Agree, looks like disk caching, which is considered in use but can be freed as things ask for it. What kind of storage are you using here?
I checked my converged group, the GUI memory use looks like it includes the cache as well, just not as extreme as yours :) If you wanted to confirm, you could clear it with
sync; echo 3 > /proc/sys/vm/drop_caches
and you should get all the cache memory back without rebooting. Presumably, it’s all holding filesystem data for quick access, so not a terrible thing, just alarming to see 97% utilization. Unless it kicked off unneeded MOM runs if you’re using same page merging and ballooning, then its doing unneeded work and maybe slowing your VMs… You’d probably see it in top if this was the case, it’s fairly CPU intensive.
Ultimately, you may want to file a bug/feature request to have the GUI differentiate cache from “regular’ in use memory for clarity.
On Dec 13, 2018, at 12:06 AM, Tony Brian Albers <
tba@kb.dk> wrote:
Screenshots attachedWe have 2 hosts and 24 running vms, all vms are pretty small.As you can see, top and the dashboard does not agree.An interesting thing is that if I view the host itself in the engine,it says under "General" that Max free Memory for scheduling new VMs: 413360 MBSo maybe it's some sort of caching that's using the memory./tonyOn Wed, 2018-12-12 at 09:59 -0600, Darrell Budic wrote:Yeah, you’re right about 400G, I dropped a digit reading it out of
your top display.
So what are you seeing in the dashboard, I’m not sure I understand
the disconnect between the top you shared and what you’re seeing
there. It shows lots more than 110G in use, I gather? Or are you
seeing this on the hosts page per host mem use?
On Dec 12, 2018, at 12:34 AM, Tony Brian Albers <tba@kb.dk> wrote:
I'm not following you on the 42G available, the way I see it
there's
400+G available:
[root@man-001 ~]# free -h
total used free shared buff/cache
a
vailable
Mem: 503G 96G 19G 205M 387G
405G
Swap: 4.0G 520K 4.0G
And here's top sorted by %mem usage:
top - 07:29:00 up 104 days, 20:56, 1 user, load average: 0.59,
0.68,
0.67
Tasks: 564 total, 1 running, 563 sleeping, 0 stopped, 0
zombie
%Cpu(s): 1.5 us, 0.2 sy, 0.0 ni, 98.2 id, 0.0 wa, 0.0 hi, 0.0
si, 0.0 st
KiB Mem : 52807689+total, 20085144 free, 10132981+used,
40666195+buff/cache
KiB Swap: 4194300 total, 4193780 free, 520 used.
42491062+avail
Mem
PID USER PR NI VIRT RES SHR S %CPU
%MEM TIME+
COMMAND
5517 qemu 20 0 9128268 8.1g 14084
S 3.0 1.6 5892:07
qemu-kvm
14187 qemu 20 0 9210236 8.1g 14072
S 5.3 1.6 6586:00
qemu-kvm
12791 qemu 20 0 9272448 8.1g 14140
S 14.2 1.6 17452:10
qemu-kvm
135526 qemu 20 0 9117748 8.1g 13664
S 2.3 1.6 5874:48
qemu-kvm
7938 qemu 20 0 9129936 8.1g 13744
S 2.3 1.6 22109:28
qemu-kvm
11764 qemu 20 0 9275520 8.1g 13720
S 3.3 1.6 10679:25
qemu-kvm
12066 qemu 20 0 9360552 8.1g 13708
S 3.0 1.6 10724:34
qemu-kvm
11153 qemu 20 0 9113544 8.1g 13700
S 15.6 1.6 19050:12
qemu-kvm
12436 qemu 20 0 9161800 8.1g 13712
S 16.2 1.6 21268:00
qemu-kvm
6902 qemu 20 0 9110480 8.0g 13580
S 0.7 1.6 1804:16
qemu-kvm
7621 qemu 20 0 9203816 4.8g 14264
S 1.7 1.0 3143:35
qemu-kvm
6587 qemu 20 0 4880980 4.1g 13744
S 0.7 0.8 2354:56
qemu-kvm
7249 qemu 20 0 4913084 1.6g 13712
S 0.7 0.3 1380:38
qemu-kvm
111877 qemu 20 0 1911088 1.1g 14076 S 0.3 0.2
419:58.70
qemu-kvm
4602 vdsm 0 -20 4803160 114184 13860
S 1.3 0.0 2143:44
vdsmd
4058 root 15 -5 1154020 38804 9588
S 0.0 0.0 0:00.81
supervdsmd
818 root 20 0 84576 35356 34940
S 0.0 0.0 1:05.60
systemd-journal
3602 root 20 0 1496796 32536 9232 S 0.0 0.0
123:53.70
python
2672 root 20 0 358328 30228 7984
S 0.0 0.0 0:14.76
firewalld
4801 vdsm 20 0 1640996 28904 5484
S 0.0 0.0 1265:14
python
Rebooting a host doesn't help, (I've tried that earlier) the only
thing
that works is to stop all vm's, reboot all hosts at the same time
and
start vm's again. Then memory usage shown in the dashboard slowly
increases over time again.
/tony
On Tue, 2018-12-11 at 14:09 -0600, Darrell Budic wrote:
That’s only reporting 42G available of your 512, ok but something
still using it. Try sorting the top by memory %, should be ‘>’
while
it’s running.
