tpautoconf -replace_drive doesn't stick

[CBC_Companies]

I replace drives in my SL8500 library. These LTO5 drives are shared with SSO over three or four Solaris media servers.

Historically, I run tpautoconf -report_disc and the drive shows as missing, with a new drive at the bottom, and the missing and new drives have the same path on my media server.

Then, I run tpautoconf -replace_drive and presto, we are all set.

I do this now, ( I replaced three drives on Friday ) and it works! for a while...

After the replace_drive command, when I run the report_disc command, there are no drives missing. Some time later ( and this seems to be random, maybe when it is accessed from a different media server) the drive drops again. It is as if the emm database is not being updated on all servers.

I tried running the report_disc command on all media servers, once the replace_drive command runs, it does not show as missing on another server.

I tried running the replace_drive command anyway on all media servers, it still drops.

The only resolution is to do a hard scan, which requires a NetBackup recycle on each media server.

Anyone else having this issue?

[Marianne]

My experience has always been that NBU Device management service (ltid) needs to be restarted on all media servers.

[Genericus]

thanks, I will try that next time. I have not had to do that in the past - only on one of the four.

Next time I will try it on all four.

thanks!