Also, check your mount retention time for the device class. I seem to remember that being an issue during large restores with requested data residing on multiple tapes. For some reason I recall the restore operation only using one tape drive for the restore, even though there were other drives available to mount tapes on. I would think that if the restore required 3 mounts it would mount all three tapes on available drives but that did not seem to be the case.
I updated my mountretention to 0 and as soon as it was done with that tape, it unloaded it and immediately mounted the next one. Remember to put it back when you're done. This is probably a hack so if someone has any input I'd be interested to hear it. I think the only way to get multiple mounts would be to multi thread your restore. Regards, Denis L. L'Huiller 973-360-7739 [EMAIL PROTECTED] Enterprise Storage Forms -> http://admpwb01/misc/misc/storage_forms_main.html Robin Sharpe <Robin_Sharpe@B To: [EMAIL PROTECTED] ERLEX.COM> cc: Sent by: "ADSM: Subject: Re: "Think time" between switching tapes during restore Dist Stor Manager" <[EMAIL PROTECTED] ST.EDU> 01/15/2002 11:39 AM Please respond to "ADSM: Dist Stor Manager" I would guess the server is figuring out what tape to mount next. Can you measure server cpu% and i/o during that time? Robin Sharpe Berlex Labs Jeff Connor <connorj@NIAGARAM OHAWK.COM> To: [EMAIL PROTECTED] cc: (bcc: Robin Sharpe/WA/USR/SHG) 01/15/02 11:05 AM Subject: Please respond to "Think time" between switching tapes during restore "ADSM: Dist Stor Manager" Does anyone know what TSM is doing just prior to switching tapes during a restore? First our environment: - TSM for AIX V4.1.4.0 - AIX 4.3.3 on an H80 - Client being restored is a Windows 4.0 server sp6a two 700MHz xeon with 512MB RAM - TSM Windows client 4.2.1.15 - 100Mbit Ethernet nic and switch locked at 100Mbit full duplex - 3494 with 3590 Fibre Channel attached drives using K tapes. The specifics of what we are seeing is: 1. Nothing else is running on the TSM server except our restore session. 2. The DSMC process, viewed through task manager on the Windows client shows 0 to 1% CPU and system idle 99%+ 3. Q SESS from the admin CLI shows the restore session in a RUN state with send and recv byte counts not moving. 4. After the session sits in a RUN state for up to 5 minutes a state change for the session to MEDIAW occurs. A new tape is mounted in a minute or less and the session continues. So my question is, what is going on during that time between tapes when the session appears to halt but says it's running? Anyone else seen this behavior? It does not appear to be a client or server resource shortage of any kind. Thanks Jeff Connor Niagara Mohawk Power Corp
