.iscabbs.com — Sat, 04 Feb 2012 03:49:00 -0600

OpenSSL should do that for you.  Check this here link out:

http://www.ipsec-howto.org/x595.html

(River Rat/185950) Is there a linux app that will generate a X.509 Certificate?

River Rat@rss.iscabbs.com — Fri, 03 Feb 2012 15:37:00 -0600

Is there a linux app that will generate a X.509 Certificate?

Is that generated by openssl?

(Cyanide/185949) USB> Here is a better explanation of what is going on with the p...

Cyanide@rss.iscabbs.com — Fri, 03 Feb 2012 04:23:00 -0600

USB> Here is a better explanation of what is going on with the power
management software.
http://www.mjmwired.net/kernel/Documentation/usb/power-management.txt
It gives an explanation of what they are doing in the previous website.

RRat> Im not sure I really like systemd. I understand what they are trying to
do.. It seems kludgy, and I think solaris got kludgy when they started to
introduce it also.. Im not sure Solaris is an OS I would try to copy... :)

Wang> The kvm/spice stack is getting pretty wicked. I am just wondering
how good it is for 3d games. Im not a gamer. It also isn't -that- hard to
configure anymore. I think it is 4 windows drivers, and with F16, it is
merely a checkbox in the virtual manager gui/libvirtd. Much easier, and
better then i think it was F10 or F12 when I started to test it...

(Kena/185948) Ohhhhh... and here's more info on what happens -- apparently, "s...

Kena@rss.iscabbs.com — Thu, 02 Feb 2012 10:27:00 -0600

Ohhhhh... and here's more info on what happens -- apparently, "safely remove"
actually powers off the device, which is why re-insertion or a power cycle
(which seem kind of like sledge hammer vs. fly) is the solution:
http://ubuntuforums.org/showthread.php?t=1477247

(Kena/185947) Methinks that's not what he means; the device won't exist, as th...

Kena@rss.iscabbs.com — Thu, 02 Feb 2012 10:25:00 -0600

Methinks that's not what he means; the device won't exist, as the USB module
isn't presenting it.  Here's a sneaky way -- reset power on the USB bus:
http://unix.stackexchange.com/questions/7412/how-to-reconnect-a-logically-disco
nnected-usb-device (Wups: http://tinyurl.com/6b2okvs )

(Copper Lethe/185946) Do you know the device name?

Copper Lethe@rss.iscabbs.com — Thu, 02 Feb 2012 08:50:00 -0600


Do you know the device name?

If so, you should be able to do

hdparm -z device-name

(River Rat/185945) Probably not.

River Rat@rss.iscabbs.com — Thu, 02 Feb 2012 08:14:00 -0600

Probably not.

(John Public/185944) How do RE-detect a USB device after I've "ejected" it? The The d...

John Public@rss.iscabbs.com — Wed, 01 Feb 2012 17:46:00 -0600

How do RE-detect a USB device after I've "ejected" it? The The device is still
plugged in.  So far, the only way I can trigger this is to physically
disconnect the USB device, then re-connect it. Is there a command-line or a
GUI way to pretend the device was re-plugged in?

(MrPiolo/185942) Debian Testing already has kernel 3.1.

MrPiolo@rss.iscabbs.com — Sun, 29 Jan 2012 17:17:00 -0600


Debian Testing already has kernel 3.1.
Stable backports is in 3.2 for i386 and amd64.

(Kena/185941) When Debian goes to the 3.0 kernels.

Kena@rss.iscabbs.com — Sun, 29 Jan 2012 16:42:00 -0600

When Debian goes to the 3.0 kernels.

Sadly, that's probably not an elusive answer.

(River Rat/185940) When is Debian going to goto the 3.0 kernels?

River Rat@rss.iscabbs.com — Sun, 29 Jan 2012 15:22:00 -0600

When is Debian going to goto the 3.0 kernels?

