HPE Community read-only access December 15, 2018
This is a maintenance upgrade. You will be able to read articles and posts, but not post or reply.
Hours:
Dec 15, 4:00 am to 10:00 am UTC
Dec 14, 10:00 pm CST to Dec 15, 4:00 am CST
Dec 14, 8:00 pm PST to Dec 15, 2:00 am PST
Operating System - OpenVMS
cancel
Showing results for 
Search instead for 
Did you mean: 

Replace faulty disk belonging to a shadow set

 
SOLVED
Go to solution
Samshen
Frequent Advisor

Replace faulty disk belonging to a shadow set

Hi all,
I have a faulty disk that belongs to a shadow set in a clustered VMS.
I have to change it online, as the disk is hot plugable.
I have two questions:

1. Where can I see some system logs that report disk as faulty? I have just physically seen the disk has a red light, but I want more information from the system.

2.Is there any special procedure I have to follow or just remove the old one and put the new one and the system will automatically perform the rebuild?

Thank you in advance.
5 REPLIES
John Gillings
Honored Contributor
Solution

Re: Replace faulty disk belonging to a shadow set

Samshen,

>Is there any special procedure I have to follow

General principle. Never reduce a shadowset below 2 members. If you have sufficient disk slots, I'd recommend you start by adding a 3rd member to the shadowset and let it copy. Shadowing will notice any faults on the bad disk, and copy good data from the other member. Once the copy has completed, DISMOUNT the bad disk then physically remove it. Don't forget to change your startup MOUNT commands to reflect the new membership.

> Where can I see some system logs that report disk as faulty?

Does SHOW ERROR show a non-zero error count for the disk? If so, look in the error log. Exactly how depends on your hardware and version (ANALYZE/ERROR, DIAGNOSE, Compaq Analyze, SEA etc...)
A crucible of informative mistakes
Kris Clippeleyr
Honored Contributor

Re: Replace faulty disk belonging to a shadow set

Hi,
Item 1: If the disk is really faulty, the shadow server should have trown it out of the shadowset. OTOH, you can use ANALYZE/ERROR/ELV to see what the drivers, etc. have reported; you can also check the OPERATOR.LOG file.
Item 2: If the disk is still part of the shadowset, you have to dismount it first. E.g.: If you have device DSA1, consisting of $1$DKA0: and $1$DKB0:, and $1$DKB0: is the faulty one, do the following:

$ DISMOUNT $1$DKB0:
physically replace the disk
$ INITIALIZE $1$DKB0: label
$ MOUNT/SYSTEM DSA1:/SHAD=($1$DKA0:,$1$DKB0:) label

Hope this helps,
Kris (aka Qkcl)
I'm gonna hit the highway like a battering ram on a silver-black phantom bike...
Karl Rohwedder
Honored Contributor

Re: Replace faulty disk belonging to a shadow set

You may use the SHOW ERROR command or SHOW DEVICE dsk to get the error count for the disk.
For further analysis there are tools like ANA/ERROR, DIAGNOSE or SEA (the last 2 must be installed separately).

If the disk is still active in the shadowset, you should dismount it with DISMOUNT dsk, e.g. DISMOUNT DKA100. This leaves the shadowset intact, so the users are not affected.
Then you should replace the disk and remount it with
$ MOUNT/SYS DSAnn /SHAD=(dsk) label /confirm

regards Kalle
Samshen
Frequent Advisor

Re: Replace faulty disk belonging to a shadow set

Thank you all for the reply.
I issued the command in the system:

show dev dk

Device Device Error Volume Free Trans Mnt
Name Status Count Label Blocks Count Cnt

$1$DKA11: (SYS2) Online 312
$1$DKA21: (SYS2) ShadowSetMember 0 (member of DSA1:)
$1$DKA22: (SYS2) ShadowSetMember 0 (member of DSA2:)
$1$DKA12: (SYS2) ShadowSetMember 0 (member of DSA2:)

DSA1 has been formed by $1$DKA11 and $1$DKA21. I see the faulty disk $1$DKA11 witch has 312 errors and $1$DKA21 witch is part of DSA1. Label has been APP_DATA.
Should I issue:

$ INITIALIZE $1$DKA11: APP_DATA
$ MOUNT/SYSTEM DSA1:/SHAD=($1$DKA11:,$1$DKA21:) APP_DATA

and the data will be copied from DKA21 to DKA11?

Thank you again.
Karl Rohwedder
Honored Contributor

Re: Replace faulty disk belonging to a shadow set

Samshen,

the INIT is not nec.
You may add the /CONFIRM qualifier. VMS then will show you, in what way the data will be copied, and if is it as expected, you confirm with YES and off the shadow copy goes.

regards Kalle