MSL2024 Tape Library

[Louann Duchemin]

Good Morning,

We recetnly install this tape library on a ES40's. This library is connect to the SAN switch that goes with the Fibre Channel disk subsystem. I know with the system disk, we had to do a WWIDMGR setup to see the disks, do we need to do the same with these tape drives?

After the site had a downtime, they rebooted and they were not able to see the tape drives. They had to do an IO Find, then an IO Auto.

[EdgarZamora_1]

No you don't do WWIDMGR for the tape drives. The IO FIND and IO AUTO was sufficient.

[Louann Duchemin]

I understand that the IO Find and IO Auto will give will give them access to the tape drive, however, they should not have to do this everytime they reboot the system. I trying to find a solution that will keep the access to the tape drive intact without having to do IO steps.

[EdgarZamora_1]

No it does not have to be done every time you reboot the system. Only once.

The information is stored in SYS$SYSTEM:SYS$DEVICES.DAT and unless you delete that file the information should remain through reboots.

[Louann Duchemin]

Well, it didn't remain after a system reboot. It is also happening with the AS4100 too. These tape drive are not connect directly to the node. They are connected to the SAN switches but fibre.

[EdgarZamora_1]

Are you up to date with your patches, especially FIBRE-SCSI?

Please provide hardware models and OS versions. Can you also attach the file SYS$SYSTEM:SYS$DEVICES.DAT (it's a text file)?

[EdgarZamora_1]

Also check your loop mode on your MSL. It should be set to hard, not soft.

[EdgarZamora_1]

Actually forget that loop mode comment... you are connected to a switch. I don't think that parameter applies in this case.

[Louann Duchemin]

Not sure if my last response made it.

The two types of system effected at ES40's and As4100. They are both operating at OpenVMS 7.3-2 with
DEC AXPVMSVMS732_FIBRE_SCSI V11.0.

Attached is the Sys$devices.dat file. The new tape drives are referenced in the file, but they did not get setup after the reboot.

[EdgarZamora_1]

We're probably not getting the complete story here. When you said the site had a downtime did that include the tape library (being powered off or losing power, etc?)

How may reboots did you do wherein the tape drives didn't show up after the reboot? I have a feeling it's just once (for each system). Have you tested by rebooting again?

My gut feeling is that the id's on tape library changed somehow and so when VMS booted up it didn't configure the tape drives. That's just a guess though. I strongly suggest trying another reboot with logging of the startup (enable this through sysman startup).

[Volker Halle]

Louann,

please have a look at the various SYSMAN IO *_WWID commands related to FC tapes.

Try SYSMAN> IO LIST_WWID to find out, if any FC tape WWIDs on your SAN have not yet been configured by OpenVMS.

Volker.

[Louann Duchemin]

Let me try this again. We installed the new tape drive library. We then configured the libraries into the system by doing the following:
Mc Sysman
sysman> set e/c
sysman> io find_wwid
syman> io auto
sysman> exit
$ show dev mg/full
We noted that 2 devices were created. $2$mga0 and $2$mga1

A few day later the site had to shutdown the nodes. They did not have to power off anything. They then rebooted the nodes, and after the reboot they tried to schedule a tape backup and the tape drives were no longer available. The site then did the Show dev mg/full and the two device, $2$mga0 and $2$mga1 did not list. The site then went through the SYSMAN steps again and the devices were then made available.

They should not have to do the SYSMAN steps everytime they perform a system shutdown and reboot. So, what needs to be done to prevent this from happening?

[Volker Halle]

Louann,

it should work the way you're expecting it to work. But apparently, it hasn't.

Do the systems do a full auto-configure during startup (check SYCONFIG.COM) ? Any error messages during the last startup ?

If the problem happens again, try a SYSMAN IO AUTO first. Does it solve the problem ?

Volker.

[Louann Duchemin]

Attached is a copy of the file, sysconfig.com. it would appear that we are not using the file. what is your take on it?

[EdgarZamora_1]

There is nothing in your SYCONFIG.COM and that's fine.

Please do a SYSMAN IO LIST and make sure there are no WWIDs listed. If that is the case, I strongly feel that your tape drives will show up on your next reboot.

[Volker Halle]

Louann,

I asked you to look into SYCONFIG.COM as this would be a supported method to skip auto-configuration during startup and later manually include SYSMAN IO commands, which could exclude certain type of devices.

Your SYCONFIG.COM does not contain such commands.

Volker.

[Louann Duchemin]

Here is the output from SYSMAN IO LIST:

SYSMAN> io list

%SYSMAN-I-OUTPUT, command execution on node TRIPAA
On port _PGA0:, the following tape WWIDs are not yet configured:

Target 3, LUN 1, HP MSL G3 Series
WWID=02000008:5001-10A0-008A-BC98

Target 8, LUN 1, HP MSL G3 Series
WWID=02000008:5001-10A0-008A-BB98

On port _PGB0:, the following tape WWIDs are not yet configured:

Target 3, LUN 1, HP MSL G3 Series
WWID=02000008:5001-10A0-008A-BC98

Target 8, LUN 1, HP MSL G3 Series
WWID=02000008:5001-10A0-008A-BB98

[EdgarZamora_1]

Well that was unexpected and not good...

I have, however, come up with a brilliantly plausible explanation! You used SYSMAN commands set e/c... io find...io auto...

are your two systems clustered? what account (username) did you use when you did the sysman commands? Was that a privileged account? Did you enable privileges (sysman do set proc/priv=all) before doing the io find and io auto? Please also do a SYSMAN IO LIST on the other system and post the output. I hope this is your problem and not something else.

[Louann Duchemin]

These are clustered systems. The account used for the IO FIND and IO AUTO was SYSTEM and yes this is a priviledge account.

Keep in mind that we are connecting the MSL2024 to a Fibre Channel Switch.

NODE A:
SYSMAN> io list

%SYSMAN-I-OUTPUT, command execution on node TRIPAA
On port _PGA0:, the following tape WWIDs are not yet configured:

Target 3, LUN 1, HP MSL G3 Series
WWID=02000008:5001-10A0-008A-BC98

Target 8, LUN 1, HP MSL G3 Series
WWID=02000008:5001-10A0-008A-BB98

On port _PGB0:, the following tape WWIDs are not yet configured:

Target 3, LUN 1, HP MSL G3 Series
WWID=02000008:5001-10A0-008A-BC98

Target 8, LUN 1, HP MSL G3 Series
WWID=02000008:5001-10A0-008A-BB98

NODE B:
SYSMAN> io list

%SYSMAN-I-OUTPUT, command execution on node TRIPAB
On port _PGB0:, the following tape WWIDs are not yet configured:

Target 3, LUN 1, HP MSL G3 Series
WWID=02000008:5001-10A0-008A-BC98

Target 9, LUN 1, HP MSL G3 Series
WWID=02000008:5001-10A0-008A-BB98

On port _PGA0:, the following tape WWIDs are not yet configured:

Target 3, LUN 1, HP MSL G3 Series
WWID=02000008:5001-10A0-008A-BC98

Target 9, LUN 1, HP MSL G3 Series
WWID=02000008:5001-10A0-008A-BB98

[Louann Duchemin]

Good morning,

Does anyone elase have some suggestion as to why the tape drives are nolonger reachable after a system reboot?

[EdgarZamora_1]

It may be a problem with the configuration of the MSL and not OpenVMS. Can you provide output of SHOW DEVGG/FULL from BOTH systems and also if you can provide MSL drive configuration info (maybe through Commandview MSL or OCP) that would be great.