vgextend io error

[Chad Brindley]

Torsten,

Secure Manager is in use.

I have attached the file, not sure what I am looking at but I assume its LUN number and device world wide name?

Chad

p.s. Learning alot here, really appreciate your time and help on this.

[Torsten.]

Let's pick a line:

NODEWWN 50060b00002fed42 21 WC

This means the server HBA with WWN 50060b00002fed42 is allowed to do WRITE and CONFIGURE on LUN 21.

You need such entry for each server WWN and LUN.





All the DEFAULT lines mean no access for anyone.






So this sets "NO ACCESS by default for all the LUNs", then "overwrites" with "WWN abc is allowed to write to LUN 21", clear enough?

[Chad Brindley]

After looking at the table this looks fine.

I ran an ioscan to find out the HBA device and then an fcmsutil and the LUNS+security is assigned correctly.

[Torsten.]

Well, I can't check this, I don't know your WWN.

IIRC the armtopology command will also help you on this.


The output of "armdsp -a would be nice ...

[Chad Brindley]

Hi Torsten,

Been away for a few days so just coming back to this.

I have attached the armdsp output for you to look through, it looks okay to me but maybe it will help you understand what I have here.

Chad

[Chad Brindley]

Hi,

I finished extending to the alternate paths and did not have an issue again which was strange, maybe it was due to data volume so my question is.

/dev/dsk/c4t2d7
/dev/dsk/c9t2d7 Alternate Link
/dev/dsk/c6t2d7 Alternate Link
/dev/dsk/c11t2d7 Alternate Link

available
Total PE 25596
Free PE 596
Autoswitch On

/dev/dsk/c4t5d1
/dev/dsk/c6t5d1 Alternate Link
/dev/dsk/c11t5d1 Alternate Link
/dev/dsk/c9t5d1 Alternate Link

available
Total PE 25596
Free PE 25596
Autoswitch On

If you notice when I did the vgextend originally I started off through controller 4 for both disks

c4t2d7 and then c4t5d1, is this going to cause an issue, a bottleneck?

If so how can I change that?

Regards,

Chad

[Torsten.]

Hi Chad,

if you want to sort the order of

/dev/dsk/c4t5d1
/dev/dsk/c6t5d1 Alternate Link
/dev/dsk/c11t5d1 Alternate Link
/dev/dsk/c9t5d1 Alternate Link


just do a

# vgreduce ... /dev/dsk/c4t5d1

then

# vgextend ... /dev/dsk/c4t5d1

the result will look like this

/dev/dsk/c6t5d1
/dev/dsk/c11t5d1 Alternate Link
/dev/dsk/c9t5d1 Alternate Link
/dev/dsk/c4t5d1 Alternate Link


Sort them until the primary link is using the other FC HBA.

[Chad Brindley]

Thanks Torsten, most helpful. This did the trick.

I will assign points soon as this thread has really assisted me to complete this task.

Strange how the IO error dissapeared, I think it must have been an issue on those days. The volume group hosts an oracle database and maybe it was just extremely busy on those days causing it to timeout.