Re: [CentOS] broken GFS

2008-05-14 Thread Johnny Hughes

Doug Tucker wrote:

Tru,

I work at a university.  They don't provide enough money for test
environments :).  Just kinda odd, last time kernel update, gfs updated
at the same time so all was well.  But twice now kernel has upgraded
with no GFS so it went bye-bye.  Is the GFS being installed, compiled
against particular kernel headers, or could I just copy the /fs/gfs
and /fs/gfs_locking to the new kernel /lib/modules (or symlink for that
matter) and be lucky enough it would work?




Please be aware that redhat releases GFS at a different time than the, 
usually 2-3 days later (at the earliest).


In this case, here are the upstream release dates:

kernel - 5/7/2008
gfs kmods - 5/9/2008

my point is that even upstream does not release these  at the same time.

What you should do (and what everyone who has kmods on c4 should do) is 
to exclude kernels from automatic updates ... then you can manually 
update the kernels and kmods together separately.


Thanks,
Johnny Hughes



signature.asc
Description: OpenPGP digital signature
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread js

Hello All;

Maybe, because XFS seems to be important, is it  possible to build xfs 
right after the kernel src build?


Is this far more longer than only build the kernel?

Ok nobody pay you to do Centos, ok.
Centos is a very good project, but i think it's not really constructive 
to say ok, pay me and I will do it :)

You don't do Centos because you need money but because you like what you do.


Of course, forget my mail if XFS is a crap to build, but if a simple  
add stuff in changelog xfs.spec;  rpmbuild -ba --sign 
mycoolXFSmodul.src.rpm is enough,
maybe You could think to build xfs in the same time a kernel update is 
available ?


Regards

js.






Johnny Hughes a écrit :

Linux wrote:

On Tue, May 13, 2008 at 1:15 AM, Tru Huynh [EMAIL PROTECTED] wrote:

On Tue, May 13, 2008 at 12:40:22AM +0300, Linux wrote:
  What a coincidence. That is the 1st time I live such a thing. Well,
  show me a way to prove.
 /var/log/messages ?

Only a small part of it.


  This log is after update  reboot:
  May 11 16:06:03 x kernel: XFS: failed to read root inode
 nothing more?

Well, that is the only unexpected part. Just to show that XFS module
was loaded for WRONG kernel. As you said, you newer saw before.


  According to this, there is a mystery in May 11 16:06:03 because
  there WAS a kmod_xfs but it was 53.1.14, not 53.1.19 as updated
  kernel.
 too bad you rebooted 1 hour before the kernel-xfs module update.

When was kernel-xfs module updated in repository? Just that time? If
so too bad CentOS folks do not update every piece of kernel as a whole
in repositories. Where is integrity?

If not, yum update does not update everything at once. I have to run
yum update twice maybe more. First it will load kernel then see that a
new kernel is available, will go and bring its modules...

Still, it is a bit annoying and confusing. I am beginning to think
whether XFS is really supported in CentOS :)


OK ... let me give you an official answer

red hat does not even release the the gfs kmods on the same day as the 
kernel, that is FULLY supported and even an added expense for rhel4.


we DO NOT update xfs (or the centosplus kernel) on the same day as the 
base centos kernels.  We are NOT going to wait to release the main 
kernel security update for a day or more to get centosplus stuff also 
done.


xfs IS NOT SUPPORTED in the same way as the base centos distro is and 
xfs is not in RHEL.


Our 2 million users do not want to wait for the base kernel security 
updates for 2 extra days so that a very small group of people who use 
the xfs file system can get their updates at the same time.


It might take even longer to get these built as no one pays me to 
build them and I have a real job and a real life  ... if you can't do 
one of these:


1.  Build your own module.
2.  Exclude the kernel and only update it when the modules are ready.

Then you can pay me $200.00 per hour and I will manage your server for 
you.


Thanks,
Johnny Hughes



___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos
  


___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Martyn Drake
On Wed, May 14, 2008 at 2:10 PM, js [EMAIL PROTECTED] wrote:

 Maybe, because XFS seems to be important, is it  possible to build xfs right
 after the kernel src build?

 Is this far more longer than only build the kernel?

Assuming that you've set it up as a module rather than actually
compiling it into the kernel itself, it should be a case of just
doing:

cd $KERNEL_SOURCE_TREE
make fs/xfs/xfs.ko
mkdir /lib/modules/$KERNEL_VERSION/kernel/fs
cp fs/xfs/xfs.ko /lib/modules/$KERNEL_VERSION/kernel/fs/xfs

which will build the XFS module and stick it in the right place.  Note
that all that does not include the XFS userspace tools...

Regards,

Martyn
-- 
Martyn Drake
http://www.drake.org.uk
http://www.mindthegapps.com
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Akemi Yagi
On Wed, May 14, 2008 at 6:39 AM, Martyn Drake [EMAIL PROTECTED] wrote:
 On Wed, May 14, 2008 at 2:10 PM, js [EMAIL PROTECTED] wrote:

   Maybe, because XFS seems to be important, is it  possible to build xfs 
 right
   after the kernel src build?
  
   Is this far more longer than only build the kernel?

  Assuming that you've set it up as a module rather than actually
  compiling it into the kernel itself, it should be a case of just
  doing:

  cd $KERNEL_SOURCE_TREE
  make fs/xfs/xfs.ko
  mkdir /lib/modules/$KERNEL_VERSION/kernel/fs
  cp fs/xfs/xfs.ko /lib/modules/$KERNEL_VERSION/kernel/fs/xfs

  which will build the XFS module and stick it in the right place.  Note
  that all that does not include the XFS userspace tools...

Making kernel modules is a bit more involved than that.  Please see:

http://wiki.centos.org/HowTos/BuildingKernelModules

if you really feel like building modules yourself.

Akemi
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Martyn Drake
On Wed, May 14, 2008 at 3:28 PM, Akemi Yagi [EMAIL PROTECTED] wrote:

 Making kernel modules is a bit more involved than that.  Please see:

 http://wiki.centos.org/HowTos/BuildingKernelModules

 if you really feel like building modules yourself.

You're quite right.  You can tell I do it often, can't you? :)

Regards,

Martyn
-- 
Martyn Drake
http://www.drake.org.uk
http://www.mindthegapps.com
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Michael Semcheski
On Wed, May 14, 2008 at 9:10 AM, js [EMAIL PROTECTED] wrote:

 Maybe, because XFS seems to be important, is it  possible to build xfs
 right after the kernel src build?

 Is this far more longer than only build the kernel?

 Ok nobody pay you to do Centos, ok.
 Centos is a very good project, but i think it's not really constructive to
 say ok, pay me and I will do it :)
 You don't do Centos because you need money but because you like what you
 do.



As a matter of not annoying volunteers and developers, you want to be
careful about asking people to do what you can't or won't do yourself.
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Doug Tucker
I intend to do that.  Kernel's removed from automatic updates.

