NOT afraid of a debian distro upgrade, partly success

hoekie · 06-01-2011, 07:04 AM

I just read Zippy's post, who is afraid of doing a distro upgrade, for a reason!. After some reading and preparations, i did do the aptitude dist-upgrade (Etch->Lenny) on a production machine. All went well, just a
couple of issues:

- The usual config suggestions,
- Some graphic pop-ups about programms to run at boot,
- and i had to remove lilo, somehow it was installed in the process, although the machine has GRUB.
so after 15min or so i was running 5.0.8 (cat /etc/debian_version).
But unfortunately it had the old kernel: 2.6.18-6-686-bigmem, which stayed the same after reboot.

Then a bit hasty i read about grub and update-grub. update-grub
discovered a new kernel (2.6.26-2-686-bigmem) and it updated /boot/grub/menu.lst. But then after reboot nothing!

I had to use rescue mode (@Zippy: dont be afraid of rescue mode, its your best friend) to the change menu.lst and reboot with 2.6.18. I cant pinpoint the problem at boot, the logfiles say:

/var/log/kern.log:
ata2.00: exception Emask 0x50 SAct 0x0 SErr 0x680900 action 0x2 frozen
ata2.00: (irq_stat 0x08000000, interface fatal error)
ata2.00: tag 0 cmd 0xc4 Emask 0x50 stat 0x50 err 0x0 (ATA bus error)
...
Buffer I/O error on device sdb2, logical block 1
...
sd 1:0:0:0: SCSI error: return code = 0x00040000
end_request: I/O error, dev sdb, sector 0
---

/var/log/messages
May 31 17:35:16 mymachine kernel: ata2: soft resetting port
May 31 17:35:16 mymachine kernel: ata2: softreset failed, retrying in 5 secs
May 31 17:35:21 mymachine kernel: ata2: hard resetting port
May 31 17:35:22 mymachine kernel: ata2: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
May 31 17:35:22 mymachine kernel: ata2.00: configured for PIO0
May 31 17:35:22 mymachine kernel: ata2: EH complete
May 31 17:35:22 mymachine kernel: SCSI device sdb: 976773168 512-byte hdwr sectors (500108 MB)
May 31 17:35:22 mymachine kernel: sdb: Write Protect is off
May 31 17:35:22 mymachine kernel: SCSI device sdb: drive cache: write back
...
May 31 17:35:53 mymachine kernel: sd 1:0:0:0: SCSI error: return code = 0x00040000
May 31 17:35:53 mymachine kernel: end_request: I/O error, dev sdb, sector 4209031
May 31 17:35:53 mymachine kernel: sd 1:0:0:0: SCSI error: return code = 0x00040000
May 31 17:35:53 mymachine kernel: end_request: I/O error, dev sdb, sector 4209036
May 31 17:35:53 mymachine kernel: sd 1:0:0:0: SCSI error: return code = 0x00040000
May 31 17:35:53 mymachine kernel: end_request: I/O error, dev sdb, sector 4209031
May 31 17:35:53 mymachine kernel: sdb : READ CAPACITY failed.
May 31 17:35:53 mymachine kernel: sdb : status=0, message=00, host=4, driver=00
May 31 17:35:53 mymachine kernel: sdb : sense not available.
May 31 17:35:53 mymachine kernel: sdb: Write Protect is off
May 31 17:35:53 mymachine kernel: sd 1:0:0:0: SCSI error: return code = 0x00040000
May 31 17:35:53 mymachine kernel: end_request: I/O error, dev sdb, sector 2097024
May 31 17:35:53 mymachine kernel: sd 1:0:0:0: SCSI error: return code = 0x00040000
May 31 17:35:53 mymachine kernel: end_request: I/O error, dev sdb, sector 2097024
...
May 31 17:37:30 mymachine shutdown[4663]: shutting down for system reboot

/var/log/syslog
pretty much the same

I suspect a driverproblem? What should be done to update the kernel?

H.

SalmonEater · 06-02-2011, 02:07 AM

Se smxi
It will save a lot of guesswork and time!

Welcome to linux!

craigevil · 06-02-2011, 06:43 PM

1) Fully upgrade etch first.
2) READ the release notes for Lenny and follow the upgrade steps
3) Read the release notes for Squeeze and follow the upgrade steps
4) Be happy!

hoekie · 06-03-2011, 05:45 AM

Quote:

Originally Posted by SalmonEater

Se smxi
It will save a lot of guesswork and time!

Welcome to linux!

Thanks, the script looks promising. I'll have to try it on local vm's first.

Quote:

Originally Posted by craigevil

1) Fully upgrade etch first.

I did.

Quote:

Originally Posted by craigevil

2) READ the release notes for Lenny and follow the upgrade steps

I did too (still have that page open), did I overread something? I read about udev, grub, some timing issues, ata2.00 bug/error, irqpoll...

Quote:

Originally Posted by craigevil

3) Read the release notes for Squeeze and follow the upgrade steps

Read those too, but didnt perform the upgrade steps yet. Now I'm wondering if it isn't easier to just start with a clean Squeeze image and get done with it, or to learn something and keep searching for a solution?

Quote:

Originally Posted by craigevil

4) Be happy!

I will, thank you.

hoekie · 06-12-2011, 03:17 AM

Turned out sdb was defect, the disk is in RAID1:

Code:

root@rescue ~ # cat /proc/mdstat
Personalities : [raid1]
md1 : active raid1 sda2[0]
      486279424 blocks [2/1] [U_]

md0 : active raid1 sda1[0]
      2104448 blocks [2/1] [U_]

unused devices: <none>

hoekie · 06-12-2011, 09:20 AM

Ok SOLVED!
I had the drive replaced and mirrored it from the working one, using:

Code:

dd if=/dev/sda of=/dev/sdb count=1 bs=512

to copy the partitiontable and:

Code:

sfdisk -R /dev/sdb

to activate it for the kernel. Then placed the drive back into the array:

Code:

mdadm /dev/md0 -a /dev/sdb1
mdadm /dev/md1 -a /dev/sdb2

The array will mirror the drive and you can follow the process with:

Code:

cat /proc/mdstat

Then it was time to try the new kernel, as proposed by grub in the first place. I changed /boot/grub/menu.lst back to original and a successful reboot made me happy!