Re: [Veritas-bu] IBM LTO-5 (8.0gbps FC) / QLogic 2562-CK (8gbps HBA) tim

Hi,

OS; RHEL 5.3 64-bit

Tape Library (in all cases): SL500 w/FW 1432

LTO-5, I’ve tried them all that are available from Oracle (4), there is an issue with encryption on these drives in our setup, currently on BBN2.

Due to the encryption bug (details below) had to stop using it for now, but we’re seeing drives timing out as noted below with no encryption.

I’ve tried A9Q5, B5BF, B6W2 and now on BBN2 (latest)

Issue w/ encryption btw incase anyone is curious:

It appears the IBM LTO-5 drive is advertising BOTH 12 and 60 bytes for the Maximum AUTH Key-associated data?
But the IBM LTO-4 drive only advertises 12 bytes?

$ grep 'Max AUTH Key-associtated data' lto4
Max AUTH Key-associtated data 12 (bytes)
$ grep 'Max AUTH Key-associtated data' lto5
Max AUTH Key-associtated data 12 (bytes)
Max AUTH Key-associtated data 60 (bytes)

You can get the output via:

/usr/openv/volmgr/bin/scsi_command -d /dev/nst0 -spi

From Oracle:

The IBM LTO5 scsi reference manual does show that Maximum Authenticated Key-Associated Data (A-KAD) Bytes
(000Ch) = 12 bytes

Justin.

From: Len Boyle [mailto:Len.Boyle AT sas DOT com]
Sent: Tuesday, April 10, 2012 8:19 AM
To: Justin Piszcz; veritas-bu AT mailman.eng.auburn DOT edu
Subject: RE: [Veritas-bu] IBM LTO-5 (8.0gbps FC) / QLogic 2562-CK (8gbps HBA) timeout question

Justin

What os are you using.

Which tape library?

Which firmware level are you using in the lto-5.

From: veritas-bu-bounces AT mailman.eng.auburn DOT edu [mailto:veritas-bu-bounces AT mailman.eng.auburn DOT edu] On Behalf Of Justin Piszcz
Sent: Tuesday, April 10, 2012 8:16 AM
To: veritas-bu AT mailman.eng.auburn DOT edu
Subject: [Veritas-bu] IBM LTO-5 (8.0gbps FC) / QLogic 2562-CK (8gbps HBA) timeout question

Hi,

Was curious if anyone had been running into timeouts with IBM LTO-5 drives in heavily utilized environments with QLOGIC 2562-CK (8GBPS) HBAs?

Never saw this in smaller environments with MPX <= 3 but now with MPX >= 6 some drives seem to be timing out and going into a “hung” state.

Was curious if anyone ever ran into this issue before?

Power cycling the drive (reseating it) fixes it and then it’s fine again as a workaround but not a fix, thoughts?

F/W on the IBM LTO-5 drives is BBN2 (latest from Oracle)

F/W on the HBA’s is 3.00 (latest from QLogic)

When the problem occurs (these errors spew continuously) until the drive is reseated (rebooting the robot does not clear out the errors)

st 3:0:0:0: timing out command, waited 7s

qla2xxx 0000:0a:00.1: scsi(3:0:0): Abort command issued -- 1 e08 2002.

st 3:0:0:0: timing out command, waited 7s

qla2xxx 0000:0a:00.1: scsi(3:0:0): Abort command issued -- 1 e0a 2002.

st 3:0:0:0: timing out command, waited 7s

qla2xxx 0000:0d:00.0: scsi(0:0:0): Abort command issued -- 1 1820 2002.

qla2xxx 0000:0d:00.0: scsi(0:0:0): Abort command issued -- 1 1821 2002.

st 0:0:0:0: timing out command, waited 900s

st0: Error 6080000 (sugg. bt 0x0, driver bt 0x6, host bt 0x8).

st 0:0:0:0: timing out command, waited 180s

qla2xxx 0000:0d:00.0: scsi(0:0:0): Abort command issued -- 1 1823 2002.

st 0:0:0:0: timing out command, waited 60s

qla2xxx 0000:0a:00.1: scsi(3:0:0): Abort command issued -- 1 e18 2002.

st 3:0:0:0: timing out command, waited 7s

qla2xxx 0000:0a:00.1: scsi(3:0:0): Abort command issued -- 1 e1a 2002.

st 3:0:0:0: timing out command, waited 7s

Justin.

_______________________________________________
Veritas-bu maillist  -  Veritas-bu AT mailman.eng.auburn DOT edu
http://mailman.eng.auburn.edu/mailman/listinfo/veritas-bu

Re: [Veritas-bu] IBM LTO-5 (8.0gbps FC) / QLogic 2562-CK (8gbps HBA) timeout question