We'll agree to disagree about the importance of not breaking an
officially supported kernel filesystem on an automated upgrade because
only a few of us are affected.  Keep in mind this is not an
unsupported XFS that someone hijacked my thread with.  I say there is
little in a new kernel that the rest of the users cannot wait 2-3
lousy days for.  Wanna stretch it to a week to meet your statement of
earliest, I can live with that and my statement still stands.  And, I
do realize this is not centos's fight, I guess my complaint is with
RedHat in this case, they should be more responsible than that.  If M$
took that policy and released official upgrades they knew would break
even a small percentage of their users, especially something as critical
as the very filesystem that your entire user data resides on, we (the
linux community) would be throwing them under the rug for it.


On Wed, 2008-05-14 at 05:44 -0500, Johnny Hughes wrote:
 Doug Tucker wrote:
  Tru,
  
  I work at a university.  They don't provide enough money for test
  environments :).  Just kinda odd, last time kernel update, gfs updated
  at the same time so all was well.  But twice now kernel has upgraded
  with no GFS so it went bye-bye.  Is the GFS being installed, compiled
  against particular kernel headers, or could I just copy the /fs/gfs
  and /fs/gfs_locking to the new kernel /lib/modules (or symlink for that
  matter) and be lucky enough it would work?
  
  
 
 Please be aware that redhat releases GFS at a different time than the, 
 usually 2-3 days later (at the earliest).
 
 In this case, here are the upstream release dates:
 
 kernel - 5/7/2008
 gfs kmods - 5/9/2008
 
 my point is that even upstream does not release these  at the same time.
 
 What you should do (and what everyone who has kmods on c4 should do) is 
 to exclude kernels from automatic updates ... then you can manually 
 update the kernels and kmods together separately.
 
 Thanks,
 Johnny Hughes
 
 ___
 CentOS mailing list
 CentOS@centos.org
 http://lists.centos.org/mailman/listinfo/centos

___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread MHR
On Wed, May 14, 2008 at 7:44 AM, Doug Tucker [EMAIL PROTECTED] wrote:
 I intend to do that.  Kernel's removed from automatic updates.

There you go.

 We'll agree to disagree about the importance of not breaking an
 officially supported kernel filesystem on an automated upgrade because
 only a few of us are affected.  Keep in mind this is not an
 unsupported XFS that someone hijacked my thread with.  I say there is
 little in a new kernel that the rest of the users cannot wait 2-3
 lousy days for.  Wanna stretch it to a week to meet your statement of
 earliest, I can live with that and my statement still stands.  And, I
 do realize this is not centos's fight, I guess my complaint is with
 RedHat in this case, they should be more responsible than that.  If M$
 took that policy and released official upgrades they knew would break
 even a small percentage of their users, especially something as critical
 as the very filesystem that your entire user data resides on, we (the
 linux community) would be throwing them under the rug for it.


1) You're top posting - please stop it.  In this email list, we bottom
post as a matter of policy and courtesy.  It's not that hard

2) This isn't really an issue of agreeing to disagree.  XFS is *not*
a Red Hat product at all.  They (RH) do not support it at all.  The
CentOS project provides XFS as an *extra* that is NOT part of the
mainline CentOS release stream.  It is only supported by the CentOS
group in the centosplus repository, which is a courtesy provided for
free by the CentOS group.

IOW, CentOS does not have to support XFS at all.  That they do is a courtesy.

Now, if you like the centosplus product and use it, remember to
follow the guidelines for it - little things like not doing automatic
updates because you already *know* that centosplus does not come out
immediately when RH releases a change that CentOS picks up and
releases as well.

All of this is clearly discussed here from time to time, so the
expectations have been set accordingly.  Please try to remember this
and manage your installations accordingly, too.

And that's *my* soapbox, from which I will now step down and shut up.
Temporarily.

:-}

mhr
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Doug Tucker
On Wed, 2008-05-14 at 11:07 -0700, MHR wrote:
 On Wed, May 14, 2008 at 7:44 AM, Doug Tucker [EMAIL PROTECTED] wrote:
  I intend to do that.  Kernel's removed from automatic updates.
 
 There you go.
 
  We'll agree to disagree about the importance of not breaking an
  officially supported kernel filesystem on an automated upgrade because
  only a few of us are affected.  Keep in mind this is not an
  unsupported XFS that someone hijacked my thread with.  I say there is
  little in a new kernel that the rest of the users cannot wait 2-3
  lousy days for.  Wanna stretch it to a week to meet your statement of
  earliest, I can live with that and my statement still stands.  And, I
  do realize this is not centos's fight, I guess my complaint is with
  RedHat in this case, they should be more responsible than that.  If M$
  took that policy and released official upgrades they knew would break
  even a small percentage of their users, especially something as critical
  as the very filesystem that your entire user data resides on, we (the
  linux community) would be throwing them under the rug for it.
 
 
 1) You're top posting - please stop it.  In this email list, we bottom
 post as a matter of policy and courtesy.  It's not that hard

I'm sorry, that last sentence was unnecessary and just rude.  I don't
tell you how to set your email client and what your preference is toward
how you like to read your email.  I find it completely annoying to have
to scroll to the bottom of a message to read a reply.  I will comply
with the group as a whole that I chose to join, I was unaware that
bottom posting was preference.  But I do not appreciate the tone, you
could have easily asked nicely or referred me to the preference policy
for me to follow.

 2) This isn't really an issue of agreeing to disagree.  XFS is *not*
 a Red Hat product at all.  They (RH) do not support it at all.  The
 CentOS project provides XFS as an *extra* that is NOT part of the
 mainline CentOS release stream.  It is only supported by the CentOS
 group in the centosplus repository, which is a courtesy provided for
 free by the CentOS group.

This is a matter of agreeing to disagree on the release of a kernel and
a supported file system.  If you had read my thread and subsequent
paragraph you're taking issue with properly, you would have gotten that.
My whole issue is around GFS, which is officially supported (someone
else hijacked this thread with XFS which got more attention), and in my
statement I said: Keep in mind this is not an unsupported XFS that
someone hijacked my thread with.  So I'm agreeing that XFS should never
be brought up in the same fashion as GFS, as it is not a supported file
system.  GFS is, and it is my opinion RH should release the 2 together.

 
 IOW, CentOS does not have to support XFS at all.  That they do is a courtesy.
 
 Now, if you like the centosplus product and use it, remember to
 follow the guidelines for it - little things like not doing automatic
 updates because you already *know* that centosplus does not come out
 immediately when RH releases a change that CentOS picks up and
 releases as well.

I already agreed and removed kernel from the update, no need to lecture.
Again, if you will take the time to read instead of knee-jerking a
reaction in some automatic defense of your feelings, you will note that
I took the aim at RedHat for the issue, and said it was not CentOS's
problem.  Read boy, read.

