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

error from dmesg output observed on 12H

SOLVED
Go to solution
viks
Advisor

error from dmesg output observed on 12H

hi all ... can somebody tell wat this error means in my dmesg output on node1 of a two node cluster sharing a 12H autoraid box.las ttime i got same sort of errors and a file system use to respond "i/o error"for bdf command.then i did a thorough Preventive Maintenance.this problem didnt repeat after that.but now similar error is there but there are no i/o errors..help pls...
regard
4 REPLIES
A. Clay Stephenson
Acclaimed Contributor
Solution

Re: error from dmesg output observed on 12H

It means than the path was switched from the primary controller X to Y on your 12H. The most common reason for this is using the default IO timeout on the physical disk. Do a pvdisplay /dev/dsk/cXtYdZ and note the IO Timeout value. If it is default (or less than 120 seconds) use pvchange to set it to at least 120 seconds. Somewhere in the range of 120-180 seconds is good. You should set the value on all the LUN's in your 12H.

Man pvdisplay, pvchange for details.

Clay
If it ain't broke, I can fix that.
Bill McNAMARA_1
Honored Contributor

Re: error from dmesg output observed on 12H

you should also get an arraydsp -a output
and logprint output just to verify.

Quickly
arraydsp
will tell you if the thing is in a READY state.

If it is the errors you are seeing are not much to worry about.. unless they start becoming more frequent that is..


Later,
Bill
It works for me (tm)
viks
Advisor

Re: error from dmesg output observed on 12H

but clay can u tell me why my system has gone very slow.i cudnt find the bottleneck.i got a cluster with two Kclass and a shared 12H box.all of a sudden the nis clients strated facing very slow transfer rate of 500KB/s.whereas if they login to their own machine as root and then ftp a file from server it is 1.5MB/s.the home directory of the nis clients lies in the node1.i mean the raid box has two VGs one mounted on node1 and one on node2.if any further input reqd pls do ask..this is critical now..help asap
regard
viks
Advisor

Re: error from dmesg output observed on 12H

hi clay..u tell me one thing.can i transfer the I/O path for a few LUNs thru the second controller.i think it is pvchange with which i can do this.i am not very sure.can u suggest me?bcoz my machine underperforms suddenly and as u told there is starvation for the I/O thruput!!
regard