changelog:
qom-get return values are
$info->{stats}->{"stat-total-memory"};
$info->{stats}->{"stat-free-memory"};
and not
$info->{"stat-total-memory"};
$info->{"stat-free-memory"};
Signed-off-by: Alexandre Derumier
---
PVE/QemuServer.pm | 21 +++--
1 file changed, 11 insert
applied, thanks!
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
changelog:
only use qom-get, no need to call qmp query-balloon
query-ballon "actual" = qom-get "stat-total-memory"
query-ballon "maxmem" = config memory value (with pending, the value is
accurate now)
(I'll try to autoballoning today)
___
pve-devel
Signed-off-by: Alexandre Derumier
---
PVE/QemuServer.pm | 21 +++--
1 file changed, 11 insertions(+), 10 deletions(-)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index bb7a7f3..c884854 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -2490,17 +2490,12 @@ sub vm
now zfs will wait 5 sec if error msg is "dataset is busy"
Signed-off-by: Wolfgang Link
---
PVE/Storage/ZFSPoolPlugin.pm | 18 --
1 file changed, 16 insertions(+), 2 deletions(-)
diff --git a/PVE/Storage/ZFSPoolPlugin.pm b/PVE/Storage/ZFSPoolPlugin.pm
index 5cbd1b2..d8721a4 100
>>This can be retrieve too with qom : stat-total-memory
>>so,only "actual" balloon value need to an extra qmp query.
Oops, sorry, this is wrong.
qom "stat-total-memory" = qmp "actual"
I need to rebase my patch for a v3
- Mail original -
De: "aderumier"
À: "dietmar"
Cc: "pve-devel
>>only max_mem is missing from orignal qmp query-balloon or qom-get.
>>No sure we need to keep a patch just for this. We can use vmid.conf value.
>>(with pending, now we also have true value)
This can be retrieve too with qom : stat-total-memory
so,only "actual" balloon value need to an extr
Signed-off-by: Alexandre Derumier
---
PVE/QemuServer.pm | 30 --
1 file changed, 24 insertions(+), 6 deletions(-)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index bb7a7f3..01f671d 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -2486,22 +2486,37 @
changelog:
qom values return are stat-free-memory && stats-total-memory,
not free_mem && total_mem
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
>> Maybe it's time to remove this patch, and use qemu qom implementation ?
>>
>>Would be great, but current code needs 'actual, max_mem'?
just send a patch.
the actual value is return by original query-balloon.
only max_mem is missing from orignal qmp query-balloon or qom-get.
No sure we need to
This patch use qom-get to retrieve guest balloon statisctics,
instead to use the proxmox query-balloon patch,
which seem to not work anymore when the vm is started with -machine option.
I don't have tested it too much, so help is welcome for testing
_
Signed-off-by: Alexandre Derumier
---
PVE/QemuServer.pm | 30 --
1 file changed, 24 insertions(+), 6 deletions(-)
diff --git a/PVE/QemuServer.pm b/PVE/QemuServer.pm
index bb7a7f3..2bb907c 100644
--- a/PVE/QemuServer.pm
+++ b/PVE/QemuServer.pm
@@ -2486,22 +2486,37 @
On 03/06/2015 05:57 AM, Dietmar Maurer wrote:
So this regex will never match!
oh, just saw that the colon is added inside run_command - so just
ignore this.
+
+ if ($ret =~ m/.*: dataset is busy.*/){
+
+ for(my $i = 0; $ret && $i < 5; $i++){
+ sleep(1);
+
> As far I remember, we have make the virtio-balloon-fix-query.patch patch,
> because at this time, it was impossible to use qom guest-stats to retrieve all
> values in 1 call
> (And we wanted to avoid to call 5-6 times qom guest-stats)
yes, we use that inside QemuServer::vmstatus, so that needs t
> On March 6, 2015 at 5:26 AM Alexandre DERUMIER wrote:
>
>
> Seem to be a proxmox bug finally (or maybe proxmox use a path not used anymore
> by qemu),
>
> using qom balloon guest-stats (with the guest is started with -machine ..)
Maybe you can try to find out what we do differently?
___
> -run_command($cmd, outfunc => $output, timeout => $timeout);
> +run_command($cmd, errmsg => "ERROR", outfunc => $output, timeout =>
> $timeout);
Note: you use "ERROR" here - without collon.
>
> return $msg;
> }
> @@ -291,7 +291,26 @@ sub zfs_create_zvol {
> sub zfs_delete_zvol
@Dietmar:
As far I remember, we have make the virtio-balloon-fix-query.patch patch,
because at this time, it was impossible to use qom guest-stats to retrieve all
values in 1 call
(And we wanted to avoid to call 5-6 times qom guest-stats)
Now,it seem that we can get all values directly. (excl
Seem to be a proxmox bug finally (or maybe proxmox use a path not used anymore
by qemu),
using qom balloon guest-stats (with the guest is started with -machine ..)
my $stats = vm_mon_cmd_nocheck($vmid, 'qom-get',
path => "machine/peripheral/balloon0",
now zfs will wait 5 sec if error msg is "dataset is busy"
Signed-off-by: Wolfgang Link
---
PVE/Storage/ZFSPoolPlugin.pm | 23 +--
1 file changed, 21 insertions(+), 2 deletions(-)
diff --git a/PVE/Storage/ZFSPoolPlugin.pm b/PVE/Storage/ZFSPoolPlugin.pm
index 5cbd1b2..999e43
maybe this commit
http://git.qemu.org/?p=qemu.git;a=commit;h=91ba21208839643603e7f7fa5864723c3f371ebe
http://lists.nongnu.org/archive/html/qemu-devel/2014-09/msg05176.html
virtio-balloon: fix virtio-balloon child refcount in transports
diff --git a/hw/virtio/virtio-pci.c b/hw/virtio/virtio-pci.c
Thanks Michael
That's seem to confirm that the problem come from the "-machine " option.
