StoreEver Tape Storage
1748171 Members
4174 Online
108758 Solutions
New Discussion

MSL5052 + Vertias 9.1 Backup Exec

 
Sean_Streets
Occasional Contributor

MSL5052 + Vertias 9.1 Backup Exec

I have an issue with an installation of a MSL5052 + NSR1200 - 160. Has anyone come across it????

Basically....

They have a MSL5052ST connected to a SAN using a e1200-160 NSR.

2 clustered nodes running W2k SP4 + all MS latest hot fixes all SAN connected running Backup Exec 9.1 + SSO+ Library Expansion Option + latest veritas device drivers.
1 server running W2k SP4 + MS hot fixes and Backup Exec 9.1 + SSO+ Library Expansion packs with master ADAMM database on + veritas device drivers.

All nodes can see the tape library and send jobs to it.

When the other node tries to send a job. Backup Exec reports that the library is offline and has a robotic error. Although the 1st server continues to complete the job and library continues and completes as if there isn't a problem.

It's a pain because it doesn't happen consistently. Which is making it really difficult to pin down and correct.

After trying various things including
Dumping and repairing the ADAMM database.
Different terminator configuration.
Uninstalling Veritas and installing a fresh copies on all nodes.
Different SCSI maps on the NSR.
Disabling the RSM service Win 2K (which should be done as it's running SSO)
Cataloguing old tapes from the existing backup. This worked all ok which made me more puzzled.
MSL library diags. All reports all ok.
Upgrading firmware on Robotic Arm to 4.27 and the SDLT drives ver 5252. All upgraded all ok.
Uninstalling low level drivers using tapeinst.exe. Reinstalling drivers from latest veritas device pack.

Hardware involved is Proliant DL360G3, DL560
NSR 1200-160 internal to MSL 5052 with 2 SDLT drives. SAN architecture is switched fabric. No zoning has been setup and currently there are only only nodes connected to an EVA3000.

Software installed is W2K + SP4 + hotfixes, Veritas 9.1 Backup Exec + hotfixes.
1 REPLY 1
Marino Meloni_1
Honored Contributor

Re: MSL5052 + Vertias 9.1 Backup Exec

It seems there is some kind of communication problem between the two node with BE, I had some similar problem when media management where performed on the library, like adding cartridge or changing cartridge. the next operation reported this error. Veritas suggested to always use the mail slot to manage the media if a mail slot is configured. So try this first.
Second, I do not know if you have just two servers on the SAN or more, but if you have more servers, be sure that all Windows one have the RSM disabled, and the insight agents related to scsi and FC disabled too.

Related to Windows RSM, with windows 2003 there is also a Qarticle explaining that even if you disable the RSM service, some background tasks may still working (like TUR - Test Unit Ready)causing trouble in san environment .