Hey all,

We're using A.06.11 installed on a 2K3 server as cell manager and talking to a number of devices, including an HP ESL 4-drive library and an EMC array that we use as a virtual tape library. I am having a problem where DP has just screwed up the SCSI IDs of most of the devices. Of course this means that every job scheduled for the devices whose ID changed bombed. Any ideas why?

Hello MB,
I guess I am a little confused regarding the details. Data Protector will not "lose" any paths unless the MMDB has been whacked. SCSI paths routinely change though, so you have options. First, I would set persistent binding on all of your HBAs that provide routes from your Media Servers to your backup targets, both physical tape and virtual devices. The other thing that you can do to prevent your backup jobs from failing, is to enable (checkbox) the "Automatically discover changed SCSI address" option in the tape drive settings tab within the Devices & Media context.

Thanks Mark. As it turned out, I found the problem. A fiber cable connecting one of my SCSI drives in my ESL to the SAN died, so upon rebooting the cell manager, Data Protector moved all of the SCSI ID's of all of the devices on that controller by one and threw every device on the same SCSI controller as the failed drive off by one since one device was now missing. And we can't use persistent bindings because of a bug that causes random unreadable characters to be appended to the paths when the bindings get created. It's all good now.