Am 17.12.2013 09:40, schrieb Dietmar Maurer:
>>> Tue Dec 17 07:58:55 2013: Mounting local filesystems...done.
>>> Tue Dec 17 07:58:55 2013: Activating swapfile swap...done.
>>> Tue Dec 17 07:58:55 2013: Cleaning up temporary files
>>> Tue Dec 17 07:58:55 2013: Setting kernel variables ...sysctl
Am 17.12.2013 09:38, schrieb Dietmar Maurer:
>>> What exactly are the advanced features?
>>>
>>
>> for example flow support
>
> What I do not understand is why this is advanced? AFAIK iptables has much
> more features than openflow?
> Or what do you talk about exactly?
Let's assume you have bond
Am 17.12.2013 08:08, schrieb Dietmar Maurer:
> I get the following warnings when booting with 3.10 kernel:
>
>
>
> ---
>
> Tue Dec 17 07:58:55 2013: Mounting local filesystems...done.
> Tue Dec 17 07:58:55 2013: Activating swapfile swap...done.
> Tue Dec 17 07:58:55 2013: Cleaning up t
Am 17.12.2013 09:21, schrieb Dietmar Maurer:
>> My point is that it could be fine to give user choice (at least for advanced
>> users,
>> not everyone need advanced features of openvswitch)
>
> What exactly are the advanced features?
>
for example flow support
__
Am 17.12.2013 07:56, schrieb Alexandre DERUMIER:
>>> it just works for me with vanilla 3.10 and the additional patch. BUT
>>> without VLAN filtering i don't use it.
>
> Don't you use special setup with bridge on top of another bridge ? (It was
> about gvrp support If I remember)
This was neede
Hi,
wouldn't it be a good idea to integrate http://kanaka.github.io/noVNC/?
It's pretty fast and there would be no need for java anymore...
Greets,
Stefan
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listin
> Am 11.12.2013 um 07:44 schrieb Dietmar Maurer :
>
> I just re-activated the pvetest repository. I hope this makes it easier for
> developers to test.
Thanks! This makes a lot of things easier.
Stefan
>
> - Dietmar
> ___
> pve-devel mailing list
Hi,
i wanted to test spice and installed virt-viewer under ubuntu.
If i try to connect via spice to a machine and start virt-viewer. I get
this error message while using firefox:
"Verbindung zu libvirt mit URI [keine] konnte nicht hergestellt werden"
Any ideas what's wrong?
Chroma has another
Hi,
what's the disadvantage of having SPICE (qxl) as the default graphics
card for every vm? Wouldn't it make sense to do so?
Greets,
Stefan
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Hi,
Am 09.12.2013 11:25, schrieb Dietmar Maurer:
> Hi Stefan,
>
> I applied your patch, but mad some modification
>
> https://git.proxmox.com/?p=pve-manager.git;a=commitdiff;h=496ec71600f4ec8d3a8c46a94b68423399e36a6c
>
> Please can you test?
I get these errors:
{
'errors'
Am 06.12.2013 11:26, schrieb Dietmar Maurer:
>>> yeah but still how do arrays and hashes be encoded? The key and values
>>> are url encoded - sure but how are values of type array encoded?
>>
>> Oh, I guess easiest way would be to add the ability to post 'JSON' data (use
>> application/json instead
Am 06.12.2013 10:31, schrieb Dietmar Maurer:
>> GET
>> https://cloud1-1203.de-nserver.de:8006/api2/json/nodes/cloud1-
>> 1203/execute?commands={JSON
>> ENCODED STUFF}
>>
>> sorry never used pve api with get and arrays.
