Topic: Upgrade from 7.3 to 7.4 kernel panic

[mudtoe]

Hi folks:

I was trying to upgrade from 7.3 to 7.4 using the upgrade ISO image and I'm having some problems.  After the upgrade completes and it reboots I'm getting a kernel panic error that looks like SME 7.4 is unable to mount the file system.  I verified that the file system is still there by booting into the 7.4 rescue environment and it does mount the file system under /mnt/sysimage and I'm able to access the files from there, so there hasn't been any damage do the file system.  Fortunately SME is running as a VM and I took a snapshot before the upgrade, so I can go back easily, but I'm not sure what's wrong.  I looked at the changes made to grub.conf from 7.3 to 7.4 and I don't see anything obvious, but then again I'm not a real Linux expert.  I'm attaching images of the console showing the 7.4 panic error, and the grub.conf files for the 7.3 and 7.4 installations. 

Suggestions welcome.

mudtoe

=========================

Kernel Panic Error:




SME 7.3 grub.conf:




SME 7.4 grub.conf:

[gzartman]

Please post this over in the bug tracker:  http://bugs.contribs.org/

If there is a problem with SME 7.4, the bug tracker is the place to get it solved.

[mudtoe]

Well, for the record, if I do the upgrade via the yum upgrade screen instead of booting the 7.4 CD and doing the upgrade from the CD, everything works find and I don't get the kernel panic error.
 
mudtoe

[jptechnical]

Well, for the record, if I do the upgrade via the yum upgrade screen instead of booting the 7.4 CD and doing the upgrade from the CD, everything works find and I don't get the kernel panic error.
 
mudtoe

Can you tell me if the yum upgrade screen you mentioned is the same 'Software Installer' page in the server/manager? I haven't upgraded any systems a full release, just updates via the server-manager. Just curious if installing the updates via the https://hostname/server-manager/cgi-bin/yum is acceptable as opposed to booting from the CD. I have several installs in semi-remote locations... trying to spare some panicy service calls.

Thanks.

[cactus]

Well, for the record, if I do the upgrade via the yum upgrade screen instead of booting the 7.4 CD and doing the upgrade from the CD, everything works find and I don't get the kernel panic error.
 
mudtoe

Also please report that to the bugtracker. Since no additional updates have been released AFAIK the update from CD should do the same as per server-manager.

[Stefano]

Can you tell me if the yum upgrade screen you mentioned is the same 'Software Installer' page in the server/manager? I haven't upgraded any systems a full release, just updates via the server-manager. Just curious if installing the updates via the https://hostname/server-manager/cgi-bin/yum is acceptable as opposed to booting from the CD. I have several installs in semi-remote locations... trying to spare some panicy service calls.

Thanks.

such an upgrade should be done via shell access.. not via web interface.

IMHO

Ciao
Stefano

[cactus]

such an upgrade should be done via shell access.. not via web interface.

I am curious to know why as they should both do the same.

[mudtoe]

Can you tell me if the yum upgrade screen you mentioned is the same 'Software Installer' page in the server/manager? I haven't upgraded any systems a full release, just updates via the server-manager. Just curious if installing the updates via the https://hostname/server-manager/cgi-bin/yum is acceptable as opposed to booting from the CD. I have several installs in semi-remote locations... trying to spare some panicy service calls.

Yes, I did the update from the Software Installer page in the Server Manager.   I have no idea why doing it from the CD resulted in the kernel panic error and doing it from the software installer didn't.  I did it from the CD first because I figured it might be better to do a big update like that when the server itself wasn't running; however it turned out that the other way around was the right way.  Also, someone posted about doing it from the command line, and the only reason I didn't do it that way was that there were some updates listed for Cyrus (I'm using Toltec Connector so I had to install Cyrus as the IMAP software) that I didn't want to apply at the same time as a version upgrade.  I also have no idea if the presence of Cyrus somehow screwed up the CD upgrade; however, since the kernel panic error had something to do with not being able to mount the file system I wouldn't think that the presence of Cyrus would cause that kind of error.  If Cyrus was going to be an issue for the CD upgrade I would expect problems with IMAP after the upgrade, as opposed to a problem with the whole file system.
 
mudtoe

[gustlsoft]

Hello,
I ran into the same problem. My SME Server is a virtual machine too (on ESXi 3.5U3).
Unfortunatly I didn't make a snapshot before updating.
So I googled around and found a posting in the VMware Community:
Red Hat 4 virtual machine with "kernel" panic after update 4 on ESX 3.0.1
http://communities.vmware.com/thread/67781;jsessionid=ACB1F00C46A833E15B862E2FD73DAE1E?tstart=0
Have a look at tsightler's Post from Jan 18, 2007 10:46 AM
This worked for me!
Delete all SCSI_HOSTADAPTER lines from /etc/modprobe.conf, leaving one line:
alias scsi_hostadapter mptscsih
Then rename the /boot/initrd..................img files to something like initrd...............img.old
and do a fresh:
mkinitrd /boot/initrd-2.6.9-78.0.8.EL.img 2.6.9-78.0.8.EL
Reboot, configure vmware-tools and enjoy your fresh SME 7.4

[Stefano]

I am curious to know why as they should both do the same.

just because, via console, you can have a better control of what is going on..

I repeat, IMHO.

Ciao
Stefano

[m]

FYI:
On a fresh SME 7.4 install and SME 7.4 updated from 7.3 from ISO image, both running on ESXi3.5 Update3, I have these scsi_hostadapter aliases:

Code: [Select]

alias scsi_hostadapter mptbase
alias scsi_hostadapter1 mptscsi
alias scsi_hostadapter2 mptspi
alias scsi_hostadapter3 mptsas
alias scsi_hostadapter4 mptscsih
alias scsi_hostadapter5 ata_piix


[mark]

On Friday I updated a 7.3 with yum update which appeared to go smoothly untill the reboot when it refused to boot saying there were disk errors (there weren't) - The only way I could get it to boot was to select the previous kernel then it seemed  fine.

I updated my office server today using yum update from 7.3 and it ignored the signal-event reboot (but not the subsequent reboot cmd) then came up without internet access. Another 'signal-event post-upgrade' and warm boot didn't fix the problem but a subsequent hard boot brought everything to life.

I certainly wont be updating my remote servers for a while.

And before someone jumps on his/her high horse and tells me to use the bug tracker - I have tried several times and simply find it too confusing - I can't/wont be bothered.

cheers

Mark

[cactus]

And before someone jumps on his/her high horse and tells me to use the bug tracker - I have tried several times and simply find it too confusing - I can't/wont be bothered.

But I will ride that horse, it is not very much harder than writing in the proper board on the forums. If you do not report it as a bug it is not likely to be fixed. So please, take a plunge in the bugtracker and report where you get stuck so we can guide you. The bugtracker is the single and only place to report problems if you want them fixed.
With the use of open source like this you are morally obliged to do something back IMHO, which in your case would be take a little effort to learn how to work the bugtracker to report a new bug and follow-up when questions are asked. So I suggest you saddle up!

[okepc]

I have exact the same problem as the topic starter.
Running a scsi raid.
I have had this problem before some while ago and there is an old bug somewhere in the bug tracker.
As of now im running on the previous kernel.

Is there already filed a bug and so what's the link?

Dirk

[okepc]

Old but..

http://bugs.contribs.org/show_bug.cgi?id=698