(Wang Master/185939) Cyanide:

Wang Master@rss.iscabbs.com — Sat, 28 Jan 2012 15:17:00 -0600

Cyanide:
i havne't tried.  i gave up gaming on linux and now just boot to windows for
all my gaming

(River Rat/185938) January 27, 2012, 10:08 AM . The Fedora Project is currently mou...

River Rat@rss.iscabbs.com — Sat, 28 Jan 2012 13:32:00 -0600

January 27, 2012, 10:08 AM . The Fedora Project is currently mounting a
concerted effort to merge Linux filesystem directories into a more organized
structure, an effort known as /usr merge.

This is not really new news: I covered this back in November. But a new posting
on FreeDesktop.org from systemd developer Lennart Poettering is seeking to
dispel many of the arguments against such a merge, while touting the merger's
advantages.

Poettering is not the initial proposer of /usr merge--that designation goes to
fellow Red Hat developers Harald Hoyer and Kay Sievers, as an effort to clean
up the mess that was made when the /sbin and /bin directories were first split
off from each other. The Grand /usr-fication Theory will, when implemented,
essentially pull in every component of the operating system to a single mounted
volume.

When all of the binaries and libraries are on such a volume, it would be far
simpler to run multiple instances of the operating system on different machines
on a network, as well as facilitate the use of snapshots. With the onset of the
btr filesystem (btrfs) in the Linux world, such a unification would be a huge
advantage.

In his post this week, Poettering takes pains to note that this filesystem
change is neither dependent on systemd, or Fedora alone. systemd is a new init
daemon developed by Poettering designed to improve services management on Linux
systems--particularly mobile ones that have variable needs depending on where
or when the system is used. (Carla Schroder has an excellent series of articles
on Linux.com that highlights the features of systemd.)

"systemd supports both systems with split and with merged /usr, and the /usr
merge also makes sense for systemd-less systems," Poettering wrote on
FreeDesktop.org. "That said we want to encourage distributions adopting systemd
to also adopt the /usr merge."

Poettering highlights four broad issues why such a merge makes sense to him:

"Improved compatibility with other Unixes/Linuxes in behaviour: After the /usr
merge all binaries become available in both /bin and /usr/bin, resp. both /sbin
and /usr/sbin (simply because /bin becomes a symlink to /usr/bin, resp. /sbin
to /usr/sbin). That means scripts/programs written for other Unixes or other
Linuxes and ported to your distribution will no longer need fixing for the file
system paths of the binaries called, which is otherwise a major source of
frustration. /usr/bin and /bin (resp. /usr/sbin and /sbin) become entirely
equivalent.
"Improved compatibility with other Unixes (in particular Solaris) in
appearance: The primary commercial Unix implementation is nowadays Oracle
Solaris. Solaris has already completed the same /usr merge in Solaris 11. By
making the same change in Linux we minimize the difference towards the primary
Unix implementation, thus easing portability from Solaris.
"Improved compatibility with GNU build systems: The biggest part of Linux
software is built with GNU autoconf/automake (i.e., GNU autotools), which are
unaware of the Linux-specific /usr split. Maintaining the /usr split requires
non-trivial project-specific handling in the upstream build system, and in your
distribution's packages. With the /usr merge, this work becomes unnecessary and
porting packages to Linux becomes simpler.
"Improved compatibility with current upstream development: In order to minimize
the delta from your Linux distribution to upstream development the /usr merge
is key."
As I wrote in November, all of this seems like a pretty good idea, as this will
make cross-distro development a lot easier for those distros who adopt the new
filesystem architecture. Currently, different distros locate key libraries and
binaries in different locations in the Linux filesystem, for which application
developers have to compensate when they code and package their apps. An
all-in-one solution that would merge or symlink all of these locations together
would make things a whole lot easier.

In order to discount some of the objections to this proposal, Poettering listed
eleven "myths" that he then dispelled. Of these, the one that caught my eye was
this one:

"Myth #9: The /usr split is useful to have a minimal rescue system on the root
file system, and the rest of the OS on /usr."