 All of this is clearly discussed here from time to time, so the
 expectations have been set accordingly.  Please try to remember this
 and manage your installations accordingly, too.
 
 And that's *my* soapbox, from which I will now step down and shut up.
 Temporarily.

And unfortunately, all based on improper understanding of what was
written, which makes it inappropriate in a public forum.  Me thinks you
had seen enough of the other guy whining about his unsupported platform,
saw the word XFS in my paragraph, and basically quit reading and decided
to send your XFS rant at me.  I hope from a therapeutic standpoint, it
helped you in some fashion.

:D


 :-}
 
 mhr
 ___
 CentOS mailing list
 CentOS@centos.org
 http://lists.centos.org/mailman/listinfo/centos

___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread John R Pierce

Doug Tucker wrote:

My whole issue is around GFS, which is officially supported (someone
else hijacked this thread with XFS which got more attention), and in my
statement I said: Keep in mind this is not an unsupported XFS that
someone hijacked my thread with.  So I'm agreeing that XFS should never
be brought up in the same fashion as GFS, as it is not a supported file
system.  GFS is, and it is my opinion RH should release the 2 together.
  



GFS is only 'officially supported' under a seperate  contract from 
Red Hat.  And, if you're a GFS customer of Red Hat's, I'm pretty darn 
sure the first thing they do is disable kernel updates... In fact, I 
seem to recall that RHEL4 ships with kernel updates disabled, you have 
to use `up2date --force` or something to enable them.





___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread MHR
On Wed, May 14, 2008 at 11:58 AM, Doug Tucker [EMAIL PROTECTED] wrote:
 On Wed, 2008-05-14 at 11:07 -0700, MHR wrote:
 1) You're top posting - please stop it.  In this email list, we bottom
 post as a matter of policy and courtesy.  It's not that hard

 I'm sorry, that last sentence was unnecessary and just rude.  I don't
 tell you how to set your email client and what your preference is toward
 how you like to read your email.  I find it completely annoying to have
 to scroll to the bottom of a message to read a reply.  I will comply
 with the group as a whole that I chose to join, I was unaware that
 bottom posting was preference.  But I do not appreciate the tone, you
 could have easily asked nicely or referred me to the preference policy
 for me to follow.


You apparently didn't see the smiley I left out of the last sentence  :-)
I didn't mean it to be rude at all - no tone implied.  I just noticed
that you have posted several times to the list and all of them, until
now, were top posts, unlike almost everyone else.  I /was/ trying to
be nice

 This is a matter of agreeing to disagree on the release of a kernel and
 a supported file system.  If you had read my thread and subsequent
 paragraph you're taking issue with properly, you would have gotten that.
 My whole issue is around GFS, which is officially supported (someone
 else hijacked this thread with XFS which got more attention), and in my
 statement I said: Keep in mind this is not an unsupported XFS that
 someone hijacked my thread with.  So I'm agreeing that XFS should never
 be brought up in the same fashion as GFS, as it is not a supported file
 system.  GFS is, and it is my opinion RH should release the 2 together.


Yes, I've been reading the thread.  I you didn't mention GFS in the
specific post to which I was replying, but you're right, it's there in
prior posts.  So all of my commentary about XFS does not apply to your
post.  Non-sequitur - mea culpa.  :-)

 I already agreed and removed kernel from the update, no need to lecture.

It was intended to be a gentle reminder.  (You've obviously never seen
me lecture)

 Again, if you will take the time to read instead of knee-jerking a
 reaction in some automatic defense of your feelings, you will note that
 I took the aim at RedHat for the issue, and said it was not CentOS's
 problem.  Read boy, read.

snip

 And unfortunately, all based on improper understanding of what was
 written, which makes it inappropriate in a public forum.  Me thinks you
 had seen enough of the other guy whining about his unsupported platform,
 saw the word XFS in my paragraph, and basically quit reading and decided
 to send your XFS rant at me.  I hope from a therapeutic standpoint, it
 helped you in some fashion.


You seem awfully touchy here - are you sure you're not lecturing me?  :-)

Take a breath, relax, you were not under attack, lecture or anything
rude.  I meant it with the best of intentions - I usually do.

mhr
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Doug Tucker
On Wed, 2008-05-14 at 12:38 -0700, MHR wrote:
 On Wed, May 14, 2008 at 11:58 AM, Doug Tucker [EMAIL PROTECTED] wrote:
  On Wed, 2008-05-14 at 11:07 -0700, MHR wrote:
  1) You're top posting - please stop it.  In this email list, we bottom
  post as a matter of policy and courtesy.  It's not that hard
 
  I'm sorry, that last sentence was unnecessary and just rude.  I don't
  tell you how to set your email client and what your preference is toward
  how you like to read your email.  I find it completely annoying to have
  to scroll to the bottom of a message to read a reply.  I will comply
  with the group as a whole that I chose to join, I was unaware that
  bottom posting was preference.  But I do not appreciate the tone, you
  could have easily asked nicely or referred me to the preference policy
  for me to follow.
 
 
 You apparently didn't see the smiley I left out of the last sentence  :-)
 I didn't mean it to be rude at all - no tone implied.  I just noticed
 that you have posted several times to the list and all of them, until
 now, were top posts, unlike almost everyone else.  I /was/ trying to
 be nice

It's not that hard would have gotten you b**ch slapped even with a
smile on your face in person.  Just stick to polite, it's not that
hard :D.

 
  This is a matter of agreeing to disagree on the release of a kernel and
  a supported file system.  If you had read my thread and subsequent
  paragraph you're taking issue with properly, you would have gotten that.
  My whole issue is around GFS, which is officially supported (someone
  else hijacked this thread with XFS which got more attention), and in my
  statement I said: Keep in mind this is not an unsupported XFS that
  someone hijacked my thread with.  So I'm agreeing that XFS should never
  be brought up in the same fashion as GFS, as it is not a supported file
  system.  GFS is, and it is my opinion RH should release the 2 together.
 
 
 Yes, I've been reading the thread.  I you didn't mention GFS in the
 specific post to which I was replying, but you're right, it's there in
 prior posts.  So all of my commentary about XFS does not apply to your
 post.  Non-sequitur - mea culpa.  :-)
 
  I already agreed and removed kernel from the update, no need to lecture.
 
 It was intended to be a gentle reminder.  (You've obviously never seen
 me lecture)

touche!

 
  Again, if you will take the time to read instead of knee-jerking a
  reaction in some automatic defense of your feelings, you will note that
  I took the aim at RedHat for the issue, and said it was not CentOS's
  problem.  Read boy, read.
 
 snip
 
  And unfortunately, all based on improper understanding of what was
  written, which makes it inappropriate in a public forum.  Me thinks you
  had seen enough of the other guy whining about his unsupported platform,
  saw the word XFS in my paragraph, and basically quit reading and decided
  to send your XFS rant at me.  I hope from a therapeutic standpoint, it
  helped you in some fashion.
 
