Re: [pve-devel] novnc update

2015-08-28 Thread Dietmar Maurer
But It's really usefull for embedded console. OK, applied. Thanks! ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

Re: [pve-devel] ZFSPlugin snapshot question

2015-08-28 Thread Dietmar Maurer
Yes, I will look in to this this weekend. Forgot to mention. Since snapshots by nature is read-only you have to set option wp=true to have iscsi properly advertise read-only to the mount util. I think Linux will make some internal optimization given it knows writes is not supported. Just

Re: [pve-devel] [PATCH] fix: start kvm with os type other

2015-08-28 Thread Dietmar Maurer
applied with modification suggested by alexandre https://git.proxmox.com/?p=qemu-server.git;a=commitdiff;h=d853f40aefc6298aa892f8861fb246825202044c please test. ___ pve-devel mailing list pve-devel@pve.proxmox.com

Re: [pve-devel] ZFSPlugin snapshot question

2015-08-28 Thread Dietmar Maurer
On August 28, 2015 at 12:19 PM datanom.net m...@datanom.net wrote: On 2015-08-28 10:25, Dietmar Maurer wrote: Another approach would be: 1.) create a clone 2.) backup cloned VM with 'stop' mode That way we do not even need to implement 'snapshot' backups ;-) Any comments

Re: [pve-devel] ZFSPlugin snapshot question

2015-08-28 Thread datanom.net
On 2015-08-28 10:25, Dietmar Maurer wrote: Another approach would be: 1.) create a clone 2.) backup cloned VM with 'stop' mode That way we do not even need to implement 'snapshot' backups ;-) Any comments about that idea? Will creating a clone instead of a snapshot not take considerably

[pve-devel] [PATCH] fix to prevent uninitialized variable massage.

2015-08-28 Thread Wolfgang Link
--- PVE/Storage/ZFSPlugin.pm | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/PVE/Storage/ZFSPlugin.pm b/PVE/Storage/ZFSPlugin.pm index 3878342..52e9090 100644 --- a/PVE/Storage/ZFSPlugin.pm +++ b/PVE/Storage/ZFSPlugin.pm @@ -275,7 +275,7 @@ sub alloc_image { my $volname

[pve-devel] [PATCH] HA resources check before destroying vm

2015-08-28 Thread Alen Grizonic
--- PVE/API2/Qemu.pm | 4 1 file changed, 4 insertions(+) diff --git a/PVE/API2/Qemu.pm b/PVE/API2/Qemu.pm index 513b224..a6ad831 100644 --- a/PVE/API2/Qemu.pm +++ b/PVE/API2/Qemu.pm @@ -1142,6 +1142,10 @@ __PACKAGE__-register_method({ my $storecfg = PVE::Storage::config(); +

[pve-devel] [PATCH] HA resources check before destroying lxc

2015-08-28 Thread Alen Grizonic
--- src/PVE/API2/LXC.pm | 4 1 file changed, 4 insertions(+) diff --git a/src/PVE/API2/LXC.pm b/src/PVE/API2/LXC.pm index 35f2c24..303f0ce 100644 --- a/src/PVE/API2/LXC.pm +++ b/src/PVE/API2/LXC.pm @@ -532,6 +532,10 @@ __PACKAGE__-register_method({ my $storage_cfg =

[pve-devel] avoid mointing same filesystem multiple times

2015-08-28 Thread Dietmar Maurer
Seems there is no protection against this dangerous situation: # qemu-img create -f raw test.raw 4G # mkfs.ext4 test.raw # mkdir mp1 # mkdir mp2 # mount -o loop test.raw mp1 # mount -o loop test.raw mp2 So how can we make sure that this never happens? We use loop mounts in our pct code.

[pve-devel] [PATCH v2] HA resources check before destroying lxc

2015-08-28 Thread Alen Grizonic
[PATCH v2] error message changed --- src/PVE/API2/LXC.pm | 3 +++ 1 file changed, 3 insertions(+) diff --git a/src/PVE/API2/LXC.pm b/src/PVE/API2/LXC.pm index 35f2c24..935493b 100644 --- a/src/PVE/API2/LXC.pm +++ b/src/PVE/API2/LXC.pm @@ -532,6 +532,9 @@ __PACKAGE__-register_method({

Re: [pve-devel] [PATCH] HA resources check before destroying vm

2015-08-28 Thread Dietmar Maurer
What about LXC Containers? On August 28, 2015 at 1:26 PM Alen Grizonic a.grizo...@proxmox.com wrote: --- PVE/API2/Qemu.pm | 4 1 file changed, 4 insertions(+) diff --git a/PVE/API2/Qemu.pm b/PVE/API2/Qemu.pm index 513b224..a6ad831 100644 --- a/PVE/API2/Qemu.pm +++

