[pve-devel] updated 3.10.0 kernel
Just uploaded a 3.10.0 kernel update top pvetest: pve-kernel-3.10.0-5-pve_3.10.0-19_amd64.deb Please test, - Dietmar ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] updated 3.10.0 kernel
Thanks,I'll try it tomorrow - Mail original - De: Dietmar Maurer diet...@proxmox.com À: pve-devel@pve.proxmox.com Envoyé: Mercredi 15 Octobre 2014 08:23:51 Objet: [pve-devel] updated 3.10.0 kernel Just uploaded a 3.10.0 kernel update top pvetest: pve-kernel-3.10.0-5-pve_3.10.0-19_amd64.deb Please test, - Dietmar ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
[pve-devel] pveproxy issue
Hi Dietmar, I am using Proxmox (great product) in a lab context for my students. Maybe you remember our mail exchange last year where we had an issue with identical mac address generation when all of my students created a new machine at the same time. You bug-fixed the issue very quickly. Thanks again. This year I have a new issue. Before starting this year's course I updated Proxmox to the latest version. When the students (11 in all) are working all together on the Proxmox GUI, pveproxy hangs after a quite a short time and they get Communication failure(0), Loading... Server Offline? or Time out messages. I have to restart the pveproxy service to get back to a working GUI. All the machines are working fine when accessing them via noVnc or remote desktop so the problem is only related to pveproxy. Observations: I did not have that problem last year. Is there any parameters related to pveproxy that got changed due to the upgrade? I discovered the keep-alive value in /usr/bin/pveproxy, and I tried values from 10 to 900 (default is 100), with no positive effect. I installed nginx in a container. Accessing the Proxmox GUI via nginx improved the situation a lot. I still get same hangs but they appear a lot later, less often and recover after a while. I do not have to restart pveproxy. Unfortunately I have no solution yet to acces the novnc this way. Is there anything else I can try? Guy ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
[pve-devel] Add functionalities
Hello, we plan to add some functionnalities to qm command : qm copy_disk vmid disk storage [OPTIONS] will act seamlessly as move disk, but won't change active disk at the end qm change_disk vmid disk storage [OPTIONS] will allow to change disk used by Proxmox Our needs : We have 3 Proxmox servers, that share a unique SAN Lvm volume thru iSCSI IBM DS3300. In case of NAS system failure, we don't want to wait for restore... So we plan to prediodicaly copy VM disks to a NAS volume. (qm copy_disk) When a failure occurs, we have just to change disk, for the copy (qm change_disk). We plan also to make this behaviour available from Web interface (as schedule backup). We will use Developer Documentation and publish it when available. Please let us know if something similar already exist, or you are in plan to deliver it. We want to push it as standard feature of Proxmox. Best regards Serge NOEL ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] pveproxy issue
When the students (11 in all) are working all together on the Proxmox GUI, pveproxy hangs after a quite a short time and they get Communication failure(0), Loading... Server Offline? or Time out messages. I have to restart the pveproxy service to get back to a working GUI. Is there any error message in /var/log/syslog? Is there a reliable way to reproduce the behavior? ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Add functionalities
In case of NAS system failure, we don't want to wait for restore... So we plan to prediodicaly copy VM disks to a NAS volume. (qm copy_disk) When a failure occurs, we have just to change disk, for the copy (qm change_disk). So you don’t want to use normal backup, because restore takes too long? ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Add functionalities
Exactly, we are working on disaster scenarii and we want to find quick solution in case of SAN failure. In this case, we imagine that SAN will be unavailable, and we hope to be able to restart with NAS (iSCSI) as replacement target, if we are able to make copy of all VM, (it seems OK, i tried to copy lvm volume group with dd : it's works) during normal operations. Doing so, in case of failure, all we have to do is : to say to Proxmox to use disk from different iSCI (or NFS) target. Regards 2014-10-15 16:41 GMT+02:00 Dietmar Maurer diet...@proxmox.com: In case of NAS system failure, we don't want to wait for restore... So we plan to prediodicaly copy VM disks to a NAS volume. (qm copy_disk) When a failure occurs, we have just to change disk, for the copy (qm change_disk). So you don’t want to use normal backup, because restore takes too long? ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Add functionalities
On Wed, 15 Oct 2014 17:02:25 +0200 Serge NOEL serge.noel2...@gmail.com wrote: Exactly, we are working on disaster scenarii and we want to find quick solution in case of SAN failure. In this case, we imagine that SAN will be unavailable, and we hope to be able to restart with NAS (iSCSI) as replacement target, if we are able to make copy of all VM, (it seems OK, i tried to copy lvm volume group with dd : it's works) during normal operations. Doing so, in case of failure, all we have to do is : to say to Proxmox to use disk from different iSCI (or NFS) target. Isn't that dangerous? I mean your approach will surely cause the loss of data since all changes between the time of creation of the fall-back disk an the incident will be lost! If you need uptime like 99.999% you should drop you single point of failure SAN an move on to clustered storage since clustered storage is the only viable solution to your high demand on uptime. -- Hilsen/Regards Michael Rasmussen Get my public GnuPG keys: michael at rasmussen dot cc http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xD3C9A00E mir at datanom dot net http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xE501F51C mir at miras dot org http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xE3E80917 -- /usr/games/fortune -es says: This fortune intentionally says nothing. pgp8E1TGud6Ia.pgp Description: OpenPGP digital signature ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Add functionalities