 
 You seem awfully touchy here - are you sure you're not lecturing me?  :-)
 
 Take a breath, relax, you were not under attack, lecture or anything
 rude.  I meant it with the best of intentions - I usually do.

Bad thing about email, it's hard to grasp tongue in cheek humor and tone
isn't it?  Didn't you see my bfg at the end of my response?

Do you honestly, like having to scroll down with the rolly thing on your
mouse 9 times to get to the reply only to find it is not something you
cared to read?  I say toss it at the top in my face where I can ignore
it with less effort.

:D
BFG!


 mhr
 ___
 CentOS mailing list
 CentOS@centos.org
 http://lists.centos.org/mailman/listinfo/centos

___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread John R Pierce

Doug Tucker wrote:

Do you honestly, like having to scroll down with the rolly thing on your
mouse 9 times to get to the reply only to find it is not something you
cared to read?  I say toss it at the top in my face where I can ignore
it with less effort.
  


the other key part of bottom posting is to delete all but what you're 
replying to.   noone needs to see the whole thread quoted in every 
message, just enough context to frame the response.  And, delete the 
.SIG stuff on the end, too.



___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Doug Tucker
On Wed, 2008-05-14 at 12:37 -0700, John R Pierce wrote:
 Doug Tucker wrote:
  My whole issue is around GFS, which is officially supported (someone
  else hijacked this thread with XFS which got more attention), and in my
  statement I said: Keep in mind this is not an unsupported XFS that
  someone hijacked my thread with.  So I'm agreeing that XFS should never
  be brought up in the same fashion as GFS, as it is not a supported file
  system.  GFS is, and it is my opinion RH should release the 2 together.

 
 
 GFS is only 'officially supported' under a seperate  contract from 
 Red Hat.  

And?  It's official.  In fact, ext3 is only officially supported from
them these day without a $$$ contract.  Which is why we're all here!  :D


 And, if you're a GFS customer of Red Hat's, I'm pretty darn 
 sure the first thing they do is disable kernel updates... In fact, I 
 seem to recall that RHEL4 ships with kernel updates disabled, you have 
 to use `up2date --force` or something to enable them.

Yes, but kernel is disabled from EL4 reguardless of filesystem, so GFS
has nothing to do with that.  YOu can just edit the up2date file to
remove that.  I merely believe that GFS filesystem updates should be
released in conjuntion with kernel with all the other filesystems built
in, treating it no differently since it is officially supported, just
not put in the standard kernel build to put separation between it and
the $$ extra product.  And that is merely, an opinion.


 
 
 
 ___
 CentOS mailing list
 CentOS@centos.org
 http://lists.centos.org/mailman/listinfo/centos

___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Doug Tucker
On Wed, 2008-05-14 at 13:00 -0700, John R Pierce wrote:
 Doug Tucker wrote:
  Do you honestly, like having to scroll down with the rolly thing on your
  mouse 9 times to get to the reply only to find it is not something you
  cared to read?  I say toss it at the top in my face where I can ignore
  it with less effort.

 
 the other key part of bottom posting is to delete all but what you're 
 replying to.   noone needs to see the whole thread quoted in every 
 message, just enough context to frame the response.  And, delete the 
 .SIG stuff on the end, too.
 

I'm still annoyed.  Forgot to mention I hate having to move my cursor in
a different location than where it is when I hit the reply button before
I can type too.  Honestly, I see zero benefit in this.  And looking at
my other tech threads (isc.org and opennms.org) and everyone appears to
be top posting, although I guess, they could all be breaking the rules.

Humor turned off for a minute, completely and honestly, can someone
explain to me *why* this is the etiquette here?  In every fashion, I
find it sooo much harder to follow.  Does it date back to some dead text
based mail client that actually made this easier for some reason?

Left first paragraph at the top, because I find it too relevant in this
one to remove.



___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Akemi Yagi
On Wed, May 14, 2008 at 1:15 PM, Doug Tucker [EMAIL PROTECTED] wrote:

 Humor turned off for a minute, completely and honestly, can someone
 explain to me *why* this is the etiquette here?  In every fashion, I
 find it sooo much harder to follow.  Does it date back to some dead text
 based mail client that actually made this easier for some reason?

This is linked from the CentOS FAQ:

http://www.caliburn.nl/topposting.html

Akemi
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread MHR
On Wed, May 14, 2008 at 12:54 PM, Doug Tucker [EMAIL PROTECTED] wrote:
 It's not that hard would have gotten you b**ch slapped even with a
 smile on your face in person.  Just stick to polite, it's not that
 hard :D.

snicker

 Bad thing about email, it's hard to grasp tongue in cheek humor and tone
 isn't it?  Didn't you see my bfg at the end of my response?

Actually, I wasn't sure what that was

I just googled BFG and got BF Goodrich, BFG Tech, Big F**king Gun
and Big Friendly Giant, but I'm guessing you meant Big Fat Grin (or
some other F* word :-).

Seriously, though, I try to read email as if it had no tone (unless
the language or emoticons make it abundantly clear) and always, always
take a deep breath before I respond - shot myself in the foot enough
times to remember a few of them when I want to blast off.  I also try
to proof the responses, and frequently delete them so I can wait a
while before I write anything.

But, technically, since we're writing and not speaking, it's all
tongue-in-cheek, isn't it?

;^)

 Do you honestly, like having to scroll down with the rolly thing on your
 mouse 9 times to get to the reply only to find it is not something you
 cared to read?  I say toss it at the top in my face where I can ignore
 it with less effort.


I'll let others handle the arguments here - I just try to go with the
flow, and I can read it either way.

I will admit, though, that some of the posts here contain WAY too much
back-data.  Edit, edit, edit!

'nuff said!

;^)))

mhr

 BFG!

YEAH!
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Linux
On Wed, May 14, 2008 at 9:58 PM, Doug Tucker [EMAIL PROTECTED] wrote:

 This is a matter of agreeing to disagree on the release of a kernel and
 a supported file system.  If you had read my thread and subsequent
 paragraph you're taking issue with properly, you would have gotten that.
 My whole issue is around GFS, which is officially supported (someone
 else hijacked this thread with XFS which got more attention), and in my
 statement I said: Keep in mind this is not an unsupported XFS that
 someone hijacked my thread with.  So I'm agreeing that XFS should never
 be brought up in the same fashion as GFS, as it is not a supported file
 system.  GFS is, and it is my opinion RH should release the 2 together.

Sorry pal, it's me who stole your thread with XFS. I feel obliged to
give an answer although which I do not have to but I'll.

I've been so far away from CentOS/RHEL that I even did not know the
difference between XFS and GFS which is officially supported by Redhat
guys. And CentOS' guys kindness about giving us a chance to use XFS is
really attracks my appreciation. Up to this was for my apology.

