There are many reasons why your backups may be slow, and while they are running causing performance problems but I would in the first instance get a good undersatanding of how your VMware storage and backups are configured. There are many articles available online about how best to set up Backup Exec 2012 to back up VMware (including applications/databases within VMs) such as:
BE2012 - Design/strategy - VMware based with GRT OR regular backups with periodic VMware backups | Symantec Connect Community
As you have recently procured the solution I would recommend working with the supplier in the first instance to resolve - as this appears to be a backup issue I would hope they prioritise the issue for you. To ensure that we are not recommending you look at the wrong thing I would suggest testing performance/pings, etc when the backup is not running - I would expect the results to be significantly different if the backups are the cause.
You'd be surprised, the exchange services have a higher priority than ping responses, so they can blat everything else out of the picture.
Originally Posted by DMcCoy
Anyway, I'd agree it looks like a network config issue - the CPUs could still do with a life to recommended levels anyway.
You've two nics on different vlans it appears - which is on your network, and how are the nics setup in windows?
You're right we do seem to have 2 NIC's on different VLAN's. Im guessing because we are part of the academy's forest aswell. This is whats in windows
Attachment 15428Attachment 15429Attachment 15430
I'd speak to the company that set it up and ask if there's any reason for not using the vmxnet3 card type - if not, ensure the vmware tools are up to date and change the network adapters for VMXNET3 ones.
are their other vms on the host showing similar symptoms? might be worth checking their isn't a speed/duplex mismatch between the host and the switch also - or that one isn't hardcoded to a slower speed.
But also as ergo said, trying it without the backup running is also worthwhile.
Where are the actual physical hosts and virtual hosts in relation to each other? If the backup is using all of a 10Mb WAN link could explain it.
Backup was stopped and exchange restarted.... suprise suprise back to <1ms pings and outlook working perfectly. I have a feeling as soon as backup runs again its going to kill it. They did allocate 16GB to exchange too.
What should the network card speed be set to? Can we just add another network card VMXNET3 copy settings across (ip, subnet etc) and remove the E1000 cards?
Yeah, the card switch should be pretty easy, as above.
Originally Posted by Darylrese
The card speed should be autonegotiate - unless there's a pressing reason for it not to be
Backup rerun, only took an hour this time and exchange still working as it should be. Network useage now virtually nothing compared to maxed out as before.
Will look into adding another 2 vCPU's and the VMXNET3 cards to help speed things up in future. Ping is <1ms or =1ms now which is fine and roll up 7 for service pack 1 has been installed! :)
Cheers for your help guys!
Glad you got it sorted :)
Originally Posted by Darylrese