Disk Arrays
cancel
Showing results for 
Search instead for 
Did you mean: 

LUN SHIFTING IN MODEL20 ARRAY

Manoj P.U.
Advisor

LUN SHIFTING IN MODEL20 ARRAY


Hi all,

I have a Nike array (Model 20) with dual controllers (SPs) with three luns. This are connected to 2 servers which is in cluster
(Only one server access it at a time)
Lun1 and lun2 are in single VG and Lun3 in another VG

A0, B0, C0, D0, E0, A1, B1 = lun1
A2, B2, C2, D2, E2, A3, B3 = lun2
C3, D3, E3, C1, D1 = lun3
E1 is hotspare.

LUN 1 and LUN 2 is bound to SPA and LUN 3 is bound to SPB.

Auto-Trespass is enabled for all LUNS.

Alternate PV is configured for all volume groups for load balancing.

Problem is while doing heavy I/O there will be continuous change in ownership of LUNS 1. At the time I am getting SCSI read errors in syslog.
The above setup was working without any problems for the last 4 years. Suspecting the back plane we replaced it last week. Then there was no problem. It worked fine for last one week without any problem. This week the same problem started. Then we rebooted the Array. After that there was no problem this is the 3 rd day without any problem.
Had anyone faced these types of error? Expecting your views and suggestions for solving this problem.

Cheers

Manoj PU
Manoj P.U.
1 REPLY
Bill McNAMARA_1
Honored Contributor

Re: LUN SHIFTING IN MODEL20 ARRAY

There could be most likely a failing disk in one of the luns.

What you should do is save the UEL for both SPs to file via

mstm > select nike 20 sp
tools (I think in expert or info) > run or log (it's been a while..)

repeat for the other sp and compare the logs.
Note that times can be out between two sps.

You can also get the UEL from entering into "FE" mode on the presentation screen
Crtl P (release all keys)
Shift F (release F) then E (release E and shift)

Then you'll have extra info available in the UEL. You should attach this here or send it to your support engineer.. there are tools that can examine the UEL.

Later,
Bill
It works for me (tm)