Event 1051 means that there were enough slow reads that the GSAN was thinking about putting the data partition into the suspended state, but decided not to do so. That's what MSG_ERR_UNNECESSARY indicates.
Event 1058 means that GSAN finally put the respective to suspended state. So, looking at the Event, it seems that diskid=0 is now in suspended state.
The "disk 0" is not suspended. It is just a warning message. This Event <1058> is just a warning confirming that the previous event <1051> did not suspend the grid. I didn't mean to say that the disk=0 is actually suspended. Basically, you may ignore these alerts.
I've an internal document which describes this event stating the above.
These events can be ignored, until we see an event <22632> - A server disk has become suspended due to degraded disk throughput. It is important to investigate the reason for the degraded disk throughput, and take corrective action if necessary.
Dis denotes "dispatchers" on the grid. A dispatcher could be any session on the data nodes (backup, cp, hfs, any command that requires a session to data nodes). Ideally, you may see 1 dispatcher on either of the data nodes which denotes the session used by the command "status.dpn" you just ran.
On the other hand, Suspend denotes if the disk on the avamar data nodes is suspended (read-only) for backups or not. A status "false" indicates that the node is ready to perform a backup or any write operation; whereas a status "true" denotes it is not ready for a backup or write operation.
These are two different columns: "Dis" and "Suspend".
The numbers are displayed under "Dis". The value "true" or "false" denotes the status of "Suspend". These two are totally different.
"Dis" shows number of sessions (or dispatchers) running on the respective node. It could be any number including 0 (meaning no session). Please see the above post for details on the sessions on the nodes.
"Suspend" shows if the disk is ready for backups (false) or not (true).
So, the true is bad, false is good, if false =0 means disk is idle, if it is 1, 2 ,3 means number of disks ready or doing something, the disks=logiacl partition like data01, 02....is it right?
arif_ahmad
20 Posts
0
May 5th, 2011 08:00
Event 1051 means that there were enough slow reads that the GSAN was thinking about putting the data partition into the suspended state, but decided not to do so. That's what MSG_ERR_UNNECESSARY indicates.
Event 1058 means that GSAN finally put the respective to suspended state. So, looking at the Event, it seems that diskid=0 is now in suspended state.
Thanks,
Anser Arif
xshzu
16 Posts
0
May 5th, 2011 10:00
Thanks for quick response,
What does the "diskid=0 is suspended" impact grid?
All the status checking of grid shown fine. What do I miss there?
can you explain it further?
arif_ahmad
20 Posts
1
May 5th, 2011 10:00
Hi,
The "disk 0" is not suspended. It is just a warning message. This Event <1058> is just a warning confirming that the previous event <1051> did not suspend the grid. I didn't mean to say that the disk=0 is actually suspended. Basically, you may ignore these alerts.
Thanks,
Anser Arif
rpervan
266 Posts
0
May 5th, 2011 12:00
Ahmed,
Do you have some document which describe such codes 1051, 1068 ... etc ?
thanks,
Rej
arif_ahmad
20 Posts
0
May 5th, 2011 13:00
I've an internal document which describes this event stating the above.
These events can be ignored, until we see an event <22632> - A server disk has become suspended due to degraded disk throughput. It is important to investigate the reason for the degraded disk throughput, and take corrective action if necessary.
Thanks,
Anser A Arif
xshzu
16 Posts
0
May 12th, 2011 13:00
When I do status.dpn the "Dis Suspend 0 false or 1 false" in the grid, is this same thing? can you interpret the column?
thx
arif_ahmad
20 Posts
1
May 12th, 2011 13:00
Dis denotes "dispatchers" on the grid. A dispatcher could be any session on the data nodes (backup, cp, hfs, any command that requires a session to data nodes). Ideally, you may see 1 dispatcher on either of the data nodes which denotes the session used by the command "status.dpn" you just ran.
On the other hand, Suspend denotes if the disk on the avamar data nodes is suspended (read-only) for backups or not. A status "false" indicates that the node is ready to perform a backup or any write operation; whereas a status "true" denotes it is not ready for a backup or write operation.
I hope it makes sense now.
Thanks,
Anser Ahmad Arif
arif_ahmad
20 Posts
0
May 13th, 2011 11:00
Ohh, I guess we are on a wrong track..
These are two different columns: "Dis" and "Suspend".
The numbers are displayed under "Dis". The value "true" or "false" denotes the status of "Suspend". These two are totally different.
"Dis" shows number of sessions (or dispatchers) running on the respective node. It could be any number including 0 (meaning no session). Please see the above post for details on the sessions on the nodes.
"Suspend" shows if the disk is ready for backups (false) or not (true).
I hope this makes sense.
Thanks,
Anser Ahmad Arif
xshzu
16 Posts
0
May 13th, 2011 11:00
So, the true is bad, false is good, if false =0 means disk is idle, if it is 1, 2 ,3 means number of disks ready or doing something, the disks=logiacl partition like data01, 02....is it right?
xshzu
16 Posts
0
May 13th, 2011 12:00
Got it, forgive me working too hard starting to image things.
Thanks again!