[pve-devel] [PATCH v2] HA resources check before destroying vm

2015-08-28 Thread Alen Grizonic
[PATCH v2] error message changed --- PVE/API2/Qemu.pm | 3 +++ 1 file changed, 3 insertions(+) diff --git a/PVE/API2/Qemu.pm b/PVE/API2/Qemu.pm index 513b224..73a1572 100644 --- a/PVE/API2/Qemu.pm +++ b/PVE/API2/Qemu.pm @@ -1142,6 +1142,9 @@ __PACKAGE__-register_method({ my $storecfg =

Re: [pve-devel] [PATCH] HA resources check before destroying lxc

2015-08-28 Thread Alen Grizonic
The code in PVE::HA::Config::vm_is_ha_managed assumes that the LXC insertion in the HA resources file /etc/pve/ha/resources.cfg starts with vm: so if we decide to change the insert to ct:, some code has to be added! On 08/28/2015 03:03 PM, Alen Grizonic wrote: --- src/PVE/API2/LXC.pm | 4

[pve-devel] [PATCH v2 2/2] HA resources check before destroying lxc

2015-08-28 Thread Alen Grizonic
[PATCH v2] error message changed --- src/PVE/API2/LXC.pm | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/src/PVE/API2/LXC.pm b/src/PVE/API2/LXC.pm index 303f0ce..935493b 100644 --- a/src/PVE/API2/LXC.pm +++ b/src/PVE/API2/LXC.pm @@ -532,9 +532,8 @@

[pve-devel] [PATCH v2 2/2] HA resources check before destroying vm

2015-08-28 Thread Alen Grizonic
[PATCH v2] error message changed --- PVE/API2/Qemu.pm | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/PVE/API2/Qemu.pm b/PVE/API2/Qemu.pm index a6ad831..73a1572 100644 --- a/PVE/API2/Qemu.pm +++ b/PVE/API2/Qemu.pm @@ -1143,8 +1143,7 @@ __PACKAGE__-register_method({

[pve-devel] qemu kvm_steal_time bug

2015-08-28 Thread Alexandre DERUMIER
Hi, I have add this bug, which occur after live migration: https://code.google.com/p/ganeti/issues/detail?id=986 https://groups.google.com/forum/#!topic/ganeti/D2Yos3BShmc https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=785557#64 qemu 2.4 or 2.1 and guest kernel 3.16 (debian jessie). The

Re: [pve-devel] ZFSPlugin snapshot question

2015-08-28 Thread Dietmar Maurer
Yes, i guess so. It also has other problems, because not all storage type supports cloning from snapshots ... So it was an interesting idea, but does not really solve all problems :-/ To solve the problem for the ZFSPlugin, I think it may help to introduce another storage API call to

Re: [pve-devel] ZFSPlugin snapshot question

2015-08-28 Thread Dietmar Maurer
While thinking about this, I came to a totally different topic. Would it make sense to add an Linux iSCSI (LIO) backend (LunCMD driver)? It is really easy to setup a ZFS box with our installer, so one could easily use such box as external iSCSI storage then? Comments?

Re: [pve-devel] [PATCH v2 2/2] HA resources check before destroying lxc

2015-08-28 Thread Dietmar Maurer
[PATCH v2] error message changed applied, thanks! (I changed coding style slightly). ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel

Re: [pve-devel] qemu kvm_steal_time bug

2015-08-28 Thread Alexandre DERUMIER
Is it only with 3.16 kernel? Haven't seen this yet. Yes, I have reproduce it on 3 debian jessie kernel 3.16. But other jessie migration have worked before... I have asked to the qemu mailing, waiting for more information. - Mail original - De: Stefan Priebe s.pri...@profihost.ag À:

Re: [pve-devel] ZFSPlugin snapshot question

2015-08-28 Thread Michael Rasmussen
On Fri, 28 Aug 2015 19:35:51 +0200 (CEST) Dietmar Maurer diet...@proxmox.com wrote: I just thought we can write a PVE::Storage::LunCmd::LIO, so that we can use a normal Linux/Proxmox box (with zfsonlinux) with the PVE::Storage::ZFSPlugin. This makes sense. I guess it is already

Re: [pve-devel] ZFSPlugin snapshot question

2015-08-28 Thread Lindsay Mathieson
I found this on reddit: https://www.reddit.com/r/debian/comments/34t0gc/whats_your_favourite_thing_about_jessie/cqxuyy2 The targetcli author replies. I'm a bit confused as to iscsi server support in proxmox - there's tgt, but appears to be missing init scripts? Iet doesn't work with kernel

Re: [pve-devel] ZFSPlugin snapshot question

2015-08-28 Thread Dietmar Maurer
I guess it is already possible with PVE::Storage::LunCmd::Iet, but LIO should provide better performance? I have been thinking about this as well because Iet is deprecated in newer distributions (younger Debian Wheezy) But I just noticed that there is no targetcli package for jessie! I