>
>
> I guess you should use PUT for the batch command (and encode the
> meth
Am 06.12.2013 10:04, schrieb Stefan Priebe - Profihost AG:
>
> Am 06.12.2013 10:00, schrieb Dietmar Maurer:
>>> mind you help me:
>>>
>>> something like:
>>> pvesh get /nodes/cloud1-1203/execute -commands ["a","b"]
>>>
>
Am 06.12.2013 10:00, schrieb Dietmar Maurer:
>> mind you help me:
>>
>> something like:
>> pvesh get /nodes/cloud1-1203/execute -commands ["a","b"]
>>
>> results in:
>> commands: type check ('array') failed
>
> You cannot execute such command with pvesh (pvesh only handles string
> parameters)
>
Am 06.12.2013 09:25, schrieb Dietmar Maurer:
>> Initiating the 250 connection took 2s. But EACH /qemu/123/config request took
>> 0.1s resulting in 25s. My client is most of the time waiting for an answer
>> of the
>> pveproxy server.
>
> OK, then the batch approach is maybe faster - feel free to
Hi Dietmar,
Am 05.12.2013 17:47, schrieb Dietmar Maurer:
>>> I don't really like to add such API extensions, because that is just a speed
>> optimization?
>>
>> Yes. As creating and establishing SSL connections is really expensive.
>
> But you need to create the connection only once - then you
Am 05.12.2013 09:30, schrieb Dietmar Maurer:
>> or just directly something like
>> /nodes/$node/qemu_all_configs
>
> I don't really like to add such API extensions, because that is just a speed
> optimization?
Yes. As creating and establishing SSL connections is really expensive.
> Instead we c
Am 05.12.2013 05:58, schrieb Dietmar Maurer:
>
>> is there a way to get all vm configs with one api request? Or at least all vm
>> configs from one node?
>
> no, sorry
i would like to implement one and though about a new path
qemu/all =>
{
GET config
}
which can or might be filled later. But t
Am 03.12.2013 09:55, schrieb Dietmar Maurer:
>> this was from alexandre - i can clean this up. Will send a new patch in a few
>> seconds.
>
> OK, please update to include latest updates. Also do not touch Alexandres
> 'Signed-Off'
done - see v5
___
pv
Am 03.12.2013 09:46, schrieb Dietmar Maurer:
> I am totally confused now -why does this patch modifies 2 files:
>
>
> new file: debian/patches/internal-snapshot-async-qemu1.7.patch
> modified: debian/patches/internal-snapshot-async.patch
this was from alexandre - i can clean this up.
yes
Mit freundlichen Grüßen
Stefan Priebe
Bachelor of Science in Computer Science (BSCS)
Vorstand (CTO)
---
Profihost AG
Am Mittelfelde 29
30519 Hannover
Deutschland
Tel.: +49 (511) 5151 8181 | Fax.: +49 (511) 5151 8282
URL: http://www.profihost.com | E-Mail: i.
Am 30.11.2013 16:09, schrieb Dietmar Maurer:
> Thanks! I will test that on Monday.
did you find the time to test?
>> Subject: Re: [pve-devel] [PATCH] internal snasphot async port to qemu 1.7 v4
>>
>> This one works completely fine for me and fixes some bugs in the code
>> which does not work an
Am I correct that we can add support for CPU and memory hot plug with qemu 1.7.
Greets
Stefan
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Am 29.11.2013 11:11, schrieb Dietmar Maurer:
>> Anything left to work on for
>> pve-qemu-kvm.git 1.7?
>
> you can help testing when we have finished the patches.
sure - will do so.
Stefan
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.
Sorry for beeing to slate to jump on - was a bit busy the last days.
Anything left to work on for
pve-qemu-kvm.git 1.7?
Greets,
Stefan
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Yes, i would recommand:
http://search.cpan.org/~mstrout/Object-Remote-0.003002/
It's absolutely great!
Stefan
Am 25.11.2013 10:50, schrieb Alexandre DERUMIER:
> Ok, I found a way to only use ssh.
>
> I can create the lun with a small perl script (5lines of code) on the proxmox
> server side.
>
> Am 23.11.2013 um 14:28 schrieb Michael Rasmussen :
>
> On Sat, 23 Nov 2013 07:16:28 +
> Dietmar Maurer wrote:
>
>>> I agree, limiting IO from the VM during backup can have advantages.