Sure it is, but : we will make one copy per week, like a full backup we will make backup of datas in a classical way (rsync or something similar...) So we plan to use it as a far recovery process (wich provide a base machine). It doesn' t allow us to bypass backup, just a way to quick recover services... More, only file server has changing files, others are TSE they stay relatively stable on time, and loosing one week of change is better, that being blocked during one day... So we are Everything is a couple of choices, nor the best, nor the worst Ideas are welcome... Serge 2014-10-15 17:22 GMT+02:00 Michael Rasmussen m...@datanom.net: On Wed, 15 Oct 2014 17:02:25 +0200 Serge NOEL serge.noel2...@gmail.com wrote: Exactly, we are working on disaster scenarii and we want to find quick solution in case of SAN failure. In this case, we imagine that SAN will be unavailable, and we hope to be able to restart with NAS (iSCSI) as replacement target, if we are able to make copy of all VM, (it seems OK, i tried to copy lvm volume group with dd : it's works) during normal operations. Doing so, in case of failure, all we have to do is : to say to Proxmox to use disk from different iSCI (or NFS) target. Isn't that dangerous? I mean your approach will surely cause the loss of data since all changes between the time of creation of the fall-back disk an the incident will be lost! If you need uptime like 99.999% you should drop you single point of failure SAN an move on to clustered storage since clustered storage is the only viable solution to your high demand on uptime. -- Hilsen/Regards Michael Rasmussen Get my public GnuPG keys: michael at rasmussen dot cc http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xD3C9A00E mir at datanom dot net http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xE501F51C mir at miras dot org http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xE3E80917 -- /usr/games/fortune -es says: This fortune intentionally says nothing. ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] Add functionalities
Hi, a better solution would be to work with lvm snapshots and deltas. You just have to move the small delta. And therefore you can do this every hour or even more aften.. i dont know i fit is also possible with vzdump but it is also capable of making deltas. But ist not the same like applieng the delta direct tot he storage target like on lvm... Like this it sould be possible to archieve something like vmware has allready. Some tool like this would be actually cool in proxmox for some guis with smaller environments. For bigger ones and maybe for you using ceph instead of your san would be better -Ursprüngliche Nachricht- Von: pve-devel [mailto:pve-devel-boun...@pve.proxmox.com] Im Auftrag von Michael Rasmussen Gesendet: Mittwoch, 15. Oktober 2014 17:23 An: pve-devel@pve.proxmox.com Betreff: Re: [pve-devel] Add functionalities On Wed, 15 Oct 2014 17:02:25 +0200 Serge NOEL serge.noel2...@gmail.com wrote: Exactly, we are working on disaster scenarii and we want to find quick solution in case of SAN failure. In this case, we imagine that SAN will be unavailable, and we hope to be able to restart with NAS (iSCSI) as replacement target, if we are able to make copy of all VM, (it seems OK, i tried to copy lvm volume group with dd : it's works) during normal operations. Doing so, in case of failure, all we have to do is : to say to Proxmox to use disk from different iSCI (or NFS) target. Isn't that dangerous? I mean your approach will surely cause the loss of data since all changes between the time of creation of the fall-back disk an the incident will be lost! If you need uptime like 99.999% you should drop you single point of failure SAN an move on to clustered storage since clustered storage is the only viable solution to your high demand on uptime. -- Hilsen/Regards Michael Rasmussen Get my public GnuPG keys: michael at rasmussen dot cc http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xD3C9A00E mir at datanom dot net http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xE501F51C mir at miras dot org http://pgp.mit.edu:11371/pks/lookup?op=getsearch=0xE3E80917 -- /usr/games/fortune -es says: This fortune intentionally says nothing. ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel
Re: [pve-devel] pveproxy issue
No unfortunately not, some times only a warning: Oct 15 09:26:04 misspiggy pveproxy[996374]: WARNING: proxy detected vanished client connection Oct 15 09:26:04 misspiggy pveproxy[996374]: WARNING: proxy detected vanished client connection Oct 15 09:28:22 misspiggy pveproxy[994305]: received terminate request Oct 15 09:28:22 misspiggy pveproxy[994305]: worker 994306 finished Oct 15 09:28:22 misspiggy pveproxy[994305]: worker 994307 finished Oct 15 09:28:22 misspiggy pveproxy[994305]: worker 996374 finished Oct 15 09:28:22 misspiggy pveproxy[994305]: server closing Oct 15 09:28:24 misspiggy pveproxy[997353]: starting server Oct 15 09:28:24 misspiggy pveproxy[997353]: starting 3 worker(s) Oct 15 09:28:24 misspiggy pveproxy[997353]: worker 997354 started Oct 15 09:28:24 misspiggy pveproxy[997353]: worker 997355 started Oct 15 09:28:24 misspiggy pveproxy[997353]: worker 997356 started Oct 15 09:28:25 misspiggy pveproxy[997356]: WARNING: proxy detected vanished client connection Oct 15 09:29:54 misspiggy pveproxy[997353]: worker 997354 finished Oct 15 09:29:54 misspiggy pveproxy[997353]: starting 1 worker(s) I don't have a reliable way to reproduce the behavior. Students connect and pveproxy hangs after a relative short time. Guy On 15/10/2014 16:09, Dietmar Maurer wrote: When the students (11 in all) are working all together on the Proxmox GUI, pveproxy hangs after a quite a short time and they get Communication failure(0), Loading... Server Offline? or Time out messages. I have to restart the pveproxy service to get back to a working GUI. Is there any error message in /var/log/syslog? Is there a reliable way to reproduce the behavior? ___ pve-devel mailing list pve-devel@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-devel