will HSG80 autospare action be reported to VMS?

[Verne Britton]

I am running both WEBES and DECevent (maybe that is not good) ... yesterday I had a disk go bad in a 3 disk RaidSet and autospare kicked in and replaced it.

But I cannot find any notification on my Alpha of this activity ... I thought a year or two ago I did get some communication from the HSG80 to my VMS 8.2 system when another disk went bad.

Any thoughts on how this should work and what could be wrong with my setup?

We have dual controllers in our three ES45s with dual fiber switches and dual HSG80 cards (dual fibers to everything) in each of our four disk cabinets (grand total of eight HSG80s on-line).

Verne

[Volker Halle]

Verne,

both WEBES and DECevent are using OpenVMS error logs to trigger event reporting (either by scanning ERRLOG.SYS or by using the unsupported $DERLMB - declare error log mailbox - system service).

Was there ANY errlog entry on ANY of the systems connected to this HSG80 at this time of that event ? If not, no chance for WEBES or DECevent to notify you.

Are you current on FIBRE_SCSI patches ?

Volker.

[marsh_1]

hi,

have you got an swcc agent installed ?

http://sysdoc.doors.ch/HP/646406.pdf

hth

[Hoff]

You'll usually need two to three error-reporting tools installed on OpenVMS (ELV by default in recent releases, and the DECevent and SEA/WEBES/etc add-on tools) to capture and decode most of the hardware errors, plus controller-specific tools for specific cases and for configuration, plus (potentially) local customizations.

IIRC, the additional software pieces you'll want to look at here are the StorageWorks Command Console (SWCC) software:

http://labs.hoffmanlabs.com/node/564

and/or the StorageWorks Array Controller Software (ACS) (agent) software:

http://h18006.www1.hp.com/products/storageworks/acs/platformsolutionkits/

There may well be some other agent or other monitoring stuff needed here for this controller or for your local environment; I've been more likely to pull out HSG boxes in recent times than to install them. Dig into the HSG manual for details here.

I've encountered cases where the monitoring agents have failed to start for some reason, such as a corrupt configuration. Dig through your local configuration and your startup logs, if you have some of these tools installed.

[Wim Van den Wyngaert]

Every hour or so, we did a set host/scsi to 1 of the disks connected on the hsg80. Then a "show failed" and if you find something, report it in out monitoring software. Also batteries and other stuff were monitored like this.

set ho/scsi needs a freeware image.

Wim

[Hoff]

The SET HOST /SCSI widget and CMDSCRIPT pointers are here:

http://labs.hoffmanlabs.com/node/564
http://labs.hoffmanlabs.com/node/366