>>> On the flip side loosing 50% of the IO
>>
>> This 50% loose has nothing to do with the new backup al
OK known ubuntu bug - sorry.
https://bugs.launchpad.net/ubuntu/+source/chromium-browser/+bug/1250550
Am 13.11.2013 12:01, schrieb Stefan Priebe - Profihost AG:
> Big thanks for the signed app.
>
> Strangely it works fine to me under firefox but does not in chromium
> under ubuntu 1
Big thanks for the signed app.
Strangely it works fine to me under firefox but does not in chromium
under ubuntu 12.04 - i get:
Exception Exception in thread "Thread-3" in thread "Thread-3"
java.lang.NoClassDefFoundError: Could not initialize class
sun.nio.ch.FileChannelImpljava.lang.NoClassDefFo
Am 14.10.2013 08:07, schrieb Alexandre DERUMIER:
> Hi,
>
> I was a little off this last week,
>
> I'm currently working on livemigration+storage migration at the same time
> (because I'll need it in some months, migration vms to a new datacenter with
> new storage).
>
> I think I'll have a fir
Am 10.10.2013 05:33, schrieb Dietmar Maurer:
>> I get values between 500ms and 5s depending on connection. One of the biggest
>> problems seems to be the tasks and ressources list. Both of them have gzip
>> compressed 25kb and pve requests them very often.
>>
>> Wouldn't it be possible to have som
Am 10.10.2013 05:24, schrieb Dietmar Maurer:
> Does that really solve the problem for you (at least partly)?
Yes def. CPU Load was at 100% before while PVE was opened (just through
WLAN and VPN). NOw it is at 2% - 25%.
>> -Original Message-
>> From: pve-devel-boun...@pve.proxmox.com [mail
Am 08.10.2013 10:44, schrieb Dietmar Maurer:
>> But right now there is no callback in:
>> PVE.data.UpdateQueue.queue
>
> It is not very hard to add one ;-)
If you're familiar with JS it def. isn't ;-)
OK my idea would be to add a 2nd param cbf and pass it down to start_update:
Ext.apply(m
Am 08.10.2013 10:14, schrieb Dietmar Maurer:
>> But what's the advantage then? We just introduce an additional callback.
>> Maybe i miss an important part in your idea.
>
> You patch is not 100% clean, because it can still issue multiple calls.
*urg* sure.
But right now there is no callback in:
P
Am 08.10.2013 07:34, schrieb Dietmar Maurer:
>>> Am 08.10.2013 06:33, schrieb Dietmar Maurer:
Wouldn't it be simpler to add a callback which gets called after the load?
if (PVE.Utils.authOK()) {
PVE.data.UpdateQueue.queue(me, function(success) {
Am 08.10.2013 06:33, schrieb Dietmar Maurer:
> Wouldn't it be simpler to add a callback which gets called after the load?
>
> if (PVE.Utils.authOK()) {
> PVE.data.UpdateQueue.queue(me, function(success) {
> load_task.delay(config.interval, run_load_task
Am 01.10.2013 10:40, schrieb Dietmar Maurer:
> applied, thank!
>
>> -Original Message-
>> From: pve-devel-boun...@pve.proxmox.com [mailto:pve-devel-
>> boun...@pve.proxmox.com] On Behalf Of Stefan Priebe
>> Sent: Montag, 30. September 2013 11:51
>> To: pve-devel@pve.proxmox.com
>> Subject:
Am 30.09.2013 13:16, schrieb Dietmar Maurer:
>>> If we do not use '-w', we need to add 'use warnings' to all library files
>>> (I guess
>> it is not used everywhere)?
>>
>> If the modules are compatible hey should contain use warnings.
>
> I am just thinking about our library files in PVE/API/*
Am 30.09.2013 12:06, schrieb Dietmar Maurer:
> q
>> -w enabled warnings for all modules - even when these modules comes from the
>> OS and are not compatible with warnings.