I'm currently digging inside qemu code to find what does it change.
- Mail original -
De: "datanom.net"
À: "pve-devel"
Envoyé: Jeudi 5 Mars 2015 17:30:00
Objet: Re: [pve-devel] balloon bug in qemu
On Thu, 5 Mar 2015 16:44:05 +0100 (CET)
Alexandre DERUMIER wrote:
> Note that balloon itself is working
>
> # info balloon
> balloon: actual=1024 max_mem=1024
>
> # balloon 512
>
> # info balloon
> balloon: actual=512 max_mem=1024
>
>
Before migration:
# info version
2.1.3
# info balloon
ba
Note that balloon itself is working
# info balloon
balloon: actual=1024 max_mem=1024
# balloon 512
# info balloon
balloon: actual=512 max_mem=1024
- Mail original -
De: "aderumier"
À: "Andrew Thrift"
Cc: "pve-devel"
Envoyé: Jeudi 5 Mars 2015 16:17:49
Objet: Re: [pve-devel] balloon
ok,I speak too fast.
It's not related to this patch.
on current pve-qemu-kvm 2.1.
balloon is working fine
# info balloon
balloon: actual=1024 max_mem=1024 total_mem=1002 free_mem=941 mem_swapped_in=0
mem_swapped_out=0 major_page_faults=120 minor_page_faults=215272
last_update=1425568324
But
I need to do more tests,
but it seem that this commit (applied on qemu 2.2 but not on qemu 2.1)
http://git.qemu.org/?p=qemu.git;a=commit;h=22644cd2c60151a964d9505f4c5f7baf845f20d8
fix the problem with qemu 2.1.
(I have tested with the patch, balloon works fine, I need to test without the
patch t
in proxmox virtio-balloon-fix-query.patch,
we have
hw/virtio/virtio-balloon.c
+
+if (!(balloon_stats_enabled(dev) && balloon_stats_supported(dev) &&
+ dev->stats_last_update)) {
+ return;
+}
+
+info->last_update = dev->stats_last_update;
+info->has_last_update = t
Hi Alexandre,
This may be the cause of the crashes we have been experiencing.We
reported them here:
http://forum.proxmox.com/threads/21276-Kernel-Oops-Panic-on-3-10-5-and-3-10-7-Kernels
These only started happening since we moved to qemu-2.1.x and we get the
same output:
# info balloon
ball
On 03/05/2015 02:13 PM, Dietmar Maurer wrote:
$msg .= "$line\n";
};
-run_command($cmd, outfunc => $output, timeout => $timeout);
+if ($method eq "destroy") {
+
+eval {run_command($cmd, errmsg => 1, outfunc => $output, timeout
=> $timeout);};
+
+if(my $err = $@) {
+
On 03/03/2015 01:17 PM, Wolfgang Link wrote:
now zfs will wait 5 sec if error msg is "dataset is busy"
Signed-off-by: Wolfgang Link
---
PVE/Storage/ZFSPoolPlugin.pm | 28 ++--
1 file changed, 26 insertions(+), 2 deletions(-)
diff --git a/PVE/Storage/ZFSPoolPlugin.p
applied both patches, thanks!
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Hi,
I have see a bug report here:
http://forum.proxmox.com/threads/2-RAM-Problem-since-Upgrade-to-3-4?p=108367&posted=1#post108367
about balloon.
on my qemu 2.2
#info balloon
balloon: actual=1024 max_mem=2048 total_mem=985 free_mem=895 mem_swapped_in=0
mem_swapped_out=0 major_page_faults=
storage.cfg
---
rbd mystorage
monhost 10.5.0.11:6789; 10.5.0.12:6789; 10.5.0.13:6789
drive_add auto "file=rbd:poolceph1/vm-101-disk-8:mon_host=10.5.0.11\\:6789;
10.5.0.12\\:6789;
10.5.0.13\\:6789:id=admin:auth_supported=cephx:keyring=/etc/pve/priv/ceph/cephzimbra.keyring,if=none,id=d
we need to escape ":" used to defined mon ports
"10.5.0.11:6789; 10.5.0.12:6789; 10.5.0.13:6789"
->
"10.5.0.11\:6789; 10.5.0.12\:6789; 10.5.0.13\:6789"
Signed-off-by: Alexandre Derumier
---
PVE/Storage/RBDPlugin.pm | 11 +++
1 file changed, 3 insertions(+), 8 deletions(-)
diff --gi
>>Ok,I think we can do it like this.
>>I think that only qemu command line need this, but I'll do tests to be sure.
I have checked, it's not so easy.
I think we should keep the addslashes in rbdplugin
if you have a storage.cfg with monhost with spaces and ports
monhost 10.5.0.11:6789; 10.5.0
Hi Detlef,
I have seen problems similar to this previously, but only on test clusters.
What version of packages?
My diagnose was that there was some process gone crazy that affected all
the node. I don't think I had CTs there. Did you have VMs running?
On 05/03/15 07:55, Detlef Bracker wrot
Oh, I means I have found from where this was comming!
I had before on one container a hacker and I have down this container
(blocking customer) and I have forgot
to set the onboot=no and yesterday I have restart the host, but this is
interessting - and 999 is only a clone from
280 for analysis and
36 matches
Mail list logo
