Hi Harold,

You are getting the errors on HBA2 - which is not for any Data access - but for NDMP Backups only - HBA2 is the AUX0 port on the data mover.

Since, the data mover AUX1 port was zoned to the Tape unit one time - it is there in the bind table of the data mover.

Please check whether you have any device created on AUX0 port or not - you may run server_devconfig server_x -l -s -n command to list the non-disk devices already configured.

If not already done, please remove the zones for the Data Mover AUX port (AUX0 and AUX1 - check whether there was any zoning for AUX1 as well).
Then run the devconfig probe and create commands to build the updated device list -
run these commands in the order -

server_devconfig server_x -p -s -n

server_devconfig server_x -c -s -n

This will ensure that no device is configured on the AUX0 or AUX1 ports.

However, the entries for the Tape units in the data mover bind table will be still there - if you want them to get removed from the bind table permanently, please get in touch with the Celerra Support or your EMC CE. However, the entries in the bind table will not cause any issue - they will be in bind pending state and should not generate any error message in the server log. Also, clearing bind table process is a very rare option and EMC support must be consulted/involved.

Hope this helps,
Thanks,
Sandip

Sandip,

Thanks, that's what I needed. I did the -probe and -create, and yes the tape devices still appear when I run the -list. I went ahead and blocked the port on the switch and the logs collected a bunch of FCDMTL 2 [1.4.1] type errors, and one HBA 2 OFFLINE error, but that appears to be one-time trauma. I'm monitoring the logs to make sure the errors don't start flooding again.

I'm assuming (danger!) this is somewhat typical and will allow me to recover the ports. Anything else I need to watch for?

Harold

Sandip,

OK, I spoke too soon. Lots of errors still being logged, like;

008-11-20 09:10:32: DRIVERS: 4: FCDMTL 2 [1.4.1] TPM Hndle API Event: cntx=0x4
33b270, evnt=0x400e, info=0x0
2008-11-20 09:10:32: DRIVERS: 4: FCDMTL 2 [1.4.1] TPM Lup Failure: st=0xa000000
, flg=0x84, info=0x400e
2008-11-20 09:10:32: DRIVERS: 4: FCDMTL 2 [1.4.1] CALLBACK: type 44004, word0 0
000000
*LOG DATA LOST!*
2008-11-20 09:13:43: LIB: 4: 121 log messages have been lost because log buffer
is full.

So, I unblocked the port again. Still want to get rid of these ports, but don't want to flood the logs.

Any additional help would be a good thing.

Thanks,
Harold

Hi Harold,

I think in order to get rid of these messages you really need support to dialin and cleanup the binding tables on the data mover.

regards
Rainer

Sandip and Ranier,

Yeah, I was afraid it wouldn't be easy. I'll open up a call and post back when and if I get a resolution. Thanks for the help.

Harold

Sandip & Rainier,

I posted a support call and got a call back within about 2 hours. Remarkable for a non-critical issue! The support person, Renaldo, dialed in and did his magic and then had to fail over/back to get the errors to clear. Now I'm not flooding the error logs and my ports are unplugged and available.

Thanks for the help.

Harold

Awesome - thanks for the update and great feedback. 3 cheers for Celerra Support !!!! Especially Renaldo.

At the end, it's so pleasant to see everyone happy

Have a great weekend.
Regards,
Sandip