cancel
Showing results for 
Search instead for 
Did you mean: 

Mirrored stripesets

admin1979
Super Advisor

Mirrored stripesets


In one of the TRU64 systems , we have HSZ80 RAID controller connected. One of the shares has giving advfs domain panic.

The share name is data9.

# scu show edt

CAM Equipment Device Table (EDT) Information:

Device: BD009635C3 Bus: 0, Target: 0, Lun: 0, Type: Direct Access
Device: BB00913466 Bus: 0, Target: 2, Lun: 0, Type: Direct Access
Device: TLZ10 Bus: 0, Target: 4, Lun: 0, Type: Sequential Access
Device: BD0366349C Bus: 1, Target: 1, Lun: 0, Type: Direct Access
Device: BD0366459B Bus: 1, Target: 3, Lun: 0, Type: Direct Access
Device: BD01864552 Bus: 1, Target: 5, Lun: 0, Type: Direct Access
Device: SuperDLT1 Bus: 2, Target: 5, Lun: 0, Type: Sequential Access

Device: HSZ80CCL Bus: 3, Target: 0, Lun: 0, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 1, Lun: 0, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 1, Lun: 1, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 1, Lun: 2, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 1, Lun: 3, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 1, Lun: 4, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 1, Lun: 5, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 1, Lun: 6, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 1, Lun: 7, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 2, Lun: 2, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 2, Lun: 3, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 2, Lun: 4, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 2, Lun: 5, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 2, Lun: 7, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 3, Lun: 0, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 3, Lun: 2, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 3, Lun: 3, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 3, Lun: 4, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 3, Lun: 5, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 3, Lun: 6, Type: Direct Access
Device: HSZ80 Bus: 3, Target: 3, Lun: 7, Type: Direct Access

Device: DEC WSE BSD/V1 Bus: 3, Target: 7, Lun: 7, Type: Processor
Device: CRD-8402B Bus: 4, Target: 0, Lun: 0, Type: Read-Only Direct Access

# showfdmn data_9

showfdmn: unable to get info for domain 'data_9'
showfdmn: error = I/O error




So here is the actual relevant configuration :

Volume Name for which Domain panic occurrs: /dev/rz25c

Domain#Fileset: data_9#data

Mounted on: /database/dump

When connected to the controller , here is the relevant part :-

hsz80>show storagesets

STRIPE3 stripeset MIRROR13 D100
MIRROR14
MIRROR15
MIRROR16

And,

hsz80>show mirrorsets


MIRROR13 mirrorset DISK30800 STRIPE3
DISK40900

MIRROR14 mirrorset DISK50900 STRIPE3
DISK60800

MIRROR15 mirrorset DISK10900 STRIPE3
DISK20800

MIRROR16 mirrorset DISK50800 STRIPE3
DISK60900


As can be seen , the stripset3 is consist of 4 mirrorsets 13,14,15 & 16 and each containing 2 mirrored disks.
Now it looks like 1 of these 8 disks has gone bad or going bad. Please note that there is no listing in the output
of "show failedsets".

Now the question is which of the disk out of 8 is gone bad and how do we avoid the data loss? And how to replace the faulty one by breaking the mirror or stripeset?

Please suggest.
6 REPLIES
Pieter 't Hart
Honored Contributor

Re: Mirrored stripesets

if it's Tru64 V5 up
try "sysman event_viewer"

to see the eventlog and to determine what was the cause of the doamin panic.

instead of "scu show edt"
try "hwmgr -show scsi"
and at the controller "show units"

at the HSZ try
"show this" (this_controller)
or "show other"
maybe not a drive but one controller of a redundant pair has failed?

admin1979
Super Advisor

Re: Mirrored stripesets

Hello,

Acutally its TRU64 V4.0G.
Please find the relevant logs in attachment and suggest.


Thanx,
Admin
Johan Brusche
Honored Contributor

Re: Mirrored stripesets


Hello,

Is this a standalone Tru64 4.0G or is it in an ASE-cluster ?

What's in /var/adm/messages at the 1st time the error occured?

Did you look into binary.errlog with DECevent or dia ?

Try scu..

scu> sbtl b 3 t 1 l 0
scu> tur
scu> show inq

Any results from
# /sbin/advfs/verify -f -d data_9
or
# /sbin/advfs/fixfdmn data_9

__Rgds, Johan.

_JB_
admin1979
Super Advisor

Re: Mirrored stripesets

Thanx for the reply.

From the previous attachment , it is clear that its a clustered environment.

Moreover please find the attachment for the completed messages & binary.errlog logs.

It is a production env so just trying to be very sure before I try anything on it.

Please let me know your views on this.

Thanx again.

Admin
admin1979
Super Advisor

Re: Mirrored stripesets

Missed the binary logs in previous post.
Here is the binary error log.
Johan Brusche
Honored Contributor

Re: Mirrored stripesets


I do not have WEBES available here so cannot look into the binary.errlog .

Is rz25 part of an ASE disk service ?
If not, how do you make sure the other node in the cluster cannot grab the disk ?

Rgds,
Johan.

_JB_