MSL5052 keeps going to "sleep"

[David Frow]

Hello,

We have a dedicated backup SAN with a fibre-attached MSL5052 (using an E1200), and 2GB Brocade-based switches, approx. 20 servers are also attached to the SAN. HBA's are a mix of QLA2310's and Compaq/Emulex cards.

We use Syncsort Backup Express 2.15D, however two or three times a week the library goes into a state of "Initialisation", I get in in the morning, reboot the library via the web interface and all is fine till it happens again.

I have gone through each piece of hardware - Library, drives, E1200, switches and HBA's and they are all at the latest frmware, this hardware has been in approx. 2 years now, before this would only happen maybe once a month but now its much more frequent. Any suggestions?

[Phillip O'Hara]

Can you provide the versions of FW you have in the library, tape drives and router?

[Morne Stevens]

Sorry, I have no solution but have exactly the same problem with my MSL5026 library and is also running a mixture qlogic and emulex cards. Any tips would be greatly appreciated. My library now goes offline every day.

OS: WIndows2003 and Window2000
BE9.1
HBA Drivers: v5-4.82a16
HBA Firmware: v3.92a1
MSL5026SL v4.30
MDR2 v1187

[David Frow]

FYI switch version is 3.1.2, HBA = 392a2, Library = 4.27, Tape Drive = V082/CPQ-2

I found the problem - we had disabled the Fibre Agent from the Insight Manager agents on each server on the fibre, but our Netware servers had CPQFCASA.NLM loaded - also a fibre agent, I unloaded and deleted this NLM from each server and the library ahs not gone to sleep for over a week.

My guess is that the agents stay active even afer the server has finished its backup and released the tape drive, the next server tries to acquire the drive and this causes the library to re-initialise

[Marino Meloni_1]

In SAN environment, the drives can be seen by several servers, and you should prevent the ones that are not allowed to contact the tape drives. These servers can cause problems with polling during backups and can set the drives in "reserved mode" when they are not owned by the right server. At this point the server that have to use it cannot communicate with the device.
In a Windows environment there is some known sw that cause this problem, onw is in fact the FC insight agent with some versions of the Agent, another is RSM, the services of windows, and especialy with W2003, due to a bug of windows, it may start even if the services is disabled in Computer management

Remember that these sw could also be on the same server that the backup application and cause the same problem. Also valid for a direct attached configuration.