To which Poettering responds:

"Fact: On Fedora the root directory contains ~450MB already. This hasn't been
minimal since a long time, and due to today's complex storage and networking
technologies it's unrealistic to ever reduce this again. In fact, since the
introduction of initrds to Linux the initrd took over the role as minimal
rescue system that requires only a working boot loader to be started, but not a
full file system."

Poettering seemed to take on the issue that I and others have raised about this
proposal, as well as the adoption of the new syslog system, Journal, by the
Fedora Project. There are concerns by many, including me, that Fedora's (and
subsequently Red Hat's) adoption of these new techniques and tools are an
attempt to move forward in the marketplace and set themselves apart from other
Linux distros as an "easier" platform for which to develop.

"Multiple other Linux distributions have been working in a similar direction,"
Poettering wrote.

That is not in dispute, but there is also no denying that the Fedora Project is
by far the most aggressive proponent of such moves. Initially, I was critical
of this effort, mostly because of my bias towards the Linux Standard Base and
concerns that Red Hat was yet again thumbing its nose at the rest of the Linux
community and going off and doing it's own thing.

But maybe there's something else afoot here.

A few times, Poettering mentioned this change would mean better compatibility
with Solaris 11. Normally, that would be a point of trivial interest. But after
running up against a lot of people from Joyent this past weekend at SCALE 10X,
some dots are getting connected.

Joyent is a cloud services provider that provides a complete software stack for
its customers. with its SmartOS operating-system layer based on Solaris. Joyent
is coming out will all guns blazing of late, with lots of presence at events
like SCALE 10X, Node Summit, and Monki Gras, not to mention some aggressive
positioning of Solaris' Dtrace and other features.

Again, taken alone, this is all well and good. you certainly can't blame Joyent
for being enthusiastic about their product.

But, given all of these little breadcrumbs, not to mention more than a little
groundswell about Oracle Solaris these days, now one wonders: are Fedora/Red
Hat trying to shift themselves (and the rest of Linux) into a better position
to combat Solaris?

It might not be a bad idea. If Solaris were to take off (and the cloud is a
great platform on which one could deploy bazillions of Solaris instances pretty
darn quickly), one area of competition would definitely be application
compatibility. If Linux can maintain better compatibility with Solaris (and any
other UNIX contender), then that point of contention would be come moot--or at
least weaker.

Suddenly, what seemed to be a power-play by Red Hat might ultimately be of
great benefit to the overall Linux ecosystem. Sure, Red Hat would reap the
initial benefits, but if they were the ones who saw this challenge coming, I
certainly wouldn't begrudge them that.

Speculation, to be sure, but broader unification with UNIX-based platforms has
to have some reason driving it, and this seems as likely as any.

Read more of Brian Proffitt's Open for Discussion blog and follow the latest IT
news at ITworld. Drop Brian a line or follow Brian on Twitter at
@TheTechScribe. For the latest IT news, analysis and how-tos, follow ITworld on
Twitter and Facebook

(John Public/185937) Enjoy yr vacation.

John Public@rss.iscabbs.com — Fri, 27 Jan 2012 17:38:00 -0600

Enjoy yr vacation.
I follow GCC very closely. I'm glad Fedora is pre-testing the next version.
There's nothing like a full distro to test GCC. Methinks if the QA is a
complete linus distro, then it doesn't get much more solid than that!

(Cyanide/185936) Okay going from way back.. was/am on vacation...

Cyanide@rss.iscabbs.com — Fri, 27 Jan 2012 15:47:00 -0600

Okay going from way back.. was/am on vacation...

JP> while the compiler shouldnt need to boast a 10x performance boost in
linking.. It was definately needed and a very noteworthy fix. Fedora 17 is
moving towards it, or at least they already began/did a mass compile with it to
help shake out some bugs in it..

Raub> I -think- the ldap.conf in /etc/openldap is for the ldapclient software..
and the /etd/ldap.conf is for the nss implemenation. Although they might be the
exact same thing except one is read first like the sasl conf files.

Wang> How does gaming in a kvm vm using the spice libraries compare to actual
hardware? *curious*

Commercial unix> It still has a place, especially for places who need support.
Like at the state, they had a db policy where the software had to be IN
production for like 5 years in a similar place, and the company had to exist
for more then 10 years before they would even consider the contract.

