How does disk I/O activity affect network speed?

[filcap]

Hi!

In my company they are experiencing performance problems with some VM that have very high disk I/O activity.

These machines are running on a VMWare ESX 3.5 cluster, with shared storage on a high performance fiber channel SAN (no iSCSI).

They have noticed that during this intensive disk read/write activity the network connection of this VM slows down: the ping latency is much higher (compared to the same parameter when the VM is not accessing the disk).

I know I'm not giving too many details... Actually I don't have direct access to the cluster and I don't have much more information... I'm looking for the relationship between the disk activity and the latency on the ping response. Maybe this could be a starting point for a further investigation about the poor performances of these VMs.

Any idea about that would be greatly appreciated!

Thanks a lot!

Filippo

[Rumple]

One of the things you are going to want to do is look at the perf charts inside the VM...specifically look at the Disk read and Disk write avg Disk queue...if this number goes beyond 1 then you are going to see serious performance issues inside that VM (and even physical machines).

Once the Operating system has delays writing to the disk, everything thats swapped to the pagefile will have delays reading and writing causing a cascade effect which can rtesult in high ping times as the system cannot respond to the requests to acknowledge the packets and send results,etc.

[AndreTheGiant]

Check also storage side data and statistics.

Which kind of RAID do you have? How many LUNs?

How many physical disks? Are all FC at 15k?

Andre