Unsolved
This post is more than 5 years old
2 Posts
0
5672
September 26th, 2011 02:00
Syslog Error Message-> kernel: Buffer I/O error on device sda, logical block 0
Hi
The monitoring service detected the following error in /var/log/messages
kernel: Buffer I/O error on device sda, logical block 0
May I know how do I troubleshot this?
ps . Running Avamar 4.0
Thank you.
chiam
No Events found!
rpervan
266 Posts
0
September 26th, 2011 03:00
Hi,
You could start OS system scan for this particular drive with 'chkdsk' command to check consistency on drive.
Regards,
.r
Mark_Bellows
240 Posts
0
September 26th, 2011 06:00
Chiam,
I presume you are receeving this message from the Avamar unit. This is likely a disk error. The logs will state which node and which disk is giving the error message. Take a look at the front of the Avamar grid and see if there are any amber or red lights on any of the drives. If so, there is a drive that needs to be replaced.
I would then recommend calling support and asking them to send someone out to replace the drive.
Since Avamar is closed system with a propriary application on it, running commands, such as chkdsk, is not recommended with out the assistance of support. The gsan would need to be shut down in order to properly protect the data before a chkdsk is run.
In all of the time I have supported Avamar, I have never run a chkdsk - or any other such process - on a grid. The hardware is simpley replaced.
Let us know how you make out.
Mark
Mark_Bellows
240 Posts
0
September 26th, 2011 07:00
By the way, you should consider getting the grid upgraded to the 5.x version - 4.x is not going to be supported for much longer.
tcchiam
2 Posts
0
September 26th, 2011 18:00
Hi Mark
Thanks for your advice.
If there is an amber light in one of the disk, can I just swap out the disk and it will rebuild itself automatically?
Regards
Chiam
bcouhbor
59 Posts
0
September 26th, 2011 21:00
By chkdsk, do you guys really mean fsck, considering it's UFS?
As Mark said, i would avoid running an FSCK, and contact EMC support. Chances are, you have a faulted disk. The disks are hot swappable as far as i know, but the drive needs to be swapped out with an identical drive.
Contact EMC support and get someone on site to perform the disk replacement. I would refrain from doing it yourself.
Mark_Bellows
240 Posts
0
September 27th, 2011 10:00
Chiam,
If the disk has an amber light, it should be changed.
Submit a service request with EMC and put the information about the amber light and which node it is on. They will bring the correct drive out.
Also, be sure to include the serial number of the node that has the failed drive. This can be found by removing the front cover and looking for the label over the little display panel. On rare cases the label is UNDER that display panel, but it there somewhere on it or near it.
Mark