Alert: Control Domain Memory Usage

garyabrahams

@fasterfourier
Can I ask what version of XCP-ng that you are running along with the OS version / Kubernetes version you are running? Still trying to work out what may be causing this.

olivierlambert

@garyabrahams What about alt kernel?

garyabrahams

@olivierlambert
I installed that this morning and running my tests for a few hours (so only a short period of time)
So far I have seen this on dom0 (via free -m)
I'm running XCP-ng 8.0

29.1.2021 9:07 AM -- 749 used - 6688 free (initial with all vms running, but no programs)
29.1.2021 9:15 AM -- 767 used - 6673 free (vms running, all tests running)
29.1.2021 9:46 AM -- 778 used - 6654 free
29.1.2021 10:51 AM -- 793 used - 6630 free

Linux cpt-dc-xen02 4.19.68 #1 SMP Fri Sep 27 10:14:57 CEST 2019 x86_64 x86_64 x86_64 GNU/Linux

olivierlambert

So is it better than before?

garyabrahams

Still checking. Going to run for another day and see.

Should the memory usage in DOM0 be dropping as part of normal use?

I understand ram usage going up and down just like a normal OS, but a constant increase in memory usage doesn't seem right to me or am I misunderstanding how this work?

Gary

olivierlambert

"Depends". But you shouldn't have invisible RAM used.

fasterfourier

@garyabrahams

Sorry I was unclear, but we are not running Kubernetes in our environment. We are running Citrix Hypervisor 8.2 LTSR.

fasterfourier

I have another observation to throw in the thread here. In working with Citrix support on our dom0 memory exhaustion issue in CH8.2LTSR, they are focusing on several of our VMs that had dynamic memory control enabled, which is deprecated in CH8.x. They believe this is related to the control domain memory exhaustion.

I have disabled this on all VMs that I can find with the feature enabled and will continue to monitor. I don't have much hope that this is the underlying issue, since we are seeing the memory issue on our pool master, which could only have hosted a VM with DMA enabled for very brief periods of time while other VMs were shuffled around for maintenance.

Does this track with anyone else here?

stormi

This is... Surprising. I thought xl top would allow to rule that out very fast (and it did earlier in this thread when I suspected something related to dom0 memory ballooning). Unless Xen leaks the memory in some way that would not be visible to itself. I don't know if that is even possible and I don't see how that would relate at all with the memory used by dom0, even if there was such a leak related to domU DMC.

fasterfourier

@stormi

I am also suspicious of this diagnosis, and I think this is likely related to checking off the "misalignments" in our configuration before escalating the case to the next level of troubleshooting support. That said, I figured I'd run it by the group here to see if there's any correlation between users with dynamic memory on their VMs and this issue.

garyabrahams

The boxes that I have do not have dynamic memory (never used it), and we are getting the issue.

Some feedback on my test box running 8.0 alternative kernel. Been running it for a week and getting this.

As you can see there was a increase in memory of the first few days, but then it seemed to level off. I'll continue to do some tests, then I'm intending to upgrade to 8.2 and see if I can replicate (both with the standard and alternative kernels).

I'll provide feedback once I have it.

Gary

garyabrahams

I have another production box that has this issue.. and noticed this

[10:41 host ~]# free -m
              total        used        free      shared  buff/cache   available
Mem:           7913        6445          77         210        1390         284
Swap:          1023          41         982

[10:41 host ~]# ps -ef | grep sadc | wc -l
6337

[10:41 host ~]# ps -ef | grep CROND | wc -l
6337

[10:41 host ~]# ps -ef | grep 32766
root       306 32766  0 Jan31 ?        00:00:00 /usr/lib64/sa/sadc -F -L -S DISK 1 1 -
root     32766  2898  0 Jan31 ?        00:00:00 /usr/sbin/CROND -n

Not sure why I have 6337 processes for CROND and sadc, but going to do some investigations

dave

[10:36 xs03 ~]# free -m
              total        used        free      shared  buff/cache   available
Mem:          11921       11322         171         151         427         175
Swap:          1023          37         986
[10:36 xs03 ~]# ps -ef | grep CROND | wc -l
1

BTW: All my affected pools never had dynamic memory.

stormi

@garyabrahams I think this is a separate issue that would deserve a separate thread, though it's interesting to have mentioned it here just in case someone else would have noticed something similar (I don't remember anyone mentioning such proliferation of processes in this thread).

Now, maybe that host also is affected by the memory leak, but for now nothing allows to think both issues are related. Or maybe the lack of free memory is what caused the processes to never quit.

fasterfourier

Update: we have disabled dynamic memory on all VMs in our pool and the issue is still occurring.I expect this to be sent to the citrix developers shortly, since the normal support team has exhausted their troubleshooting options.

garyabrahams

An update from my side.
I have tried 8.0 alt kernel, 8.2 standard kernel and 8.2 alt kernel and in each case the memory usage increased over time

Below the first increase is 8.0 alt kernel, 2nd increase was 8.2 standard and 3rd 8.2 alt kernel.

olivierlambert

I don't think there's enough time to really be sure about the memory leak. It's normal to see raising RAM usage, what's not is to consume all the dom0. Can you wait a bit longer between 2 tests?

OlivierD

Hello,
I can confirm we encountered this issue on several hosts in the past few months.

Config :

Server is Dell PowerEdge M630
local storage on SSD + remote SR on ISCSI
kernel used: kernel-4.19.19-6.0.11.1.xcpng8.1.x86_64
we use ixgbe (intel-ixgbe-5.5.2-2.xcpng8.1.x86_64) and network cards are Intel 82599 10 Gigabit dual Port (with bonding on XCP NG).
firmwares up to date (less than 6 months, when we updated to XCP-NG 8.1+).

Additional informations :

we don't have any VM with memory ballooning
shutting down VMs does not free memory
we don't do many operations each day (less than 5 reboots/stop/start).
size of pools does not matter (bug happened with two hosts and on another pool with 10 hosts).

According to some messages, it seems kernel-alt fixes the issue ... We'll try to switch kernel when we encounter the issue again.

olivierlambert

Yes, please, keep us posted

Thank for your feedback!

fasterfourier

One more observation here. This issue does not occur on a different pool of ours that's also running CH8.2LTSR. That pool has lower loading overall, 2 hosts instead of 7, and does not contain any NICs using the ixgbe driver. Other aspects of the pool are identical.