BUT (a big one);

People who prepare and maintain a distro have (and should have) many
concerns in mind. Security is one of them and integrity is another.
But in this situation, integrity is simply ignored (on the behalf of
GFS situation because I backed down from my XFS related complains)

Disabling kernel upgrades simply solves the situation but raises some
other questions about What else can be broken with security
apprehensions?

I do not know which one to choose:
- Absolutely not-working server because of missing updates
- Maybe will be attacked server because of missing security updates.
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Doug Tucker

 This is linked from the CentOS FAQ:
 
 http://www.caliburn.nl/topposting.html
 
 Akemi

LOL!  This is just TOO good.


1. Because it is proper Usenet Etiquette.

...all but dead...I run a usenet server here, had 3 logins last
month...user base is over 4000...

2.We use a good news reader like Forte Agent.

OMG.  I haven't used a usenet reader in 10 years for anything.  Assumed
Forte Agent went out of development years ago.

I'll stop there, there is not a single thing on that page I can agree with 
anymore, 
technology, email and the web have moved on beyond that ideology of old.

I'm already at about 50% of the time reading email on my iphone mail app.  Like 
it
or not for the religious users (and I'll count myself there in many 
categories), eventually
most of our mail will be read on a handheld device.  So the 2 line preview pane 
at the top 
before deciding to atually open the message becomes very relevant, which does 
not lend
itself useful in bottom posting.  I can't remember the last time I saw a 
desktop user
regardless of client not read their mail using the preview pane.  They need 
to just 
rename that, as people even rarely click to open the message anymore.  Again, 
not good
when bottom posting.  I got poo-poo'd off about my GFS/kernel release schedule, 
for being 
in some small minority.  So, where are bottom posters, in terms of majority 
these days?
Maybe it's time, to update with the times?

Go ahead, let the bashing begin!

I'm off to another building, taking my email in my pocket with me...

___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread John R Pierce

Linux wrote:

People who prepare and maintain a distro have (and should have) many
concerns in mind. Security is one of them and integrity is another.
But in this situation, integrity is simply ignored (on the behalf of
GFS situation because I backed down from my XFS related complains)

Disabling kernel upgrades simply solves the situation but raises some
other questions about What else can be broken with security
apprehensions?

I do not know which one to choose:
- Absolutely not-working server because of missing updates
- Maybe will be attacked server because of missing security updates.
  


specific to GFS...   GFS is a clustered file system.  You do NOT run 
automatic updates willy-nilly on a production cluster, there's just far 
too many ways it can go bad.  You test them on a staging environment 
before approving their deployment, then you have to have a specific 
process for applying the patches to the cluster, and if they are major 
patches, this usually involves bringing the cluster down, applying the 
tested and approved patches to all cluster members, then bringing the 
cluster back up one node at a time, then going back live for 
production. If the patches are minor, you may be able to do a 
rolling upgrade, where you bring down one cluster member, patch it, put 
it back online, then bring down the next, etc...   The cluster 
administrator have to determine the appropriate maintenance process, 
then follow it religiously.






btw, what is WITH all these lame gmail addresses?   linuxlist ?   
centoslist ??   Do I call you Mr Linux, or Mr List ?



___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Johnny Hughes

Doug Tucker wrote:

On Wed, 2008-05-14 at 12:37 -0700, John R Pierce wrote:

Doug Tucker wrote:

My whole issue is around GFS, which is officially supported (someone
else hijacked this thread with XFS which got more attention), and in my
statement I said: Keep in mind this is not an unsupported XFS that
someone hijacked my thread with.  So I'm agreeing that XFS should never
be brought up in the same fashion as GFS, as it is not a supported file
system.  GFS is, and it is my opinion RH should release the 2 together.
  


GFS is only 'officially supported' under a seperate  contract from 
Red Hat.  


And?  It's official.  In fact, ext3 is only officially supported from
them these day without a $$$ contract.  Which is why we're all here!  :D


But, RHCS and RHGFS are not part of RHEL, and not part of base CentOS 
(before centos-5 that is).  It is an addon repository. We do update it, 
but it takes a back seat to the main centos repo.


Regardless ... I am building those updates and they should be released 
after I QA them sometime later today.





And, if you're a GFS customer of Red Hat's, I'm pretty darn 
sure the first thing they do is disable kernel updates... In fact, I 
seem to recall that RHEL4 ships with kernel updates disabled, you have 
to use `up2date --force` or something to enable them.


Yes, but kernel is disabled from EL4 reguardless of filesystem, so GFS
has nothing to do with that.  YOu can just edit the up2date file to
remove that.  I merely believe that GFS filesystem updates should be
released in conjuntion with kernel with all the other filesystems built
in, treating it no differently since it is officially supported, just
not put in the standard kernel build to put separation between it and
the $$ extra product.  And that is merely, an opinion.


Sure ... the reason they want you to manually update the kernel is that 
for all but the most basic of systems, you have to think BEFORE you 
update it.


All I am saying is that GFS (and any other ADDED repo besides Base or 
Updates) will get updates ... however they are not normally going to be 
as fast as the Base and Updates repos.  That is just how it goes.


Thanks,
Johnny Hughes



signature.asc
Description: OpenPGP digital signature
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Johnny Hughes

Doug Tucker wrote:

This is linked from the CentOS FAQ:

http://www.caliburn.nl/topposting.html

Akemi


LOL!  This is just TOO good.


1. Because it is proper Usenet Etiquette.

...all but dead...I run a usenet server here, had 3 logins last
month...user base is over 4000...

2.We use a good news reader like Forte Agent.

OMG.  I haven't used a usenet reader in 10 years for anything.  Assumed
Forte Agent went out of development years ago.

I'll stop there, there is not a single thing on that page I can agree with anymore, 
technology, email and the web have moved on beyond that ideology of old.


I'm already at about 50% of the time reading email on my iphone mail app.  Like 
it
or not for the religious users (and I'll count myself there in many 
categories), eventually
most of our mail will be read on a handheld device.  So the 2 line preview pane at the top 
before deciding to atually open the message becomes very relevant, which does not lend

itself useful in bottom posting.  I can't remember the last time I saw a 
desktop user
regardless of client not read their mail using the preview pane.  They need to just 
rename that, as people even rarely click to open the message anymore.  Again, not good
when bottom posting.  I got poo-poo'd off about my GFS/kernel release schedule, for being 
in some small minority.  So, where are bottom posters, in terms of majority these days?

Maybe it's time, to update with the times?

Go ahead, let the bashing begin!

I'm off to another building, taking my email in my pocket with me...


OK ... you are officially an ass .. I will no longer reply to your mails 
or help you in any way.






signature.asc
Description: OpenPGP digital signature
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Linux
On Wed, May 14, 2008 at 11:50 PM, John R Pierce [EMAIL PROTECTED] wrote:
 btw, what is WITH all these lame gmail addresses?   linuxlist ?   centoslist
 ??   Do I call you Mr Linux, or Mr List ?

