Drive is down and cannot be upped !!

[oskhan]

hello all

I have run into this scenario and I cannot get my drive status to go from DOWN to UP.

I took the following actions.

I checked the media logs, and realized that one of the full backups supposed to run on saturday were not completed. from that point onwards there were around 15 messages "could not write KEEPALIVE to COMM_SOCK " followed by "no storage units available for use (213)" periodically.

I checked up this error message and found the following troubleshooting document.

http://ftp.support.veritas.com/pub/support/products/NetBackup_Enterprise_Server/277475.pdf

since my master and media server are both the same(v440) I did not have to check for any patch differences

till step 1.3.1 , I did not have to change anything, At this step however I noticed that the drive was DOWN and did not come UP ( either by using vmoprcmd or thro the GUI console)

I verified that BPCD is listening

Is there anything else I can do before calling veritas techsupport?

Also, there is one more thing that i should mention. I am pretty sure that this is the cause of the problem,

this monday I had tried to create a copy of the only backup on tape A00004. using the catalog. The process started, but when I saw that it was trying to mount both A00004 and A00007 tapes at the same time and nothing was happening, I cancelled the job. I realized that since our robot only has one drive, it cannot mount 2 tapes at once.

I never expected that this would cause our backups to stop taking place

please offer any suggestions you think might help me

 

[AquaTeenFryMan]

Kinda of an obvious suggestion, but why not power down and power up the robot, also restart all master and media servers. Again dumb suggestion, but will prolly do the trick.

 

[oskhan]

I had tried that also !

I tried rebooting stoping and restarting the media manager daemons many times. but it did not do anything.

rebooted the system, but still the same problem/

then finally took out all the tapes, and manually pressed the power switch and turned it back on. but still nothing changed

 

[GreyWanderer]

Sounds like a spurious lock file somewhere. Since it's only one drive why not delete it from the config, recycle services, add drive back....just a thought

 

[comtec17]

Can you do a robtest and move tapes in and out of the drive? Once a media is in the drive and do an mt -f /dev/rmt# stat to see if the server can use the drives.

Did you persistantly bind the tape drives to the host?

 

[oskhan]

I did as GreyWanderer suggested, deleted the drive, and added it again. Though it had not worked before, it worked like a charm today.

Anybody have any idea why this was happening?

Comtech, sorry I did not get a dchance to try robtest