>
> Interesting. Why/when is a module not compatible with warnings?
If the developer has not cleanly programmed. But as lon
Am 30.09.2013 12:30, schrieb Dietmar Maurer:
>> -w enabled warnings for all modules - even when these modules comes from the
>> OS and are not compatible with warnings. This is the reason to NOT use -w and
>> instead use "use warnings".
>
> If we do not use '-w', we need to add 'use warnings' to a
Am 30.09.2013 10:54, schrieb Stefan Priebe - Profihost AG:
> Am 30.09.2013 10:03, schrieb Dietmar Maurer:
>>> never tried but i think the server software should also handle corner cases
>>> to be
>>> safe and should not print any warnings regardless what the client
Am 30.09.2013 10:03, schrieb Dietmar Maurer:
>> never tried but i think the server software should also handle corner cases
>> to be
>> safe and should not print any warnings regardless what the client is doing.
>
> Sure. Would be great if you can find a fix for that.
>
*urg* seems to be a bug i
Am 29.09.2013 07:14, schrieb Dietmar Maurer:
>>> What browser do you use? Any way to reproduce it?
>>
>> I can't reproduce with a browser it happens when using
>> https://metacpan.org/release/Net-Proxmox-VE
>
> Oh, I do not know that software. Does it work when you use PVE/API2Client.pm
> instead
Hello,
i'm trying to upgrade my machines from squeeze to wheezy.
But after upgrading i'm getting the following messages while starting
the cluster:
Starting cluster:
Checking if cluster has been disabled at boot... [ OK ]
Checking Network Manager... [ OK ]
Global setup... [ OK ]
26.09.2013 11:08, schrieb Dietmar Maurer:
>
>
>> -Original Message-
>> From: pve-devel-boun...@pve.proxmox.com [mailto:pve-devel-
>> boun...@pve.proxmox.com] On Behalf Of Stefan Priebe - Profihost AG
>> Sent: Donnerstag, 26. September 2013 09:55
&g
Hi,
while trying to update to wheezy and port and compile my packages, i've
seen that the proxmox repository does no seem to contain the following
packages / versions:
libspice-protocol-dev (>= 0.12.5)
is this expected?
Greets Stefan
___
pve-devel mai
Hi,
are there any planned qemu updates for pve? In the past I read that 1.6 was
planned.
Greets,
Stefan
___
pve-devel mailing list
pve-devel@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
vrp on
reorder_hdr on
ip a add 192.168.99.10/24 bro 192.168.99.255 dev vmbr1.3025
ip link set vmbr1.3025 up
But i'm not able to get a tap device working.
Stefan
> - Mail original -
>
> De: "Stefan Priebe - Profihost AG"
> À: "Alexandre DERUMIER&qu
port)
> mhm i could try that. So you mean instead of adding:
> vmbr1.3021@vmbr1
> and
> vmbr1v3021
>
> just use vmbr1.3021@vmbr1 and add the tap device to that one?
>
>> I really don't known how it's work, I'll try to have a look at it next week.
>
mplement new vlan code ?
>> (don't tag anymore on eth with multibridge, but use only 1 bridge with
>> tagging port)
> mhm i could try that. So you mean instead of adding:
> vmbr1.3021@vmbr1
> and
> vmbr1v3021
>
> just use vmbr1.3021@vmbr1 and add the tap de
Hi,
as Linux Kernel 3.10 ist the newest long term kernel i'll start
migrating all my machines to it.
Everything works fine except for tap devices with VLANs on top of bridges.
I'm pretty sure that this i related to this commit:
https://git.kernel.org/cgit/linux/kernel/git/stable/linux-stable.git
Am 19.08.2013 um 17:14 schrieb Dietmar Maurer :
>> That piece of code was a previous test, i did not want to send it, sorry for
>> that. Why don't you want to translate those expressions ? I think the
>> interface should be completely in the user's language, or not translated at
>> all. An half t
OK works fine with vanilla 3.10.5 kernel instead of the proxmox one.