Nothing to do with gmail. About calling me, it's a nice thing but
probably not needed. And I also know about usenet etiquette.

Well, I post in so many different threads and I do not want someone
googling and finding all about me. Besides, you people have the right
to ignore my-type people since not using real (or reallike nick)names.

Instead of deceiving people with different names on different
platforms, I prefer being honest about hiding my i.d. and I think this
is also my right (as your ignoring right)

But if it'll satisfy someone, I can choose some real-looking nicknames
from now on :)

Thanks...
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Jim Perrin
On Wed, May 14, 2008 at 4:58 PM, Johnny Hughes [EMAIL PROTECTED] wrote:

 OK ... you are officially an ass .. I will no longer reply to your mails or
 help you in any way.


Yes. When I signed on with CentOS it was explicitly written into my
requirements that *I* be the only 'official' ass. Yes, a non-compete
clause is involved, so can all just STEP OFF!

:-P

/humor


-- 
During times of universal deceit, telling the truth becomes a revolutionary act.
George Orwell
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread jleaver+centos



Yes. When I signed on with CentOS it was explicitly written into my
requirements that *I* be the only 'official' ass. Yes, a non-compete
clause is involved, so can all just STEP OFF!

:-P

/humor


  
Haha, thanks for the humorous remark!  It has been said that If you get 
too serious, you'll spoil all the fun.


I imagine that most of the folks subscribed are System Administrators, 
Engineers and Architects.  I'd also leap to the unproven assumption that 
the majority are overworked, underpaid, stressed, and stuff like that.   
If that doesn't make for a bunch of terse, grumpy, and otherwise 
friendly-and-cheer-challenged folks at times, well, you're better folks 
than myself, which, admittedly,  isn't all that difficult, and a little 
humor can go a long way.


Personally, I'd prefer top posting, I don't have an issue reading 
messages staged that way.  To me, scrolling down to see history makes a 
great deal of sense, as it means I see the most pertinent portion of a 
message first. (Arguably pertinent, however if it wasn't for the most 
recent content, the message wouldn't have been sent.)  However, I don't 
care enough to make an issue out of it, while some obviously have strong 
preferences to bottom posting.  What is REALLY not helpful is top, 
bottom, top, bottom posting, and thus I go with the norm.


Either way, while building and sending escalatory non-main-topic content 
(ie, flame wars) are as traditional as bottom posting, I think we'd be 
better off without.


Peace,

Jacob Leaver
Sr. Systems Administrator
ReachONE Internet
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Doug Tucker

 All I am saying is that GFS (and any other ADDED repo besides Base or 
 Updates) will get updates ... however they are not normally going to be 
 as fast as the Base and Updates repos.  That is just how it goes.
I can totally live with that, I was just b**ching about RH's approach.
I'm not expecting centos to do anything more, I appreciate the fact that
this exists, as it keeps me from having to use debian :).

___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread Doug Tucker

 OK ... you are officially an ass .. I will no longer reply to your mails 
 or help you in any way.

Wow.  My apologies, I thought that was actually a productive reply, not
even sure how you got offended, but I will apologize anyway, I don't
intend to ever offend anyone.

___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-14 Thread MHR
On Wed, May 14, 2008 at 3:02 PM, Doug Tucker [EMAIL PROTECTED] wrote:

 Agreed to all, and I was just having some fun and trying to bring some
 humor to everyone's day.  Thanks for having a sense of humor, I'll
 respectfully bow out now.


There you go, man
Keep as cool as you can
Face piles of trials with smiles
It riles them to believe
That you perceive
The web they weave...
And keep on thinking free
 - In the Beginning, On the Threshold of a Deam, the Moody Blues

:-)

mhr
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


[CentOS] broken GFS

2008-05-12 Thread Doug Tucker
This is the 2nd time this has happened to me.  There was a kernel
release over the weekend to .67.0.15, yet, they did not release the
updated GFS to go along with it, so when the machine rebooted, there was
no gfs file system in the new running kernel which in turn wreaked havoc
on my cluster.  I truly wish they would not do that :).  I guess I shall
have to not allow automatic yum updates from these machines.

___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-12 Thread Linux
Well, I should add a terrible story for XFS...

I did a yum update and after updating many packages I rebooted and viola...
Old xfs module ruined my 1.2TB partition. After updating to correct module and
hours of xfs_repair I had to move and rename 500 subfolders from lost+found.

I am using CentOS because I have to (for cPanel).
I am not very comfortable with it (indeed I am a pro gentooer for 5-7 years)

On Mon, May 12, 2008 at 10:51 PM, Doug Tucker [EMAIL PROTECTED] wrote:
 This is the 2nd time this has happened to me.  There was a kernel
  release over the weekend to .67.0.15, yet, they did not release the
  updated GFS to go along with it, so when the machine rebooted, there was
  no gfs file system in the new running kernel which in turn wreaked havoc
  on my cluster.  I truly wish they would not do that :).  I guess I shall
  have to not allow automatic yum updates from these machines.

  ___
  CentOS mailing list
  CentOS@centos.org
  http://lists.centos.org/mailman/listinfo/centos

___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


RE: [CentOS] broken GFS

2008-05-12 Thread Flaherty, Patrick
 This is the 2nd time this has happened to me.  There was a kernel
 release over the weekend to .67.0.15, yet, they did not release the
 updated GFS to go along with it, so when the machine 
 rebooted, there was
 no gfs file system in the new running kernel which in turn 
 wreaked havoc
 on my cluster.  I truly wish they would not do that :).  I 
 guess I shall
 have to not allow automatic yum updates from these machines.

Use the yum's exclude functionality. Man yum.conf for the syntax. I
think it will just be exclude=kernel. 

You also might want to remove the non gfs kernels from your installation
and get a staging environment for patching set up (if this is a
production system).

Best
Patrick
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-12 Thread Tru Huynh
On Mon, May 12, 2008 at 11:00:18PM +0300, Linux wrote:
 Well, I should add a terrible story for XFS...
 
 I did a yum update and after updating many packages I rebooted and viola...
You seem to enjoy living dangerously ? Don't you ever use a testing machine
before rolling the updates on a production server?
We appreciate your trust in our project, but you should always test on your
own setup.

 Old xfs module ruined my 1.2TB partition. After updating to correct module and
 hours of xfs_repair I had to move and rename 500 subfolders from lost+found.
That is the 1st time I hear such a story: if the xfs module is not installed
for your new kernel, the only thing that should happen is the inability to
mount the XFS filesystem.
 
 I am using CentOS because I have to (for cPanel).
That's trolling, CPanel is NOT CentOS...

