Disk errors while creating database

[csreenivas]

We are getting the below disk errors when our DBA team trying to create database. Can anyone help me in resolving the below errors.

/home/sybase/dba/> 00:00000:00009:2008/11/24 16:25:20.97 kernel Cannot send, host process disconnecte
d: 3976 spid: 9
01:00000:00000:2008/11/24 16:33:13.31 kernel sddone: write error on virtual disk 7 block 59064:
01:00000:00000:2008/11/24 16:33:13.31 kernel sddone: 16384 bytes passed, -713603864 returned on write for virt disk
7 block 59064
01:00000:00001:2008/11/24 16:33:13.31 server Error: 3225, Severity: 21, State: 1
01:00000:00001:2008/11/24 16:33:13.31 server I/O failed while attempting to clear pages starting at page 25600 of disk number 7.
01:00000:00001:2008/11/24 16:33:13.32 kernel

[Suraj K Sankari]

Hi,

Did you checked your syslog.log ?
Please post last 60 lines of /var/adm/syslog/syslog.log

Suraj

[unixguy_1]

Hi Csreenivas,

Can you check the anyother HDD is Problem Or not?

i thought that disk no 7 is a problem.

Pls Kindly check and revert back.

Regards,
Unixguy.

[unixguy_1]

Hi Csreenivas,

what happen?

Is it resolved?

Regards,
Unixguy.

[csreenivas]

Hi

I got below errors in syslog.log file

------ EMS Event Notification ------ Value: "CRITICAL (5)" for Resource: "/storage/events/disks/default/10_4_12.4.0" (Threshold: >= " 3") Execute the following command to obtain event details: /opt/resmon/bin/resdata -R 188219434 -r /storage/events/disks/default/10_4_12.4.0 -n 188219395 -a

Looks like disk has been failed but we didnt see any stale errors, we asked the DBA team to work on other disk where we have already created logical volumes.

[Johnson Punniyalingam]

>>>>I got below errors in syslog.log file

------ EMS Event Notification ------ Value: "CRITICAL (5)" for Resource: "/storage/events/disks/default/10_4_12.4.0" (Threshold: >= " 3") Execute the following command to obtain event details: /opt/resmon/bin/resdata -R 188219434 -r /storage/events/disks/default/10_4_12.4.0 -n 188219395 -a<<<<<<<

It says Execute the following command to obtain event details

did you ..? execute .?

# /opt/resmon/bin/resdata -R 188219434 -r /storage/events/disks/default/10_4_12.4.0 -n 188219395 -a

execute the above command and It will show event details--> please look for Recommended Action,

if you feel it refer's to Disk, please verify disk with dd command

# dd if=/dev/dsk/cxtxdx of=/dev/null bs=128k
560057+1 records in
560057+1 records out
(output should show some thing like above, also the output will take while depends on the size of you disk)

Thanks,
Johnson

[csreenivas]

Yes I haev executed and observed the below errors

Disk at hardware path 10/4/12.4.0 : Media failure

Associated OS error log entry id(s):
0x492a1ff300000000

Probable Cause / Recommended Action:
Reformatting the medium may fix the problem.
Alternatively, the medium in the device is flawed. If the medium is removable, replace the medium with a fresh one.


So we informed DBA to use other logical volumes.

Can anyone help me how to format the disk?

[Johnson Punniyalingam]

Disk at hardware path 10/4/12.4.0 : Media failure

ioscan -fnH 10/4/12.4.0 --. post the output
(above command will show you the disk path)

ioscan -fnC disk --> output need to advice more,

Thanks,
Johnson

[ARCHITECT_1]

hi

Plz check the /var/adm/syslog/syslog.log and post the output of syslog and

#ioscan -fnC disk

Also check the disk using the dd command


Thanks

[csreenivas]

I have checked already using ioscan, the disk is showing "claimed" and there are no stale errors while checked with lvdisplay command.
Using vpdisplay I gathered all the lvs on that paricular disk, and informed DBA not to create database on that lvs.

[Suraj K Sankari]

Hi Again,

If its a hot-swapable disk then unmount the filesystems and take out the disk from the slot and wait for 40 to 50 seconds then again put it into the slot.

Then mount it again and check

Some time its works

Suraj

[unixguy_1]

HI Csreenivas,

Suraj is Correct, u do like that,conformed some times it's working.

Regards,
Unixguy.

[Johnson Punniyalingam]

Suraj & Unixguy,

early post shows that disk claimed in ioscan
our suspect should be checking the "Condition" of the disk by using dd command,

if the dd command's record+in records+out return's 0, than suspected could be either the bad controller / bad disk,

Thanks,
Johnson

[Ganesan R]

Hi Sreenivas,

Since it is a critical EMS event, replace the disk as early as possible. Don't rely on disks which is intermittenly failing. Otherwise you may loose the data.

Disk may be looking good at this point of time and may fail at any time because modern days disks are so sensitive. Since you are going go create database on it, it is recommanded to engage the hardware vendor to replace the disk.