Paul,

I doubt that the 5.3.5P2 NDMP problem could account for the incomplete tape eject condition you described.

Symptoms of the 5.3.5P2 problem are that NDMP on the filer deadlocks and stops responding to network requests from the backup application. This condition can be verified by attempting to telnet to the NDMP port (10K) on the filer.

-Greg

-----Original Message----- From: Paul Mikulencak [mailto:paul.mikulencak@amd.com] Sent: Wednesday, May 03, 2000 3:17 PM To: Linn, Greg Cc: 'toasters@mathworks.com' Subject: Re: Data OnTap 5.3.5P2 and NDMP Backups

On Wed, 3 May 2000, Linn, Greg wrote:

A NDMP-based backup problem has been identified in our Data OnTap 5.3.5P2

patch

release.

During development of Data OnTap 5.3.6, a deadlock condition was discovered which impacted the NDMP Java implementation. This condition typically

manifests

itself as an NDMP connection failure during resource depletion conditions. We believe it was first introduced 5.3.5P2. The problem has been resolved in 5.3.5R2P2 and in the forthcoming 5.3.6 release.

If you are running 5.3.5P2, and are experiencing NDMP backup problems similar

to

the connection failure described above, we recommend that you upgrade to 5.3.5R2P2.

greg (and all),

i'd appreciate a pointer to a description of the "problem" in detail.

we're wrestling with something here that i've been assuming was a hardware problem on our storagetek 9740 tape changer, but which hasn't responded to anything i or the storagetek field engineer have tried.

our backups (budtool, 4.6.1, with the jumbo patch, dlt7000 drives scsi attached to each F760 filer) have been failing because fully written tapes are incompletely ejected from their drives.

nothing in the messages file on either the budtool host or the netapps in question. tape drives have been replaced (for one filer, three times), and we even replaced the robotic hand. manual loading and unloading of tapes via the storagetek touch panel work without failure. sysconfig -t always shows the drives it should, and budtool always sees them.

i have a hard time laying this at the feet of the particular revision of data ontap we're running (currently 5.3.5R2P1, formerly 5.3.4R2), but most of the hardware angles have been explored. i've been reading messages on the toasters list that indicates that a bunch of folks are having problems with their backups, but nothing quite like what we're experiencing.

anyone have any feedback?

pablo

email paul.mikulencak@amd.com ext. 602-2448 "the word pessimism bothers me pager 624-0929 because it is often used instead cube 3.277.042 of the word lucidity."

robert bresson