Mirrorset running at reduced mode

[Jorge Cocomess]

My mirrorset (2 x146Gb) currently running at the reduced mode. I have the spare drive and I would like to replace it. Would you please help me with the CLI command for the HSG 80 controllers.

Thank you in advance.

Jorge

[Ian Miller.]

If there is no replacement policy for the mirrorset then

SET mirrorsetname REPLACE=diskname

where diskname is the spare disk you want to add.

[Ian Miller.]

the HSOF V7.3 CLI Ref guide is at
http://webserver-1.wescorp.com/wescorp-library/pdf-library/DEC-CPQ-HP/MANUALS-7/EK-CLI70-RM-B01.pdf

[Andy Bustamante · Accepted solution]

The HSG 80 will rebuild your mirrorset if a disk with enough capacity is available in the spare set.

The mirrorset needs to have a replacement policy set. Assuming that's done this is an example:

SHOW FAIL ! show the fail set
DEL FAIL DISK20400 ! remove the failed disk
ADD SPARE DISK20400 ! designate a hot spare

As a rule, I prefer to have disks always in the spare set. The controller(s) will grab a replacement disk in the event of a problem.

Andy

[Anton van Ruitenbeek]

Jorge,

First you have to delete de failset and than you can add the replaced disk (can be in the same slot) to the mirrorset again.

@Andy,

As a rule I never put disks in a spare set.
This because I use only JBODs and love to use OpenVMS HBVS. There is only one limitation to HBVS. This is there may only be max. 3 drives in one shadowset. I have a disaster tolerant system, two locations and want to have on both locations 2 drives in one shadowset. Problem! I solved this by two JBODs on the location where the tapeunit is (because of the backup and reducing shadowset) and a mirrorset on the other location. This mirrorset I monitor every hour using 'SET HOST/SCSI' (actualy I monitor all the four HSZ controller every hour using this tool) and if there is something wrong I mail it to the operators/managers. By not using sparesets I did label all the disks in the cabinet and I always know where to find my disks. If there is a defective drive I can find the disk directly (I don't need a tool or show commands to find it).

AvR

[Wim Van den Wyngaert]

We use a mirrorset on both sides and shadowing between the 2 buildings.
You never know which site will get lost and Anton will lose data when there is 1 disk failure of the JBODs that is left.
When performance is not important we use a raid-5 that is shadowed to another raid-5 (to reduce # disks).

Wim