Comprehensive data protection for all workloads
Post Reply
GreenEnvy
Enthusiast
Posts: 25
Liked: 5 times
Joined: Jul 31, 2012 3:45 am
Full Name: Lee
Contact:

Help with horrible performance

Post by GreenEnvy »

Hi all,
We have the following setup:
2x Dell Poweredge R610's with 2 Intel Xeon E5620's. 16GB ram
1 Equallogic PS4000 array with 16 SATA drives
HP Procurve 3800 switches
Overland Storage snapserver dx1
ESXi 5.0
Veeam 6.1

Have a data vlan and a storage(iscsi)vlan setup. Data network is 10.3.0.0/22, storage is 192.168.99.64/26
The storage network has iSCSI running on it, and has jumbo frames enabled on everything (on equallogic, switches, esxi vswitch and kernels, and the overland storage unit, mtu is set to 9000). if I ssh into each device and ping with jumbo sized packets, they all pass through without fragmenting.

so the issue is we have a failure on our equallogic. 1 drive totally died, and another is failing. We can still access our data, but the equallogic can't do a rebuild to the spare drives (I don't know the details, equalogic support tells me it is so), it crashes when it gets to a certain spot every time. So we need to reinitialize the equallogic.

The overland storage NAS unit was just going to be used for backups and storing non-critical media, but as it's there, we decided to temporarily move all the VM's from the equallogic to it. That is proceeding ok.

However for one of the servers, we want to restore from a backup. The backup is 1TB provisioned, about 950GB actual size. Veeam is restoring now, but it's looking like it will take 4 days to complete. Windows is reporting nic activity bouncing from 0% to 5%. Oddly, the activity all seems to be on the data vlan's NIC, not the storage VLAN's nic. Any idea on things to check? Veeam backs up to the overland storage unit (which is iSCSI and attached to the veeam server as another hard disk in vsphere), and it's restoring back to that same unit in this case, since the equallogic is having issues. I understand that this will hurt performance, but that seems way too slow.
The backup is a reversed incremental, with the default settings on (compression level).

I don't think the overland unit should be that slow.
Gostev
Chief Product Officer
Posts: 31561
Liked: 6725 times
Joined: Jan 01, 2006 1:01 am
Location: Baar, Switzerland
Contact:

Re: Help with horrible performance

Post by Gostev »

Hi, this sounds like network mode restore (these are very slow because they are done through ESXi, not direct to storage). Do you have a virtual backup proxy capable of restoring VM in hot add mode? This would work much faster. Thanks.
J1mbo
Veteran
Posts: 261
Liked: 29 times
Joined: May 03, 2011 12:51 pm
Full Name: James Pearce
Contact:

Re: Help with horrible performance

Post by J1mbo »

Presumably the backup server is attached to the iSCSI network? So add host file entries for the esx(i) hosts on the backup server, using the IP address of a VMKernel port on the iSCSI LAN.
Post Reply

Who is online

Users browsing this forum: Google [Bot], Semrush [Bot] and 163 guests