Stefan
Am 06.08.2013 07:26, schrieb Stefan Priebe - Profihost AG:
> In you look at the forum it only happens on intel 3rd gen processors.
>
> Stefan
>
> This mail was sent with my iPhone.
>
> Am 06.08.
In you look at the forum it only happens on intel 3rd gen processors.
Stefan
This mail was sent with my iPhone.
Am 06.08.2013 um 03:51 schrieb Michael Rasmussen :
> On Tue, 06 Aug 2013 00:35:22 +0200
> Stefan Priebe wrote:
>
>> Hi,
>>
>> does anybody know a solution for this one:
>> http://f
Sorry wasn't up2date. The spice changes did change the file. Will send a
new one.
Am 26.07.2013 11:16, schrieb Dietmar Maurer:
> Applying: qemu-server: add support for unsecure migration (setting in
> datacenter.cfg)
> error: patch failed: PVE/QemuMigrate.pm:320
> error: PVE/QemuMigrate.pm: patch
Am 26.07.2013 11:02, schrieb Dietmar Maurer:
>>> yes, I guess that prevents the issue most times (but it is not 100%)
>>
>> mhm i can't imagine a situation where this can happen. The important part is
>> that the expiretime is BIGGER than the kvm timeout.
>
>
> Ah. What exactly is the kvm timeout
Am 26.07.2013 10:51, schrieb Dietmar Maurer:
>>> The other way. The target crashes (an another target is started using the
>> same port afterwards).
>>
>> ah OK. But when the target crashes and a new migration is started - the new
>> migration will still get a NEW port due to the expiretime logic i
Am 26.07.2013 10:37, schrieb Dietmar Maurer:
>
>
>> -Original Message-----
>> From: Stefan Priebe - Profihost AG [mailto:s.pri...@profihost.ag]
>> Sent: Freitag, 26. Juli 2013 10:36
>> To: Dietmar Maurer
>> Cc: pve-devel@pve.proxmox.com; spriebe
>
Am 26.07.2013 10:23, schrieb Dietmar Maurer:
>>> But consider the following (this can also happen with current code):
>>>
>>> - kvm crash
>>> - another migration starts shortly after using the same port now
>>> - the other side connect now to the wrong VM
>>
>> But this can happen with the tunnel t
Am 26.07.2013 10:05, schrieb Dietmar Maurer:
>>> Maybe we can use some kind of ticket to avoid unprivileged access (qemu
>> patch)?
>>
>> May be - i don't care it's a private network with no connection to the
>> internet.
>
> Maybe.
>
> But consider the following (this can also happen with curre
new patch sent
Am 26.07.2013 10:14, schrieb Dietmar Maurer:
>> OK
>>
>>> (just remove ssh parameters '-L', "$lport:localhost:$rport" if !$rport
>>> in fork_tunnel)
>
> yes (if we do not need the ssh tunnel)
>
>> That makes no sense to me as $rport is always set. Or do you mean if $raddr
>> ne "l
Am 26.07.2013 09:56, schrieb Dietmar Maurer:
>> diff --git a/PVE/QemuMigrate.pm b/PVE/QemuMigrate.pm
>> index dd48f78..be7df23 100644
>> --- a/PVE/QemuMigrate.pm
>> +++ b/PVE/QemuMigrate.pm
>> @@ -306,8 +306,8 @@ sub phase2 {
>>
>> $self->log('info', "starting VM $vmid on remote node '$self->{
Am 26.07.2013 07:14, schrieb Dietmar Maurer:
>> If you need a patch against current git Version just write me
>
> Yes, please send an updated version.
done
> I am still scared to open a tcp port without and kind of protection.
>
> Maybe we can use some kind of ticket to avoid unprivileged acces
If you need a patch against current git Version just write me
Stefan
This mail was sent with my iPhone.
