Devices in "service" status 1

[haasemi]

Hi,

we have Legato 6.1.3 on W2K.

Since the new version, sometimes one ore more tape devices in the autochanger are going on status "service" (it stand in the field enabled). In that status the devices are not available for backups.

I searched in the documentation, but I didn´t find anything about this status of devices.

Can anyone explain this?

Thanks!

 

[disldrautoldmc]

Service mode is different than the Disabled attribute because the nsrmmd process is not stopped.
No more sessions will be assigned to the device once it is placed in Service Mode.
So, if the device is disabled the NSRMMD process is stopped.
If the device is disabled, you cannot perform any of these;
If the value is set to Service, the device is temporarily disabled for maintenance purposes.
The only possible actions with enable set to Service are to unload the drive, inventory or label volumes.

If the nsrmmd is running the Device will no take more sessions but will finished those accepted already, so it not abort them.
If it is disabled, it not take more sessions, nor will it finished those accepted already, it will abort them.

So the running or not running nsrmmd is the only decision criterion to disable the device or set it to service mode.


1. When a drive reaches the max consecutive errors for whatever reason the device goes into Service Mode?
When a device goes into service mode, this allows nsrmmd to complete what it is doing and eject the volume back into its slot. After the volume has been ejected, "nsrd" will not assign any nsrmmd to that device from then on.
The reason behind the service mode was to avoid volumes getting stuck in the device when they reach the max consecutive errors. Prior to 6.1.3, if this happen, the backup session will hang because nsrd will continue to request that volume.

The reason is that prior to 6.1.3/7.0 when the device reached the max consecutive error count with active sessions and the volume inside NetWorker would disable the device. This was undesireable. With 6.1.3/7.0 the automatic transition to disabled is not allowed while an operation is in progress (running device) even if the max consecutive error count reached. The device will transition from enabled to service mode and all the current active sessions will be completed.


2. What can the user do in Service Mode with the drive that he can't do in disabled mode?
This is documented in nsr_device man page (under enabled section). Here is an excerpt that should answer this question:
If the value is set to Service, the device may not be mounted for save or recover operations. This state is used to reserve
a device for maintenance. The device can be used for administrative purposes such as volume verification, labeling or inventories if the
device is selected using the -f option. A device set in Service mode can not be used for save or recover operations. If the device is set to
Service while the device is in use, all current sessions will be allowed to complete normally, but no new sessions will be assigned to the
device. If the device is a jukebox device, the device will be unloaded after the sessions have completed.


3. If placing the drives into service mode they are still able to complete the sessions currently running. Does this mean the drive really isn't having that serious a problem that it can still accept and write streams?
You need to figure out the cause of "max consecutive error" count being reached. Look in system logs for clues like /var/adm/messages on Solaris


4. What event triggered placing the drive into service mode.?
The only 2 events are
* User manually changing the mode to service on the drive
* Max consecutive error count being reached and rather than disabling the device transition to service mode and let current sessions complete