Hello!!
First of all sorry for my english language
2 disks in RAID 1 SAS 8TB
2 disks in RAID 1 SATA 4TB
I have mounted 3 vm machines in SAS disks
-veeam vm backup and replication community edition version 9.5.4.2753 in a vm with win 2008 R2 system
-vm1 with win 2008 R2
-vm2 with win 2008 R2
In backup machinne I configured a backup of vm1 and vm2 machinnes with repository disk SATA 6 with 4TB capicity.
And then a couple of words about your setup would also help, f. ex. is it a single server with 4 disks or two different ones, what type of connectivity is involved, etc.?
is slowly to. Please I need a solution for my backup.
I'm open to hear all options.
I have a server with :
2 SAS disk in raid 1 (here is installed ESXi with 3vm: 2 win 2008 with remote desktops and 1 win 2008 with veeam backup)
ethernet with 10gb
usb available only works in 2.0 in win 2008
and setup config for veeam can be more faster than now.
According to the bottleneck statistics, the limiting factor of backup performance is your backup target speed. So if you want faster backups, you should use a different backup target. Thanks!
You should troubleshoot the storage performance issue with your server vendor. Veeam will backup as fast as your backup infrastructure is able to process the data, for example backup speed of 1-2 GB/s is not uncommon at our enterprise customers. Thanks!
I need recommendation for 2 dell servers with ESXi and 3vm per one. How mount this in your opinnion? I need to restore vm in a few minutes. Data per vm 1,5TB.
Network 10gb/s
what is your veeam proxy server specs ? (CPU, RAM)
In our case, we also had a performance issue due too several factors.
beginning = about 13hours to make a full backup (bottleneck displayed in veeam = target)
now = about 3 hours
change =
1. proxy server CPU => coreI3 instead of celeron
2. raid configuration (write cache)
3. aditionnal veeam proxy servers directly on the ESXi host (Active directory servers in our cases => more CPU allocated and RAM. 8x Cpu per VM)
Target 1 = local storage of the Veeam B&R server (physical server)
Taregt 2 = usb3.0 HDD
Target 3 = NAS (lan)
Note : if the proxy used for the job = one of the VMs, then the "bottleneck" displayed by the job report seems ok (CPU/Network/Target % of activity)
if the proxy = "physical server" then, the "bottlenecks" displayed are not "real"
gcutillas wrote: ↑May 17, 2019 8:09 am
I'm in appointed with you, but I'm ussing as target -> SATA 6 in the same server , why is too slowy
Maybe you will try to run the microsoft-tool diskspd to check how fast (or slow) your array performs. If there isn't more performance available then even a proxy reconfig wouldn't help since veeam is already using every bit of performance the repository is providing.
sogapex wrote: ↑May 20, 2019 6:12 am
Note : if the proxy used for the job = one of the VMs, then the "bottleneck" displayed by the job report seems ok (CPU/Network/Target % of activity)
if the proxy = "physical server" then, the "bottlenecks" displayed are not "real"
Bottleneck stats are real in both cases. Seems you had both proxy and repository roles assigned to a single server, that's why improving its specs helped.
In our case, the "bottlenecks" stats were definitevely "broken".
I don't know if we can get an "history" or a log of these stats over time, but before I "messed" with anything, the bottleneck displayed was not the CPU (proxy and repository at this time) whereas it was definitevely it the main bottleneck. (Intel Celeron...)
Backup time (full, usb target) = 8 hours with Celeron, 4 hours with core i3