Tru
-- 
Tru Huynh (mirrors, CentOS-3 i386/x86_64 Package Maintenance)
http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xBEFA581B


pgpoqJrT1KNlf.pgp
Description: PGP signature
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-12 Thread Doug Tucker
Tru,

I work at a university.  They don't provide enough money for test
environments :).  Just kinda odd, last time kernel update, gfs updated
at the same time so all was well.  But twice now kernel has upgraded
with no GFS so it went bye-bye.  Is the GFS being installed, compiled
against particular kernel headers, or could I just copy the /fs/gfs
and /fs/gfs_locking to the new kernel /lib/modules (or symlink for that
matter) and be lucky enough it would work?


On Mon, 2008-05-12 at 22:48 +0200, Tru Huynh wrote:
 On Mon, May 12, 2008 at 11:00:18PM +0300, Linux wrote:
  Well, I should add a terrible story for XFS...
  
  I did a yum update and after updating many packages I rebooted and 
  viola...
 You seem to enjoy living dangerously ? Don't you ever use a testing machine
 before rolling the updates on a production server?
 We appreciate your trust in our project, but you should always test on your
 own setup.
 
  Old xfs module ruined my 1.2TB partition. After updating to correct module 
  and
  hours of xfs_repair I had to move and rename 500 subfolders from lost+found.
 That is the 1st time I hear such a story: if the xfs module is not installed
 for your new kernel, the only thing that should happen is the inability to
 mount the XFS filesystem.
  
  I am using CentOS because I have to (for cPanel).
 That's trolling, CPanel is NOT CentOS...
 
 Tru
 ___
 CentOS mailing list
 CentOS@centos.org
 http://lists.centos.org/mailman/listinfo/centos

___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-12 Thread Scott Thistle
Upstream updates cluster packages about a week after OS patches. I found
that out when putting in a new cluster and 4.6 came out. The cluster
packages lagged behind a week deliberately for stability's sake.

Scott

On Mon, May 12, 2008 at 6:50 PM, Doug Tucker [EMAIL PROTECTED] wrote:

 Tru,

 I work at a university.  They don't provide enough money for test
 environments :).  Just kinda odd, last time kernel update, gfs updated
 at the same time so all was well.  But twice now kernel has upgraded
 with no GFS so it went bye-bye.  Is the GFS being installed, compiled
 against particular kernel headers, or could I just copy the /fs/gfs
 and /fs/gfs_locking to the new kernel /lib/modules (or symlink for that
 matter) and be lucky enough it would work?


 On Mon, 2008-05-12 at 22:48 +0200, Tru Huynh wrote:
  On Mon, May 12, 2008 at 11:00:18PM +0300, Linux wrote:
   Well, I should add a terrible story for XFS...
  
   I did a yum update and after updating many packages I rebooted and
 viola...
  You seem to enjoy living dangerously ? Don't you ever use a testing
 machine
  before rolling the updates on a production server?
  We appreciate your trust in our project, but you should always test on
 your
  own setup.
 
   Old xfs module ruined my 1.2TB partition. After updating to correct
 module and
   hours of xfs_repair I had to move and rename 500 subfolders from
 lost+found.
  That is the 1st time I hear such a story: if the xfs module is not
 installed
  for your new kernel, the only thing that should happen is the inability
 to
  mount the XFS filesystem.
  
   I am using CentOS because I have to (for cPanel).
  That's trolling, CPanel is NOT CentOS...
 
  Tru
  ___
  CentOS mailing list
  CentOS@centos.org
  http://lists.centos.org/mailman/listinfo/centos

 ___
 CentOS mailing list
 CentOS@centos.org
 http://lists.centos.org/mailman/listinfo/centos

___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-12 Thread Tru Huynh
On Mon, May 12, 2008 at 04:20:46PM -0500, Doug Tucker wrote:
 Tru,
 
Hi Doug,

 I work at a university.  They don't provide enough money for test
 environments :).
Sure, but they could probably understand if it's 'critical' ;)
  Just kinda odd, last time kernel update, gfs updated
 at the same time so all was well.  But twice now kernel has upgraded
CentOS can't rebuild if upstream hasn't released the corresponding src.rpm.
Upstream's GFS is often released later than the kernel RHSA...

 with no GFS so it went bye-bye.  Is the GFS being installed, compiled
 against particular kernel headers, or could I just copy the /fs/gfs
 and /fs/gfs_locking to the new kernel /lib/modules (or symlink for that
 matter) and be lucky enough it would work?
No idea.

Cheers,

Tru
-- 
Tru Huynh (mirrors, CentOS-3 i386/x86_64 Package Maintenance)
http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xBEFA581B


pgpBkO33c4XUl.pgp
Description: PGP signature
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-12 Thread Linux
On Mon, May 12, 2008 at 11:48 PM, Tru Huynh [EMAIL PROTECTED] wrote:
 On Mon, May 12, 2008 at 11:00:18PM +0300, Linux wrote:
   Well, I should add a terrible story for XFS...
  
   I did a yum update and after updating many packages I rebooted and 
 viola...
  You seem to enjoy living dangerously ? Don't you ever use a testing machine
  before rolling the updates on a production server?
  We appreciate your trust in our project, but you should always test on your
  own setup.
Indeed that was my low-value testing machine. But I cannot afford a
third machine just for testing with the same hardware. Besides, if you
suggest a VM testing, it is not a real testing, not better than you
people do...

   Old xfs module ruined my 1.2TB partition. After updating to correct module 
 and
   hours of xfs_repair I had to move and rename 500 subfolders from 
 lost+found.
  That is the 1st time I hear such a story: if the xfs module is not installed
  for your new kernel, the only thing that should happen is the inability to
  mount the XFS filesystem.
What a coincidence. That is the 1st time I live such a thing. Well,
show me a way to prove.

   I am using CentOS because I have to (for cPanel).
  That's trolling, CPanel is NOT CentOS...
Neither I am a troll, nor do I know its meaning. And I do not have an
intention to blame CentOS for anything. I have to use CentOS because
it's the best of the choises CPanel requires. I am not keen on CentOS
way of eating yoghurt. [1]

This log is after update  reboot:
May 11 16:06:03 x kernel: XFS: failed to read root inode