Am 25.07.2013 um 16:12 schrieb Dietmar Maurer :
>> I just found this tutorial to build hpn-ssh for wheezy
>>
>> http://blog.admiralakber.com/?p=248
>>
>> Don't known if it could be a good i
cpu flag ? It should reduce interrupts.
> (It's already in current qemu-server git)
Sorry ignore me. This seems to be a known bug:
https://code.google.com/p/byte-unixbench/issues/detail?id=4
Stefan
> - Mail original -
>
> De: "Stefan Priebe - Profihost AG&
Hi,
while benchmarking my kvm machines using unixbench i see the following
reproducable behaviour.
Host has 24 cores.
KVM Machine set to 2 cores: Unixbench Value 2514
KVM Machine set to 16 cores: Unixbench Value 6714
KVM Machine set to 24 cores: Unixbench Value 1742
Also i see a max of 105% CPU
Am 16.07.2013 10:05, schrieb Dietmar Maurer:
> 139895458642144:error:1408F10B:SSL routines:SSL3_GET_RECORD:wrong version
> number:s3_pkt.c:348:
Looks like you have to use TLS instead of SSLv3. Maybe force protocol?
___
pve-devel mailing list
pve-devel@
.0 \\n \\l\n\n";
Maybe
$banner .= `cat /etc/issue.net`;
is a way better.
> - Mail original -
>
> De: "Michael Rasmussen"
> À: "Stefan Priebe - Profihost AG" , "Alexandre
> DERUMIER"
> Cc: pve-devel@pve.proxmox.com
> Envo
gt;
> Oh, so maybe you are on squeeze, but simple the /etc/issue has been replaced ?
>
> what do you see if you do a :
>
> dpkg -l
>
>
> do you see squeeze packages or wheezy packages ?
all are squeeze but i don't know how that one get replaced by a wheezy
str
ages from PVE to squeeze.
Stefan
> - Mail original -
>
> De: "Michael Rasmussen"
> À: pve-devel@pve.proxmox.com
> Envoyé: Mercredi 3 Juillet 2013 18:10:20
> Objet: Re: [pve-devel] Magic auto upgrade?
>
> On Wed, 03 Jul 2013 15:12:19 +0200
> Stefan
sources.list and stuff still points to squeeze.
Also all packages are Squeeze based. Does proxmox write to /etc/issue?
Stefan
Am 03.07.2013 15:05, schrieb Stefan Priebe - Profihost AG:
> Hello list,
>
> i've a local PVE Image which i use to install my PVE machines. I boot
&g
Hello list,
i've a local PVE Image which i use to install my PVE machines. I boot
this one every night do an
apt-get update && apt-get dist-upgrade
This one has now upgraded from Debian Squeeze to Wheezy? But why? I
wanted to stay a while still at squeeze.
My other Debian images nor my productiv
Hi,
but pve-libspice-server.git does not build for me.
checking for PIXMAN... no
configure: error: Package requirements (pixman-1 >= 0.17.7) were not met:
No package 'pixman-1' found
Consider adjusting the PKG_CONFIG_PATH environment variable if you
installed software in a non-standard prefix.
Hi,
something i've seen all the time long and tried to solve by myself but i
gave up ;-(
I've problems with the PVERessourceTree with its gridscroller. Sometimes
i don't see new machines on the button the scrollbar just stops at the
old one.
Also sometimes the scrolbar does not work at all. I ca
Hi List,
i updated to latest PVE manager (39a67abca36909bf66a0636c8b10f210a17e03d4).
