Warm migration - abort and monitor
-
Hi!
I did just try your new "warm migration"-feature, but there are some questions...
- Is there any possibility to abort and monitor the migration task?
- What is the expected migration speed? I was not able to push more than 100 Mbit/s
Thank you and best wishes
KPS -
Hello,
Until we got "XO tasks" (it's a work in great progress!) it's not possible to get an "overall" monitoring on the progress.
However, you can check the progress on each XCP-ng individual tasks in the XO "Tasks" view. What takes most of the time is the initial replication.
Regarding the speed, remember it can be limited by the Dom0 VHD export speed capability. What's the hardware do you have both on source and destination? Also, the traffic has to flow to your XO, so depending where it is in your network, it can be also a bottleneck.
-
@olivierlambert
The cluster is a mid-performing-system with Xeon Gold 6154 CPUs, 192 GB memory and 10 GbE NICs. XOA is running on the same host, as the source VM. There is nearly 0 load on the systems. Did you see speeds which are far above 100 MBit/s -
I could get more than this on my setup yes. What about the destination host?
-
@olivierlambert
The destination host has the same specs. They are connected with 10 GbE on three paths. Local storage are Micron 9300 SSDs.Edit: That is the same speed, I can achieve with StorageMigration
-
That is the same speed, I can achieve with StorageMigration
Then XO isn't the bottleneck It can be anything like Dom0 usage (CPU/memory), SR usage/speed, max CPU freq/efficiency and so on.
-
@olivierlambert
My problem is, that I cannot see the bottleneck....
tapdisk on source an destination is about 8/17%
System CPU is <5%
Top shows I/O-wait at <2%
Network is only loaded at 135 Mbit/s peak, but if I transfer data manually, I am getting about 10 Gbit/s:[ ID] Interval Transfer Bitrate [ 5] 0.00-4.00 sec 450 MBytes 944 Mbits/sec receiver [ 8] 0.00-4.00 sec 1.12 GBytes 2.40 Gbits/sec receiver [ 10] 0.00-4.00 sec 1.48 MBytes 3.10 Mbits/sec receiver [ 12] 0.00-4.00 sec 3.23 MBytes 6.78 Mbits/sec receiver [ 14] 0.00-4.00 sec 2.85 MBytes 5.98 Mbits/sec receiver [ 16] 0.00-4.00 sec 3.37 MBytes 7.08 Mbits/sec receiver [ 18] 0.00-4.00 sec 2.30 MBytes 4.82 Mbits/sec receiver [ 20] 0.00-4.00 sec 410 MBytes 861 Mbits/sec receiver [ 22] 0.00-4.00 sec 2.18 MBytes 4.58 Mbits/sec receiver [ 24] 0.00-4.00 sec 12.9 MBytes 27.1 Mbits/sec receiver [ 26] 0.00-4.00 sec 425 MBytes 890 Mbits/sec receiver [ 28] 0.00-4.00 sec 2.69 MBytes 5.63 Mbits/sec receiver [ 30] 0.00-4.00 sec 2.68 MBytes 5.62 Mbits/sec receiver [ 32] 0.00-4.00 sec 1.13 GBytes 2.42 Gbits/sec receiver [ 34] 0.00-4.00 sec 1.21 GBytes 2.61 Gbits/sec receiver [ 36] 0.00-4.00 sec 3.46 MBytes 7.26 Mbits/sec receiver [ 38] 0.00-4.00 sec 3.87 MBytes 8.13 Mbits/sec receiver [ 40] 0.00-4.00 sec 2.98 MBytes 6.24 Mbits/sec receiver [ 42] 0.00-4.00 sec 3.39 MBytes 7.11 Mbits/sec receiver [ 44] 0.00-4.00 sec 12.1 MBytes 25.4 Mbits/sec receiver [SUM] 0.00-4.00 sec 4.77 GBytes 10.2 Gbits/sec receiver
But - of course: The issue is not related to your warm migration feature. It is the same for Storage Migration.
-
It's really hard to answer precisely in your case what's your bottleneck without spending more time doing a closer investigation on your setup