Upgrading Linux> I usually ahve a virtual machine which has all the of the dev
environment, and it is tested -before- upgrade of the actual physical
hardware.. Also, just installing minimal, and then adding the packages you need
keeps updates super small and makes it a lot easier to track the bugs.

For production I usually use RH/Centos. For development or major upgrades
which will take a while I typically use Fedora, or work with the project
itself, since the fixes will make it into "stable" by the time you need to
upgrade. You run into -other- things as well but keeping in line with
development makes it a lot easier to get help.

(River Rat/185935) Odd, no, that is display.

River Rat@rss.iscabbs.com — Fri, 27 Jan 2012 09:19:00 -0600

Odd, no, that is display.

Raubvogel: :)

Check this out guys.

http://www.linuxforum.com/content.php/163-Basic-Backup-Script-in-BASH

(Raubvogel/185934) River Rat, kinda nice that after 8months they decided to patch i...

Raubvogel@rss.iscabbs.com — Fri, 27 Jan 2012 08:40:00 -0600

        River Rat, kinda nice that after 8months they decided to patch it.

(Guardian Angel/185933) Considering that the cluster management software (if you're talk...

Guardian Angel@rss.iscabbs.com — Thu, 26 Jan 2012 04:11:00 -0600

Considering that the cluster management software (if you're talking pretty much
any HPC Linux cluster) is based, at least in part if not wholly, RHCS, it makes
sense. :)

If you really want Debian with slick cluster management, feel free to scratch
that itch ;)

that being said - in my experience running HPC clusters, Debian tends to end up
being so far out of date that the software demanded by the researchers isn't
able to run on the systems, or is a nightmare to get running versus biting the
fullet and running a RHEL-based cluster (a la ROCKS, or something completely
configured in house).

(Odd/185932) JXL> oh yeah, I wish I could have had debian on our cluster. P...

Odd@rss.iscabbs.com — Wed, 25 Jan 2012 22:38:00 -0600

JXL>  oh yeah, I wish I could have had debian on our cluster.  Problem is that
the cluster management software only supports redhat-based installations.  With
15 nodes and two front-ends that really starts to be worthwhile, so we just
have to accept it and deal with yum instead.

The concept is otherwise the same - we keep this thing as "stable" as we can,
but as it's used for development occasionally you'll have to take the plunge
and upgrade some software, otherwise you simply can't use or build recent stuff
anymore.

That put's us exactly in the position JP was talking about - is it easier to
upgrade often and keep the changes small, or better to wait a bit and do it
every other year or so?

(Odd/185931) River Rat> did you mean "more" instead of "display" in that awk...

Odd@rss.iscabbs.com — Wed, 25 Jan 2012 22:35:00 -0600

River Rat>  did you mean "more" instead of "display" in that awk statement?
display is a graphics-file viewer.

(River Rat/185930) Linux vendors rush to patch privilege escalation flaw after root...

River Rat@rss.iscabbs.com — Wed, 25 Jan 2012 12:10:00 -0600

Linux vendors rush to patch privilege escalation flaw after root exploits
emerge
http://www.itworld.com/244019/linux-vendors-rush-patch-privilege-escalation-fla
w-after-root-exploits-emerge

(Wang Master/185929) sweet. The bumblebee project for optimus support has come prett...

Wang Master@rss.iscabbs.com — Wed, 25 Jan 2012 05:28:00 -0600

sweet.  The bumblebee project for optimus support has come pretty far in a
fairly short period of time.  intel graphics has come along way too in terms
of acceptable 3D support.

i have a thinkpad w520 with the quadro 2000M gpu and it's decent for gaming
under windows, but under linux, in optimus mode the X server uses the intel
graphics by default, but since the nvidia gpu was never shut off, battery life
was less than 2 hours.

with bbswitch, it will now disable the nvidia gpu under linux, letting
me use the intel graphics with X and now i get 3-4+ hours easy on battery

and when i want to game, i reboot to linux and all is good.
no more fucking around with the bios settings to switch between discrete
and onboard everytime i want to reboot to game

very cool :)

(JuggernautXL/185928) Odd> Debian. It's made entirely from the perspective that you wi...