And this is the last yum.log beginning from a month before:
---
Apr 02 23:40:03 Updated: krb5-libs.x86_64 1.6.1-17.el5_1.1
Apr 02 23:40:04 Updated: cups-libs.x86_64 1:1.2.4-11.14.el5_1.6
Apr 02 23:40:04 Updated: openldap.x86_64 2.3.27-8.el5_1.3
Apr 02 23:40:05 Updated: dbus.x86_64 1.0.0-6.3.el5_1
Apr 02 23:40:11 Updated: ghostscript.x86_64 8.15.2-9.1.el5_1.1
Apr 02 23:40:12 Updated: tk.x86_64 8.4.13-5.el5_1.1
Apr 02 23:40:13 Updated: kpartx.x86_64 0.4.7-12.el5_1.3
Apr 02 23:40:13 Updated: device-mapper-multipath.x86_64 0.4.7-12.el5_1.3
Apr 02 23:40:23 Updated: cups.x86_64 1:1.2.4-11.14.el5_1.6
Apr 02 23:40:23 Updated: autofs.x86_64 1:5.0.1-0.rc2.55.el5.3
Apr 02 23:40:23 Updated: krb5-libs.i386 1.6.1-17.el5_1.1
Apr 02 23:40:24 Updated: cups-libs.i386 1:1.2.4-11.14.el5_1.6
Apr 02 23:40:33 Updated: ghostscript.i386 8.15.2-9.1.el5_1.1
Apr 02 23:40:33 Updated: dbus.i386 1.0.0-6.3.el5_1
Apr 02 23:40:33 Updated: openldap.i386 2.3.27-8.el5_1.3
Apr 02 23:40:34 Updated: tk.i386 8.4.13-5.el5_1.1
Apr 02 23:41:13 Installed: kernel.x86_64 2.6.18-53.1.14.el5
Apr 02 23:41:23 Updated: tzdata.noarch 2007k-2.el5
Apr 02 23:41:24 Updated: krb5-devel.i386 1.6.1-17.el5_1.1
Apr 02 23:41:25 Updated: krb5-workstation.x86_64 1.6.1-17.el5_1.1
Apr 02 23:41:26 Updated: krb5-devel.x86_64 1.6.1-17.el5_1.1
Apr 02 23:41:30 Updated: kernel-headers.x86_64 2.6.18-53.1.14.el5
Apr 02 23:41:30 Installed: kmod-xfs.x86_64 0.4-1.2.6.18_53.1.14.el5
May 11 00:34:48 Updated: ImageMagick.x86_64 6.2.8.0-4.el5_1.1
May 11 00:34:52 Updated: ImageMagick.i386 6.2.8.0-4.el5_1.1
May 11 00:34:57 Updated: kernel-headers.x86_64 2.6.18-53.1.19.el5
May 11 00:35:04 Updated: squid.x86_64 7:2.6.STABLE6-5.el5_1.3
May 11 00:35:04 Updated: sos.noarch 1.7-9.2.el5
May 11 00:35:36 Installed: kernel.x86_64 2.6.18-53.1.19.el5
May 11 01:28:19 Installed: hddtemp.x86_64 0.3-0.14.beta15.el5.centos
May 11 01:40:35 Installed: apt.x86_64 0.5.15lorg3.2-1.el5.rf
May 11 17:13:03 Installed: kmod-xfs.x86_64 0.4-1.2.6.18_53.1.19.el5

According to this, there is a mystery in May 11 16:06:03 because
there WAS a kmod_xfs but it was 53.1.14, not 53.1.19 as updated
kernel.

By the way, sorry for stealing some GFS thread but I see something
parallel in deep.

Thanks.

[1] There is a proverb like Every knight has his own way of eating
yoghurt which means you can eat yoghurt in different ways and also
envied people can eat it differently, which none of them is wrong. In
the end, yoghurt, a very useful nutrient, is eaten anyway.
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-12 Thread Tru Huynh
On Tue, May 13, 2008 at 12:40:22AM +0300, Linux wrote:
 What a coincidence. That is the 1st time I live such a thing. Well,
 show me a way to prove.
/var/log/messages ?

 This log is after update  reboot:
 May 11 16:06:03 x kernel: XFS: failed to read root inode
nothing more?

 Apr 02 23:41:30 Installed: kmod-xfs.x86_64 0.4-1.2.6.18_53.1.14.el5
kmod-xfs for 2.6.18_53.1.14.el5

 May 11 00:35:36 Installed: kernel.x86_64 2.6.18-53.1.19.el5
...
 May 11 17:13:03 Installed: kmod-xfs.x86_64 0.4-1.2.6.18_53.1.19.el5
and the corresponding kmod-xfs module (2.6.18-53.1.19.el5)

 According to this, there is a mystery in May 11 16:06:03 because
 there WAS a kmod_xfs but it was 53.1.14, not 53.1.19 as updated
 kernel.
too bad you rebooted 1 hour before the kernel-xfs module update.

Tru
-- 
Tru Huynh (mirrors, CentOS-3 i386/x86_64 Package Maintenance)
http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xBEFA581B


pgp33QZqvymE7.pgp
Description: PGP signature
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-12 Thread Linux
On Tue, May 13, 2008 at 1:15 AM, Tru Huynh [EMAIL PROTECTED] wrote:
 On Tue, May 13, 2008 at 12:40:22AM +0300, Linux wrote:
   What a coincidence. That is the 1st time I live such a thing. Well,
   show me a way to prove.
  /var/log/messages ?
Only a small part of it.

   This log is after update  reboot:
   May 11 16:06:03 x kernel: XFS: failed to read root inode
  nothing more?
Well, that is the only unexpected part. Just to show that XFS module
was loaded for WRONG kernel. As you said, you newer saw before.

   According to this, there is a mystery in May 11 16:06:03 because
   there WAS a kmod_xfs but it was 53.1.14, not 53.1.19 as updated
   kernel.
  too bad you rebooted 1 hour before the kernel-xfs module update.
When was kernel-xfs module updated in repository? Just that time? If
so too bad CentOS folks do not update every piece of kernel as a whole
in repositories. Where is integrity?

If not, yum update does not update everything at once. I have to run
yum update twice maybe more. First it will load kernel then see that a
new kernel is available, will go and bring its modules...

Still, it is a bit annoying and confusing. I am beginning to think
whether XFS is really supported in CentOS :)
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos


Re: [CentOS] broken GFS

2008-05-12 Thread Fajar Priyanto
On Tuesday 13 May 2008 04:20:46 Doug Tucker wrote:
 Tru,

 I work at a university.  They don't provide enough money for test
 environments :).  Just kinda odd, last time kernel update, gfs updated
 at the same time so all was well.  But twice now kernel has upgraded
 with no GFS so it went bye-bye.  Is the GFS being installed, compiled
 against particular kernel headers, or could I just copy the /fs/gfs
 and /fs/gfs_locking to the new kernel /lib/modules (or symlink for that
 matter) and be lucky enough it would work?

From my experience, if the production server is running OK, and the update is 
not security-related, then there is NO NEED to update in your situation.
If you DO want to update for whatever reason, test it first in a testbed.

Trust me. It comes from a traumatic experience :)

-- 
Fajar Priyanto | Reg'd Linux User #327841 | Linux tutorial 
http://linux2.arinet.org
07:51:43 up 54 min, 2.6.22-14-generic GNU/Linux 
Let's use OpenOffice. http://www.openoffice.org
The real challenge of teaching is getting your students motivated to learn.


signature.asc
Description: This is a digitally signed message part.
___
CentOS mailing list
CentOS@centos.org
http://lists.centos.org/mailman/listinfo/centos