Since then i only get:
[13:54:37.511] TypeError: stateinit is undefined
pointing to this line:
if(stateinit.timeframe){
context:
Ext.define('PVE.panel.RRDView', {
extend: 'Ext.panel.Panel',
gt;>
>> - Mail original -
>>
>> De: "Dietmar Maurer"
>> À: "Alexandre DERUMIER"
>> Cc: "Stefan Priebe - Profihost AG" , pve-
>> de...@pve.proxmox.com
>> Envoyé: Vendredi 7 Juin 2013 14:41:32
>> Objet: RE: [p
uot;Dietmar Maurer"
> À: "Alexandre DERUMIER"
> Cc: "Stefan Priebe - Profihost AG" ,
> pve-devel@pve.proxmox.com
> Envoyé: Vendredi 7 Juin 2013 14:41:32
> Objet: RE: [pve-devel] tasks / ressources
>
>> Maybe a tunnable value for auto-refresh time
Hi,
while debugging why PVE GUI is so slow over VPN / slow network
connection i've seen that we update:
- tasks
- ressources
- current
every second. Is this really needed? current is pretty small but tasks
and ressources has around 20kb for me.
Stefan
Am 06.06.2013 06:40, schrieb Dietmar Maurer:
> OK, fixed in git - please can you test again?
working nicely now!
Thanks!
>>> Right now when i try to roll back to them default qemu machine type
>>> from qemu 1.4 has to be set.
>>>
>>> That was the reason for using:
>>> my $minimal_qemu_machine =
Am 06.06.2013 06:08, schrieb Dietmar Maurer:
>> Right now when i try to roll back to them default qemu machine type from
>> qemu 1.4 has to be set.
>>
>> That was the reason for using:
>> my $minimal_qemu_machine = "pc-i440fx-1.4";
>
> But this only works if the snapshots are really generated with
Am 05.06.2013 11:09, schrieb Dietmar Maurer:
>> Looks good will test soon.
>>
>> Why do we have to use type= in kvm command?
>
> Because that is the correct syntax?
>
> # kvm -help|grep machine
> -machine [type=]name[,prop[=value][,...]]
>
It does not work for me as you've removed my min qemu ve
Am 05.06.2013 10:38, schrieb Dietmar Maurer:
>
>>> If you want, I can try to restructure the patch myself - maybe I see the
>> problem then.
>>
>> Thanks that would be really nice i'm under heavy load right now ;-(
>
> done - please can you test?
>
Looks good will test soon.
Why do we have to
Am 04.06.2013 12:48, schrieb Dietmar Maurer:
>>> So we have a snapshot using the actual config which has a wrong value
>>> for machine. Should then the snapshot simply overwrite that machine
>>> value? I wanted to archive that we may have to options in the feature.
>>
>> Sorry - I do not understand
Am 04.06.2013 11:33, schrieb Dietmar Maurer:
>> Yes, I was thinking of doing this without adding new extra api http request,
>> reusing current call to status api.
>
> Beside, auto-refresh can be confusing, and it is not always a good idea to do
> it.
>
> I agree that it would be useful in the s
Am 04.06.2013 10:45, schrieb Dietmar Maurer:
>> Yes and no the idea was that if we may have later a machine configuration in
>> .conf this one still forces the machine to another value.
>>
>> Example:
>> machine config option: pc-q35-1.4
>> You start the VM.
>> change machine config option: pc-q35-
Right now pve is already doing a lot of stuff regularly being very slow and
unresponsive on not so fast lab connections. Adding another 1s reload would
make it even more slow.
I would prefer something like reload after action regardless of a change or
not. So no checksums needed and only reload
Am 01.06.2013 07:38, schrieb Dietmar Maurer:
I guess we just need to add a 'Details' button to TaskProgress with
the following functionality:
1. Close TaskProgress
2. Open TaskViewer
looks quite simple. Would that fit your needs?
>>
>> Yes - but i can't do that.
>
OK,
is there any recomand way to split a patch into subpatches with git?
Stefan
Am 03.06.2013 10:53, schrieb Dietmar Maurer:
> I would prefer to split that patch into several pieces:
>
> a.) add support for --machine configuration option
> b.) add 'machine' option to vm_start
> c.) use --machine
Am 03.06.2013 10:35, schrieb Dietmar Maurer:
>> - this allows us to restore old snapshots and migrate machines running older
>> qemu versions
>
> What changed since last version?