JuggernautXL@rss.iscabbs.com — Tue, 24 Jan 2012 17:05:00 -0600

Odd> Debian. It's made entirely from the perspective that you will install it
and leave it (the server installation) in its configuration until you upgrade.
You can literally download EVERY SINGLE PROGRAM in the repositories. Therefore,
given a certain configuration, you can deploy it easily and leave it at that.
Although upgrading (ala Arch or ala Ubuntu) is great, I'm guessing the vast
majority of users of servers are not going to desire this type of arrangement.
This makes Debian a perfect choice. Application/os upgrades is not the only way
to enable security. I'm no expert, but here's my partial list of things I would
do from the perspective of using Debian as a server and sandboxing/securing as
much as possible:

1) SE Linux
2) AppArmor(not sure if this would be used with the above)
3) firewall configuration that is specific to ONLY allow traffic to/from a
whitelisted set of applications and only with the specified configurations.
4) The most minimal installation possible to run services.
5) Multiple modes of backup of system and data.
6) rootkit detector
7) Antivirus scanner if handling files that get passed on to Windows machines.

AppArmor and SE linux are great. They create these sandboxes that apps work in.
While they can't prevent vulnerabilities, in alot of cases they can prevent
these from wreaking much havok. I can see myself setting up a server to be
self-healing and self-sustaining in the sense that data and system is backed
up. Data and system is monitored regularly to ensure compliance with some
hashed value. All original system images/files/data is backed up to external
media. All applications are sandboxed to the fullest extent possible. If
inconsistencies are found, data/system files is restored. If the mandated set
of applications fails/crashes, it (they) are restarted.

It is entirely possible to run a completely secure server for a decade on a
single installation. Why someone would want to torture themselves with
installing a server every 6 months is ridiculous, especially for the type of
complex list of applications enterprise users might have... Set it and forget
it...

That said, I always think any of the above are always overkill.. at least for
casual users. Let me know if anybody has comments about the above or
suggestions (better/worse/additional).

(John Public/185927) Danix> It prob depends on how much your company pays Microsoft. ...

John Public@rss.iscabbs.com — Mon, 23 Jan 2012 18:11:00 -0600

Danix> It prob depends on how much your company pays Microsoft. MS is always
debugging some issues with SQL Server clusters for us. No, if there's one
thing you can say about Microsoft, it's that their support teams are always at
the ready, and never bored! ;)

(Mild Seven/185926) Yeah, I've activated it and added prepaid credit to it, then boo...

Mild Seven@rss.iscabbs.com — Mon, 23 Jan 2012 14:02:00 -0600

Yeah, I've activated it and added prepaid credit to it, then booted to Lubuntu
and ran Network Manager.  It might only be placebo, but I swear my downloads
and web browsing is faster under *nix than in Windows. :-)

(Kena/185925) Once you've activated your CDMA device, I believe you can use it...

Kena@rss.iscabbs.com — Mon, 23 Jan 2012 13:51:00 -0600

Once you've activated your CDMA device, I believe you can use it with the stock
Ubntu network manager.  It does, however, have to be activiated with the actual
Verizon Access Manager -- which just seems stupid.

(Mild Seven/185924) My history with *nix:

Mild Seven@rss.iscabbs.com — Mon, 23 Jan 2012 13:35:00 -0600

My history with *nix:

1991: Unix shell account at university.
1992: University deploys Linux. Computer science professor very happy about
free OS and mentions the name "Torvalds".

"Professor, how do I do this?"  Answer: "see the man pages"
lol

1999: Nuked Win9x, installed Caldera OpenLinux after 17 hours of RTFM.  It
worked. :-) Never crashed!  I felt I was on to something! :-)

2012: Installed Lubuntu (LXDE desktop). My thoughts, "This... just...works!"

