cancel
Showing results for 
Search instead for 
Did you mean: 

Syslog SCSI message meaning

Mauro Gatti
Valued Contributor

Syslog SCSI message meaning

Hi all,
i'm doing some test on SCSI device connected to my server.
When I detach a disk EMS notifies the event to syslog :

Sep 12 17:09:58 myserver EMS [3098]: ------ EMS Event Notification ------ Value: "CRITICAL (5)" for Resource: "/storage/
events/disks/default/0_3_1_0_4_1.2.0" (Threshold: >= " 3") Execute the following command to obtain event details:
/opt/resmon/bin/resdata -R 203030550 -r /storage/events/disks/default/0_3_1_0_4_1.2.0 -n 203030540 -a

that means I detached disk at HW path 0/3/1/0/4/1.2.0

but before this message syslog reports some other entries done by SCSI driver I think (isn't it?)

my server.
When I detach a disk EMS notifies the event to syslog :

Sep 12 17:09:58 myserver EMS [3098]: ------ EMS Event Notification ------ Value: "CRITICAL (5)" for Resource: "/storage/
events/disks/default/0_3_1_0_4_1.2.0" (Threshold: >= " 3") Execute the following command to obtain event details:
/opt/resmon/bin/resdata -R 203030550 -r /storage/events/disks/default/0_3_1_0_4_1.2.0 -n 203030540 -a

that means I detached disk at HW path 0/3/1/0/4/1.2.0

but before this message syslog reports some other entries done by SCSI driver I think (isn't it?)

Sep 12 17:07:47 myserver vmunix: SCSI: Unexpected Disconnect -- lbolt: 54063027, dev: bc051000, io_id: 5cf6050
Sep 12 17:09:50 myserver vmunix: SCSI: Request Timeout; Abort Tag -- lbolt: 54075375, dev: bc054000, io_id: 5cf61f7
Sep 12 17:09:54 myserver vmunix: SCSI: Request Timeout; Abort Tag -- lbolt: 54075788, dev: bc054000, io_id: 5cf61f9
Sep 12 17:09:57 myserver vmunix: SCSI: Request Timeout; Abort Tag -- lbolt: 54076054, dev: bc053000, io_id: 5cf61fc
Sep 12 17:09:58 myserver vmunix: SCSI: Request Timeout; Abort Tag -- lbolt: 54076173, dev: 1f052000, io_id: 5cf61fb

So...
What do lbolt, dev and io_id mean?
If dev stand for "device" how can I know HW path by dev?
Why detachinf only one disk I get timeout errors from other devices?

Thank You

Mauro


Ubi maior, minor cessat!
3 REPLIES
IT_2007
Honored Contributor

Re: Syslog SCSI message meaning

It is not safeway to remove SCSI devices without doing proper method. If the device is already in use then if you detach it then you receive those messages.

Same SCSI controller may have more than one disk and if you detach any one device you receive EMS alert for any disk on the controller. you can check vgdisplay -v which physical disk is missing.

Above EMS alert shows 4 disks showing SCSI lbolt errors.

ll /dev/* |egrep "510000|520000|530000|540000"
RAC_1
Honored Contributor

Re: Syslog SCSI message meaning

It is telling you that an IO operation timed out on a device. That device is 1f052000 (Hex number)
1f - decimal 31 - lsdev -b 31 - disk device.
052000-c0t5d2? Or better ll -d /dev/dsk/* | grep 052000 - this is the device.
There is no substitute to HARDWORK
Mauro Gatti
Valued Contributor

Re: Syslog SCSI message meaning

Ok for the last entry

so for this entry I can say:
lbolt: 54076173, dev: 1f052000, io_id: 5cf61fb

dev 1f052000:
1f is decimal 31 that is disk
052000 is the minor number of special file associated to disk.

But what can I say about other entries?
For instance, let me consider:
lbolt: 54063027, dev: bc051000, io_id: 5cf6050

bc is decimal 188... this time it seems the character device number of disk.
051000 is the same as above.

Why first is there block device number ad after a character devece number?
Could that depend on type of IO operation made using block or character device?

Have you got other info about lbolt and io_id?


Is it correct to think that if I detached a disk into an MSA30 I would get timeout error on all disks on the same SCSI chain until SCSI controller realizes what devices are still attached to the chain?


Thank You

RGDS

Mauro

Ubi maior, minor cessat!