Topic: kernel: ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0

[RCFerguson]

I'm getting this error consistently EVERY Sunday at approximately 00:45 local time:

Jun 13 00:44:42 server kernel: ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0
Jun 13 00:44:42 server kernel: ata1.00: (BMDMA stat 0x24)
Jun 13 00:44:42 server kernel: ata1.00: cmd 25/00:08:75:98:e3/00:00:4c:00:00/e0 tag 0 cdb 0x0 data 4096 in
Jun 13 00:44:42 server kernel:          res 51/40:00:75:98:e3/00:00:00:00:00/e0 Emask 0x9 (media error)
Jun 13 00:44:42 server kernel: ata1.00: configured for UDMA/133
Jun 13 00:44:42 server kernel: ata1: EH complete

I only found one mention of a similar error here in this forum and it didn't appear to resolve anything.  Various Google searches seem to hint at a Kernel bug, but nothing definite.  I would suspect a hard drive or controller problem, except for the fact that it always occurs once a week at the exact (ok... +/- 2 minutes) same time.  If it were a hardware error wouldn't it be random?
Server is SME 7.4 all updates applied.  Samsung 1TB hard drive.

[CharlieBrady]

It's a 'media error' on an ATA device (presumably a hard drive). The timing coincidence is likely due to some cron job running at that time, which scans the disk surface in a way that nothing else is doing.

[Stefano]

well, a s far as I understand you have sata hd.. so, maybe, the error is relative to the cd/dvd drive

[cactus]

What does /var/log/cron.* log say around Jun 13 00:44:42?

[RCFerguson]

The clamscan is the only item that only runs on Sunday. Everything else runs more often.

Jun 12 23:25:01 server crond[7941]: (root) CMD (perl /etc/e-smith/events/actions/purge-junkmail-folders)
Jun 13 00:01:01 server crond[8509]: (root) CMD (run-parts /etc/cron.hourly)
Jun 13 00:12:01 server crond[8722]: (root) CMD (/sbin/e-smith/smeserver-clamscan)
Jun 13 00:37:01 server crond[9197]: (root) CMD (/sbin/e-smith/warnquota)
Jun 13 01:12:57 server crond[9632]: (root) CMD (run-parts /etc/cron.hourly)
Jun 13 01:29:31 server crond[9770]: (root) CMD (/usr/bin/smoltSendProfile -c > /dev/null 2>&1)

It is a sATA hard drive, with an IDE CDROM drive.
I suppose if it IS a hard disk problem I could just use ghost or clonezilla to image onto a new drive

[CharlieBrady]

The clamscan is the only item that only runs on Sunday
....
It is a sATA hard drive, with an IDE CDROM drive.

Perhaps you have the clamscan configured incorrectly and it is scanning a CD. Do you have anything mounted in the CDROM drive?

[Stefano]

It is a sATA hard drive, with an IDE CDROM drive.

so,I repeat, the problem could be with the cdrom drive.. is it slave on primary ide channel?

anyway, search with google for "kernel Emask 0x9 (media error)"

I suppose if it IS a hard disk problem I could just use ghost or clonezilla to image onto a new drive

no.. you should do a backup, change your hd, reinstall and restore.. cloning softwares with SME are a no-sense

[RCFerguson]

Turned out to be the hard drive after all.  I installed a new hard drive and did the UpgradeDisk procedure... during the restore-tape procedure the system hung every time I tried it... manually copied files to a new HD until it hung on a certain file.  Deleted that file and was finally able to complete the restore-tape.