I'm very pleased with this distro.  Extremely easy to use for a noob. My old
computer is back to life again.  The ubuntu forums are very helpful.  Of
course,
I don't understand all the commands, so I mostly copy/paste them.

Scanner, webcam, etc all work just fine with Lubuntu. This time I might be able
to dump Windows for good, although I need Windows to add more prepaid credits
for my CDMA Rev. A internet connection (Verizon Access Manager doesn't work
with WINE).

(Kena/185923) Not sure what that's doing -- I don't really speak awk -- but he...

Kena@rss.iscabbs.com — Mon, 23 Jan 2012 13:00:00 -0600

Not sure what that's doing -- I don't really speak awk -- but here's the
ImageMagick link for that error:
http://rmagick.rubyforge.org/install-faq.html#delegate

(River Rat/185922) [oppie@oppie ~]$ awk 'BEGIN{FS=":"; print "digraph{"}{split($4, ...

River Rat@rss.iscabbs.com — Mon, 23 Jan 2012 11:15:00 -0600

[oppie@oppie ~]$ awk 'BEGIN{FS=":"; print "digraph{"}{split($4, a, ","); for (i
in a) printf "\"%s\" [shape=box]\n\"%s\" -> \"%s\"\n", $1, a[i], $1}END{print
"}"}' /etc/group | display
display: no decode delegate for this image format `/tmp/magick-5VEP4dPz' @
error/constitute.c/ReadImage/532.
[oppie@oppie ~]$

Am I missing something?

(Keogk/185921) We had a MS support specalist just spend 3 days at out office fi...

Keogk@rss.iscabbs.com — Mon, 23 Jan 2012 11:01:00 -0600

We had a MS support specalist just spend 3 days at out office fixing a
sharepoint issue

(Kena/185920) You ever find out the answer to that, and PLEASE let me know.

Kena@rss.iscabbs.com — Mon, 23 Jan 2012 09:24:00 -0600

You ever find out the answer to that, and PLEASE let me know.

(Raubvogel/185919) Centos question: what is the difference between

Raubvogel@rss.iscabbs.com — Mon, 23 Jan 2012 08:42:00 -0600

        Centos question: what is the difference between
/etc/openldap/ldap.conf and /etc/ldap.conf?

(River Rat/185918) 2- Duplicate installed packages from one machine to the other (R...

River Rat@rss.iscabbs.com — Mon, 23 Jan 2012 08:38:00 -0600

2- Duplicate installed packages from one machine to the other (RPM-based
systems)

ssh root@remote.host "rpm -qa" | xargs yum -y install

(Guardian Angel/185915) It's generally why UNIX and UNIX-like OS still rule the F500 DCs...

Guardian Angel@rss.iscabbs.com — Mon, 23 Jan 2012 05:32:00 -0600

It's generally why UNIX and UNIX-like OS still rule the F500 DCs. They would
prefer to pay someone to fix shit the first time than bandaid it for years.

(Danix/185914) Commercial linux> A quite good bunch of large organizations have...

Danix@rss.iscabbs.com — Mon, 23 Jan 2012 05:25:00 -0600

Commercial linux> A quite good bunch of large organizations have some kind of
"Software MUST INCLUDE support" directive in their IT policies. RHEL and the
"Commercial Linux" crowd currently fill those requirements, so it is more
likely that these orgs will fire up RHEL for their servers than say, Slackware
or Debian.

Commercial Linux vendors had one thing pretty much nailed: the "support" part
of that requirement. Sun would send us a local partner, but if stuff didn't get
fixed, they'd send local Sun people to fix it... and if those couldn't do the
job, they would send someone from Sun USA to fix it.

OTOH, Microsoft's "support" consists on routing you to a call center tended by
some generic tech from India who might or might not fix your problem. And
there's no way they'll ever send an engineer to your site. Nope, do everything
by phone.

So I'm pretty surprised that MS has *any* kind of market % in large
organizations. Or maybe that's why UNIX still rules the Fortune 500
datacenters?

(Odd/185913) JP> Yeah, good points. I agree with all of them, but it doesn'...