>
Sorry nothing i just was not sure if you got that version. Normally
you're fast in replying ;-)
Stefan
Yes that's the way I've already used don't know if you want to have the patch
Stefan
This mail was sent with my iPhone.
Am 31.05.2013 um 16:14 schrieb Dietmar Maurer :
> Yes, OK for me. Will you provide a patch or do you want me to do
> that
change?
uhu i really would li
Am 31.05.2013 15:47, schrieb Dietmar Maurer:
>>> Yes, OK for me. Will you provide a patch or do you want me to do that
>> change?
>>
>> uhu i really would like can you give me a hint where to start looking in
>> ExtJS?
>
> Simply use TaskViewer class instead of TaskProgress.
Sorry yeah sure but h
Am 31.05.2013 11:45, schrieb Dietmar Maurer:
>> I only found rbd which outputs useful progress. Is it possible to change it
>> only
>> for rbd right now?
>
> Yes, OK for me. Will you provide a patch or do you want me to do that change?
uhu i really would like can you give me a hint where to star
Am 31.05.2013 11:43, schrieb Dietmar Maurer:
> fixed:
> https://git.proxmox.com/?p=qemu-server.git;a=commit;h=2fc6bc17229563d0ba87891103fe3a25da68b189
>
> please can you test?
>
>> i discovered that if block-job-complete fails pve does not call block-job-
>> cancel. So the whole image stays open
Am 31.05.2013 11:21, schrieb Dietmar Maurer:
>> diff --git a/PVE/Storage/RBDPlugin.pm b/PVE/Storage/RBDPlugin.pm index
>> e147346..cf61ab5 100644
>> --- a/PVE/Storage/RBDPlugin.pm
>> +++ b/PVE/Storage/RBDPlugin.pm
>> @@ -103,7 +103,7 @@ sub rbd_ls {
>> };
>>
>> eval {
>> -run_command(
Am 31.05.2013 10:25, schrieb Dietmar Maurer:
>> Am 31.05.2013 10:12, schrieb Dietmar Maurer:
At least a rollback with rbd takes up to 2 minutes and there is the
nice progress output.
>>>
>>> So we want to show TaskViewer for that single case? Or for all rollback
>> actions?
>>
>> I don't
Am 31.05.2013 10:09, schrieb Dietmar Maurer:
>> Am 31.05.2013 10:04, schrieb Dietmar Maurer:
it takes a lot longer until i see output. It seems the output is heavily
>> buffered.
>>>
>>> probably a bug then (but I cant see whats wrong).
>>
>> Me too but it works great when doing print STDERR s
Am 31.05.2013 10:12, schrieb Dietmar Maurer:
>> At least a rollback with rbd takes up to 2 minutes and there is the nice
>> progress output.
>
> So we want to show TaskViewer for that single case? Or for all rollback
> actions?
I don't know how snapshots work with the other storages and if they
Am 31.05.2013 10:07, schrieb Michael Rasmussen:
> On Fri, 31 May 2013 08:04:30 +
> Dietmar Maurer wrote:
>
>>
>> probably a bug then (but I cant see whats wrong).
>>
> Is perl's internal output buffering disabled ($| = 1)?
Haven't checked but then print STDERR shift should als get buffered.
Am 31.05.2013 10:06, schrieb Dietmar Maurer:
>> The same applied for creating and rolling back snapshots. It would be nice to
>> see the task manager here too instead of the progress bar.
>
> Snapshot are usually quite fast (a few seconds), and there is no useful
> output in most cases.
At least
Am 31.05.2013 10:04, schrieb Dietmar Maurer:
>> it takes a lot longer until i see output. It seems the output is heavily
>> buffered.
>
> probably a bug then (but I cant see whats wrong).
Me too but it works great when doing print STDERR so i kept that one.
Stefan
__
501 - 600 of 1027 matches
Mail list logo
