Categories
Uncategorized

Is Ubuntu 12.04 not so Precise?

I’ve been running 10.04 in my office running two servers. One a NAS, E-MAIL, DHCP server. This was working great and was solid and bug free.

Than one day I upgraded the backup server to 12.04. Its been hell ever since. First the upgrade failed on the GRUB update. It’s never great when you upgrade and the server doesn’t boot. You then pull your hair out getting a live CD to restore GRUB.

Then I find that backuppc isĀ  broken. Well not backuppc but Samba 3.6.3 has a bug which breaks backuppc. So far, nearly one year later the fix in later Samba releases has not been back-ported to 12.04. Thanks Ubuntu! (I don’t mean that of course). The way to get around this was to do a downgrade (see this post).

Over the past 4 months I’ve been experiencing serious problems with corruption on the RAID that stores all the machine backups. This is a serious EXT4 bug. My logs are filled with

Mar 20 21:54:48 chamraid01 kernel: [386700.256102] EXT4-fs (md3): initial error at 1363470353: ext4_mb_generate_buddy:739
Mar 20 21:54:48 chamraid01 kernel: [386700.256115] EXT4-fs (md3): last error at 1363813392: ext4_mb_generate_buddy:739
Mar 20 22:32:39 chamraid01 kernel: [388971.140821] EXT4-fs error (device md3): ext4_mb_generate_buddy:739: group 1850, 26103 clusters in bitmap, 26102 in gd
Mar 20 22:32:56 chamraid01 kernel: [388987.811768] EXT4-fs warning (device md3): empty_dir:1929: bad directory (dir #17228833) - no `.' or `..'
Mar 21 00:01:28 chamraid01 kernel: [394299.785367] EXT4-fs error (device md3): ext4_mb_generate_buddy:739: group 1960, 28847 clusters in bitmap, 28846 in gd
Mar 21 01:18:56 chamraid01 kernel: [398947.816401] EXT4-fs error (device md3): ext4_mb_generate_buddy:739: group 2183, 30388 clusters in bitmap, 30387 in gd
Mar 21 02:36:16 chamraid01 kernel: [403587.581270] EXT4-fs error (device md3): ext4_mb_generate_buddy:739: group 188, 4447 clusters in bitmap, 4446 in gd
Mar 21 11:46:43 chamraid01 kernel: [436615.298941] EXT4-fs error (device md3): ext4_mb_generate_buddy:739: group 2210, 27659 clusters in bitmap, 27658 in gd
Mar 21 12:14:06 chamraid01 kernel: [438257.751369] EXT4-fs error (device md3): ext4_mb_generate_buddy:739: group 188, 4446 clusters in bitmap, 4447 in gd
Mar 21 21:56:36 chamraid01 kernel: [473207.776234] EXT4-fs (md3): error count: 16
Mar 21 21:56:36 chamraid01 kernel: [473207.776261] EXT4-fs (md3): initial error at 1363470353: ext4_mb_generate_buddy:739
Mar 21 21:56:36 chamraid01 kernel: [473207.776274] EXT4-fs (md3): last error at 1363868046: ext4_mb_generate_buddy:739
Mar 22 21:18:21 chamraid01 kernel: [557313.051610] EXT4-fs error (device md3): ext4_mb_generate_buddy:739: group 2171, 30176 clusters in bitmap, 30175 in gd

Eventually enough errors accumulate and the OS takes the RAID offline.

Currently it is enough to run a weekly script which fschk’s the drive to clean the problems and bring it back online. But I’m not sure how much data is being lost by this. This doesn’t inspire confidence in the backups.

The issues continue. Truly Ubuntu 12.04 is the worst Ubuntu release thus far.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.