Odd@rss.iscabbs.com — Sun, 22 Jan 2012 20:52:00 -0600

JP>  Yeah, good points.  I agree with all of them, but it doesn't remove the -
say - mild trepidation of typing "yum update".  One primary problem being that
we do have a double cluster frontend (so upgrading one is relatively safe), but
it's hard to duplicate the environment completely for software testing.

We'll have to prepare a little and find out at least in advance which compiler
versions we'll be getting, to find out if everything still builds.

Oh, and the other problem is it's not my job, but there's no-one else to do it.
Yum.

(John Public/185912) mainframes: yup.

John Public@rss.iscabbs.com — Sun, 22 Jan 2012 19:06:00 -0600

mainframes: yup.

(River Rat/185911) Is there businesses still using mainframes?

River Rat@rss.iscabbs.com — Sun, 22 Jan 2012 18:26:00 -0600

Is there businesses still using mainframes?
Or is it now down to pc's running linux or Sun Computers running solaris?

(Kena/185910) And this is why filesystems like zfs and btrfs are cool: take ...

Kena@rss.iscabbs.com — Sun, 22 Jan 2012 13:32:00 -0600

And *this* is why filesystems like zfs and btrfs are cool: take a snapshot, do
your upgrade.  It fails?  Get as much pertinent triage/failure info as you can,
and revert.  The only place where this can truly barf is when altering the MBR
-- that's your single point of failure.  (Though, at least on X86, it's easy
enough to make a backup of that, too, and restore/revert from bootable media.)

(John Public/185909) I'm with you on frequency of upgrades. The problem is that folks...

John Public@rss.iscabbs.com — Sun, 22 Jan 2012 10:35:00 -0600

I'm with you on frequency of upgrades. The problem is that folks do not
understand the risks accurately and unemotionally. The issue should never be
on the table WHETHER to upgrade; the issue should be WHEN and HOW. I tend to
see fear culture driving upgrade decisions. Fear prevents upgrades, and
inflates their severity:

Fear:    Upgrading introduces risk of incompatibilities,
         therefore, freezing minimizes risk. Your job is to just
         keep the power supply humming and roadblock all change.
Reality: Both upgrading and freezing move you closer to incompatibilities.
         However, freezing guarantees no improvement and every day
         closer to obsolescence. Therefore upgrading minimizes risk.
         Your job is to make smart decisions and to execute a sensible upgrade
         strategy.

Fear:    The longer you've gone without taking a specific upgrade, the
         more serious and dangerous the upgrade will be. You waited too long,
         and now the risks have compounded.
Reality: The upgrade risk is unaffected by the span of time. It's the same as
         it was the day the upgrade was released. (Possibly even less risky,
         see next fear.)

Fear:    Upgrades that jumps multiple versions are more serious and dangerous,
         i.e., now that you've waited far too long, the situation has become
         dire. You must escalate the severity of the upgrade and assemble
         a massive team to stay all weekend.
Reality: Upgrades generally improve the state of things. Chances are that
         many major issues have been resolved and performance will be
         increased. You don't need an all-out 5-alarm fire drill to pull this
         off. Just perform the upgrade as if it were any other.

Fear:    If an upgrade fails you may not be able to bring up the production
         system. By 6AM the phone will be ringing off the hook and by 10AM
         you'll walking to your car with a box of obsolete O'Reilly books
         and a Dilbert calendar from 2005. Play it safe.
Reality: You have backups, you know how to restore them, you have a
         plan already in place, and everyone knows what it is. This is what
         you are paid to accomplish.

(Odd/185908) I never thought I'd agree, but since using a production cluster ...

Odd@rss.iscabbs.com — Sun, 22 Jan 2012 09:23:00 -0600

I never thought I'd agree, but since using a production cluster I'm starting to
see why you want a) a rock-stable linux distribution and b) a commercially paid
party to give you software support.  If only to help with the upgrades within a
3-hour window.  It's getting to the point where I don't dare to do any
upgrades, which is a fairly bad thing.