new kernel is not reporting bad blocks

[tincboy]

I've a testing server which is operating on a HDD drive which has bad sectors,
It's was running 2.6.32-11-pve kernel and I could always see the number of bad sectors in /var/log/syslog and easily reallocate them with dd,
But now after updating it to 2.6.32-14-pve there are errors in /var/log/syslog which does not contain the position of bad sectors, like what you can see below:

res 41/40:08:a0:11:76/00:00:53:00:00/00 Emask 0x409 (media error) <F> ata1.00: error: { UNC }sd 0:0:0:0: [sda] Add. Sense: Unrecovered read error - auto reallocate failedres 41/40:08:a0:11:76/00:00:53:00:00/00 Emask 0x409 (media error) <F> ata1.00: error: { UNC }res 41/40:08:a0:11:76/00:00:53:00:00/00 Emask 0x409 (media error) <F> ata1.00: error: { UNC }res 41/40:08:a0:11:76/00:00:53:00:00/00 Emask 0x409 (media error) <F> ata1.00: error: { UNC }res 41/40:08:a0:11:76/00:00:53:00:00/00 Emask 0x409 (media error) <F> ata1.00: error: { UNC }res 41/40:08:a0:11:76/00:00:53:00:00/00 Emask 0x409 (media error) <F> ata1.00: error: { UNC }res 41/40:08:a0:11:76/00:00:53:00:00/00 Emask 0x409 (media error) <F> ata1.00: error: { UNC }sd 0:0:0:0: [sda] Add. Sense: Unrecovered read error - auto reallocate failedres 41/40:08:a0:11:76/00:00:53:00:00/00 Emask 0x409 (media error) <F> ata1.00: error: { UNC }res 41/40:08:a0:11:76/00:00:53:00:00/00 Emask 0x409 (media error) <F> ata1.00: error: { UNC }res 41/40:08:a0:11:76/00:00:53:00:00/00 Emask 0x409 (media error) <F> ata1.00: error: { UNC }res 41/40:08:a0:11:76/00:00:53:00:00/00 Emask 0x409 (media error) <F> ata1.00: error: { UNC }res 41/40:08:a0:11:76/00:00:53:00:00/00 Emask 0x409 (media error) <F> ata1.00: error: { UNC }res 41/40:08:a0:11:76/00:00:53:00:00/00 Emask 0x409 (media error) <F> ata1.00: error: { UNC }sd 0:0:0:0: [sda] Add. Sense: Unrecovered read error - auto reallocate failed

Is it something with new kernel? I really need that numbers to be easily found in syslog, any one has any suggestion on how to ask kernel to report them in syslog?

 

[tincboy]

Any suggestion? Maybe I should reinstall the old kernel ?

 

[tincboy]

On the same server with 2.6.32-12-pve I got errors like

Sep 21 15:34:36 c4 kernel: end_request: I/O error, dev sda, sector 2459552840
Sep 21 15:34:54 c4 kernel: end_request: I/O error, dev sda, sector 2459552840
Sep 21 15:35:12 c4 kernel: end_request: I/O error, dev sda, sector 2459552840
Sep 21 15:35:30 c4 kernel: end_request: I/O error, dev sda, sector 2459552840
Sep 21 15:35:48 c4 kernel: end_request: I/O error, dev sda, sector 2459552840
Sep 21 15:36:06 c4 kernel: end_request: I/O error, dev sda, sector 2459552840
Sep 21 15:36:24 c4 kernel: end_request: I/O error, dev sda, sector 2459552840
Sep 21 15:36:41 c4 kernel: end_request: I/O error, dev sda, sector 2459552840
Sep 21 15:36:59 c4 kernel: end_request: I/O error, dev sda, sector 2459552840

Some one from Proxmox team? Could you please let me know why the newer kernel is not reporting number of bad sector?

 

[grabol]

Probably logging is tunred off? This is just linux... not black magic

 

[tincboy]

Someone knows how to reenable the loggin?

 

[grabol]

set it in syslog config.

 

[tincboy]

I don't think updating Proxmox will case syslog config change, besides I've found no syslog.conf in my Proxmox server

 

[tincboy]

grabol would you please just skip this thread? Thanks.
I expect someone from Proxmox team could truly answer my question about new kernel

 

[dietmar]

Why don't you simply replace the bad disk? (Maybe the driver in new kernel simply does not trigger the error).

 

[tincboy]

I've backup of my data, so I'm not worried about data loss,
But I see this situation on 5 other servers ( amount 14 servers ) I just need to be able to reallocate the bad sector and buy myself time to replace the drives.
Reallocating needs the number of bad block, which was easily reachable in /var/log/syslog with older kernels of Proxmox

 

[tincboy]

Any help on this issue?
How can I ask kernel to report the badsector number?
Any help would be much apperciated

 

[snowman66]

Hi

you can use this command to verify bad blocks:

badblocks -v /dev/sda1

See: http://www.howtogeek.com/howto/37659/the-beginners-guide-to-linux-disk-utilities/