LVM message in "dmesg"

[CCIL]

i hv a N class server clusterd environment the "dmesg" shows LVM messages
entire cluster setup is still stable

[Marco Santerre]

Chances are you have several fibers attached to your disks, and in this case, the disk that seems to have trouble, has switched to its PV link.

I'm also noticing that you have a file system that's full.

[James R. Ferguson]

Hi:

You have had a couple of instances of loss of a pvlink to a device and a point where vg00/lvol6 was full when a write was attempted. Neither of these situations are particularly serious unless they occur persistently.

Regards!

...JRF...

[CCIL]

I have an alternate link to every San FC disk

root@cciln2:/>strings /etc/lvmtab
/dev/vg00
<3p1
/dev/dsk/c1t6d0
/dev/dsk/c2t6d0
/dev/vg01
/dev/dsk/c6t0d1
/dev/dsk/c6t0d2
/dev/dsk/c4t0d1
/dev/dsk/c4t0d2
/dev/vg02
/dev/dsk/c6t0d3
/dev/dsk/c4t0d3
/dev/vg03
/dev/dsk/c4t1d0
/dev/dsk/c6t1d0

every disk has two paths

[Marco Santerre]

Yes, and you probably lost one PV link.. probably not even permanently, just temporarily. But that is the reason you are getting those messages.

You can always look, using ioscan to see if you see all your disks., but I'm pretty sure it recovered itself.

[Sean OB_1]

It looks like your fiber channel had a transient error and the system switched to the alternate path to the disk.

I wouldn't worry about it too much unless you start seeing lots of this type of error, then you may need to replace the fiber adapter.

[Eugeny Brychkov]

It looks like Vritual Array connected lumped into one FC private loop. Check if these failures are generated for one FC adapter or for both. If for both then maybe something global is wrong in the loop, if only for one then maybe something wrong with FC HBA.
BTW, did you set all devices in loop correctly - disk array, library etc?
Eugeny