Problems with VMFS Metadata & SCSI Reservations

[razz0raus]

Hi All,

First question I have is about the VMFS metadata & if there is anyway to monitoring it & the copies that each ESX server is on?

The SCSI reservations question comes from a error message we have been getting trying to Vmotion some VM's to another node in our cluster.

It seemed to be too many requests being thrown at our SAN. The SAN is a USP-100 system - I/O on it was actually very small.

The problem in short was SCSI reservation conflicts when we tried to Vmotion.

Just wondering if many have had these problems?

Cheers in advance

[wobbly1]

to help with the second part of your question, ensure that each of your hosts are seeing the luns from your storage on the same IDs

[razz0raus]

Hi mate,

I actually nuttered this one out - thank you for the reply.

The host could see the LUN with out problems - but I think the HBA card is having issues & as such it will be replaced.

Also I found an error with out DR cluster - one of our hosts could see the LUN via VI - could even browse it - but in the linux shell I couldn't change in to the mounted volume - like as if it was not mounted at all. atm I am rebooting this server & warnings saying that esfsck needs to be run (Kinda like a chkdsk) which will run on reboot

Cheers again.

[biekee]

What you also could try from the commandline is the following:

esxcfg-vmhbadevs (=> watch vmhba numbers, might be vmhba1)

then

escfg-rescan vmhba1

and check the result. Not bad to repeat the rescan.

bk