On Dec 11, 2018, at 1:39 AM, Tony Brian Albers <tba@kb.dk>
wrote:
Looks ok to me:
top - 08:38:07 up 103 days, 22:05, 1 user, load average:
0.68,
0.62,
0.57
Tasks: 565 total, 1 running, 564 sleeping, 0 stopped, 0
zombie
%Cpu(s): 1.0 us, 0.5 sy, 0.0 ni, 98.5 id, 0.0 wa, 0.0
hi, 0.0
si, 0.0 st
KiB Mem : 52807689+total, 22355988 free, 10132873+used,
40439219+buff/cache
KiB Swap: 4194300 total, 4193780 free, 520 used.
42492028+avail
Mem
PID USER PR NI VIRT RES SHR S %CPU
%MEM TIME+
COMMAND
14187 qemu 20 0 9144668 8.1g 14072
S 12.6 1.6 6506:46
qemu-kvm
11153 qemu 20 0 9244680 8.1g 13700
S 4.3 1.6 18881:11
qemu-kvm
12436 qemu 20 0 9292936 8.1g 13712
S 3.3 1.6 21071:56
qemu-kvm
5517 qemu 20 0 9128268 8.1g 14084
S 3.0 1.6 5801:03
qemu-kvm
11764 qemu 20 0 9185364 8.1g 13720
S 3.0 1.6 10585:14
qemu-kvm
7938 qemu 20 0 9252876 8.1g 13744
S 2.6 1.6 21912:46
qemu-kvm
12791 qemu 20 0 9182292 8.1g 14140
S 2.6 1.6 17299:36
qemu-kvm
4602 vdsm 0 -20 4803160 <tel:20%204803160>
114132 13860
S 2.3 0.0 2123:45
vdsmd
7621 qemu 20 0 9187424 4.8g 14264
S 2.3 1.0 3114:25
qemu-kvm
12066 qemu 20 0 9188436 8.1g 13708
S 2.3 1.6 10629:53
qemu-kvm
135526 qemu 20 0 9298060 8.1g 13664
S 2.0 1.6 5792:05
qemu-kvm
6587 qemu 20 0 4883036 4.1g 13744
S 1.3 0.8 2334:54
qemu-kvm
3814 root 20 0 1450200 <tel:1450200> 25096 14208
S 1.0 0.0 368:03.80
libvirtd
6902 qemu 20 0 9110480 8.0g 13580
S 1.0 1.6 1787:57
qemu-kvm
7249 qemu 20 0 4913084 1.6g 13712
S 0.7 0.3 1367:32
qemu-kvm
It looks like it's only in oVirt-engine that there's an issue.
The
host
seems happy enough.
/tony
On Mon, 2018-12-10 at 20:14 -0600, Darrell Budic wrote:
Grab a shell on your hosts and check top memory use quick.
Could
be
VDSMD, in which case restarting the process will give you a
temp
fix.
If you’re running hyperconvered, check your gluster version,
there
was a leak in versions 3.12.7 - 3.1.12 or so, updating
ovirt/gluster
is the best fix for that.
On Dec 10, 2018, at 7:36 AM, Tony Brian Albers <tba@kb.dk>
wrote:
Hi guys,
We have a small test installation here running around 30
vms on
2
hosts.
oVirt 4.2.5.3
The hosts each have 512 GB memory, and the vms are sized
with
4-8
GB
each.
I have noticed that over the last months, the memory usage
in
the
dashboard has been increasing and is now showing 946.8 GB
used
of
1007.2 GB.
What can be causing this?
TIA,
--
--
Tony Albers
Systems Architect
Systems Director, National Cultural Heritage Cluster
Royal Danish Library, Victor Albecks Vej 1, 8000 Aarhus C,
Denmark.
Tel: +45 2566 2383 / +45 8946 2316
_______________________________________________
Users mailing list -- users@ovirt.org
To unsubscribe send an email to users-leave@ovirt.org
Privacy Statement: https://www.ovirt.org/site/privacy-polic
y/
oVirt Code of Conduct: https://www.ovirt.org/community/abou
t/co
mmun
ity-guidelines/
List Archives: https://lists.ovirt.org/archives/list/users@
ovir
t.or
g/message/SDDH2OC5RBOVYYCLGPOUF6HO676HWI5U/
--
Tony Albers
Systems Architect
Systems Director, National Cultural Heritage Cluster
Royal Danish Library, Victor Albecks Vej 1, 8000 Aarhus C,
Denmark.
Tel: +45 2566 2383 <tel:+45%202566%202383> / +45 8946 2316
<tel:+45%208946%202316>
--
--
Tony Albers
Systems Architect
Systems Director, National Cultural Heritage Cluster
Royal Danish Library, Victor Albecks Vej 1, 8000 Aarhus C, Denmark.
Tel: +45 2566 2383 / +45 8946 2316
-- Tony AlbersSystems ArchitectSystems Director, National Cultural Heritage ClusterRoyal Danish Library, Victor Albecks Vej 1, 8000 Aarhus C, Denmark.Tel: +45 2566 2383 / +45 8946 2316<2018-12-13-070032_349x526_scrot.png><2018-12-13-070259_504x81_scrot.png>