Epyc VM to VM networking slow
-
@john-c WILL DO! I've told purchasing to order the card you recommended. Let's see how long that'll take....
EDIT: ordered on Amazon. Expect to be here 1st week Nov.
Will report back then pinging you.
-
@manilx said in Epyc VM to VM networking slow:
@john-c WILL DO! I've told purchasing to order the card you recommended. Let's see how long that'll take....
EDIT: ordered on Amazon. Expect to be here 1st week Nov.
Will report back then pinging you.
The Friday (1st November 2024) after this months update to Xen Orchestra or is it the week after?
-
@manilx I've been waiting for your ping back with the report. Following you saying the first week in November 2024, now in the beginning of the 2nd week in November 2024.
I'm wondering how's it going please, anything holding it up?
-
@john-c Hi. Ordered from Amazon that day and after more than 2 weeks order was cancelled without notice from supplier. Reordered from another one and I'm still waiting....
Not easy to get one. -
@manilx said in Epyc VM to VM networking slow:
@john-c Hi. Ordered from Amazon that day and after more than 2 weeks order was cancelled without notice from supplier. Reordered from another one and I'm still waiting....
Not easy to get one.Thanks for your reply. I hope it goes well this time, anyway if it still proves difficult then you can go for another quad port 10Gbe NIC which is compatible to do the LACP 2 bond with.
If the selected quad port 10Gbe NIC is available on general sale, then you can get it through the supplier who provided you with your HPE Care Packs.
-
@john-c @olivierlambert Now we're talking!!!
Here are the results of a 2 VM Delta/NBD backup (initial one) using 2 10GB NICS in bond:
WHAT a difference, when we run XOA on an Intel host instead of an EPYC one, with backups.
I've told this from the beginning, that the slow backup speeds were due to the EPYC issue (as I got 200+ MB/s @home with measily Protectli's on 10G)
Looking on what the Synology gets: I get up to 500+ MB/s during the backup!
-
It's a good discovery that having XOA outside the pool can make the backup performance much better.
How is the problem solving going for the root cause? We too have quite poor network performance and would really like to see the end of this. Can we get a summary of the actions taken so far and what the prognosis is for a solution?
Did anyone try plain Xen on a new 6.x kernel to see if the networking is the same there?
-
Is there a difference when running
iperf3with-C bbrflag on the client side ?
In my testing with AMD EPYC and some other CPUs, results are more consistent overall with BBR, and better on the AMD EPYC side (but no miracle, it's still far from perfect).AMD EPYC 7262
vm to vm, 4 threads, iperf3
Without BBR : 4.5-6 Gbps (sometimes more; varies a lot)
With BBR : 7-8 Gbps -
@TeddyAstie That is interesting. I had a look. The default seems to be
cubic, butbbris available usingmodprobe tcp_bbr. I also wonder if different queuing disciplines (tc qdisc) can help. For example mqprio that spreads packes across the available NIC HW queues? -
@Forza the default one seems
cubicwhich in my testing causes chaotic (either good or bad) network performance on XCP-ng (even on non-EPYC platforms) where BBR is more consistent (and also better on AMD EPYC).I also wonder if different queuing disciplines (tc qdisc) can help. For example mqprio that spreads packes across the available NIC HW queues?
Regarding PV network, I don't think queue management will change anything as netfront/netback is single-queue.It's multi-queue so maybe it changes something. -
For those who have AMD EPYC 7003 (Zen 3 EPYCs), you may find in Processor settings in firmware
- Enhanced REP MOVSB/STOSB (ERMS)
- Fast Short REP MOVSB (FSRM)
Which is apparently disabled by default.
It could be interesting to enable them and see if it changes anything performance-wise. I am not sure if it's just for showing a flag, or if it changes anything in the CPU behavior though.You can also try REP-MOV/STOS Streaming to see it changes anything too.
-
I'm attaching results with Epyc 9004 / AS-1015CS-TNR-EU / MB: H13SSW
user cpu family market v2v 1T v2v 4T h2v 1T h2v 4T notes dknight-bg EPYC 9354P Zen4 server 5.10 G (130/150/250) 6.24 G (131/254/348) 11.1 G (0/131/216) 11.1 G (0/187/302) Disabled: Enhanced REP MOVSB/STOSB (ERMS), Fast Short REP MOVSB (FSRM) dknight-bg EPYC 9354P Zen4 server 6.71 G (112/223/269) 7.11 G (122/261/342) 11.3 G (0/145/190) 11.5G (0/179/282) Enabled: Enhanced REP MOVSB/STOSB (ERMS), Fast Short REP MOVSB (FSRM) I couldn't find a setting for REP-MOV/STOPS Streaming in the BIOS, nor in the MB manual.
-
In VM to VM it's already a "free" +30% perf for 1T, impressive. Also 14% for 4T, not bad.
-
@olivierlambert yes, but this is only a small improvement to overall problem. Still, compared to similar intel platform, v2v network performance is very low. Are there any news for the final solution to this problem?
-
@nicols We're on 2 HP's with AMD EPYC 7543P 32-Core and it is a pain.
Had to deploy Intel server just for running the backups, which improved a lot (but that's stupid with 2 beasts of hosts) -
No obvious solution yet, it's likely due to an architecture problem on AMD, because of CCDs and how CPUs are made. So the solution (if there's any) will be likely a sum of various small improvements to make it bearable.
I'm going to Santa Clara to discuss that with AMD directly (among other things).
-
@olivierlambert No more EPYC's here, that's for sure. In the future it'll be Intel again (even if they then belong to Broadcom and TMC :p)
-
@olivierlambert said in Epyc VM to VM networking slow:
No obvious solution yet, it's likely due to an architecture problem on AMD, because of CCDs and how CPUs are made. So the solution (if there's any) will be likely a sum of various small improvements to make it bearable.
I'm going to Santa Clara to discuss that with AMD directly (among other things).
Do we have other data to back this? The issue is not really common outside of Xen. I do hope some solution comes out from the meeting with AMD.
-
If we become partners officially, we'll be able to have more advanced accesses with their teams. I still have hope, it's just that the pace isn't on me.
-
@olivierlambert
Can we rule out extra_guest_irqs as the root cause of this problem?https://docs.xcp-ng.org/compute/#nvme-storage-devices-on-linux
