Re: [PVE-User] I try hard but...

[Dmitry Petuhov]

Of coz. But serving virtual disks over gigabit ethernet is generally bad idea too. And stable 70 megabytes per second are much better than unstable 100. Also I think that you don't need to limit iops, just throughput. 26.10.2015 18:35, Gilberto Nunes пишет: But i thing that with this limitation, transfer 30 gb of mail's over network it talks forever, don't you agree?? 2015-10-26 13:25 GMT-02:00 Gilberto Nunes <gilberto.nune...@gmail.com>: Regard badnwidth limitation, you mean like this: irtio0: stg:120/vm-120-disk-1.qcow2,iops_rd=100,iops_wr=100,iops_rd_max=100,iops_wr_max=100,mbps_rd=70,mbps_wr=70,mbps_rd_max=70,mbps_wr_max=70,size=2000G 2015-10-26 11:25 GMT-02:00 Dmitry Petuhov <mityapetu...@gmail.com>: There's issue with NFS: if you try to send over it more than network can deal with (100-120 MBps for 1-gigabit), it imposes several-second pauses, which are being interpreted like hardware errors. These bursts may be just few seconds long to trigger issue. You can try to limit bandwidth in virtual HDD config to something like 60-70 MBps. This should be enough for 1-gigabit network. But my opinion is that it's better to switch to iSCSI. 26.10.2015 16:13, Gilberto Nunes пишет: BTW, all HD is SAS with gigaethernet between the servers... I already try with gigaethernet switch in order to isoleted the Proxmox and Storage from external (LAN) traffic.... Not works at all! 2015-10-26 11:12 GMT-02:00 Gilberto Nunes <gilberto.nune...@gmail.com>: HDD config is standard... I do not make any change... I wonder why I have others VM, as I said before, with Ubuntu, CentOS, Windows 7 and 2012, and work fine! Not of all have huge big files or a lot of connections, but they stand solid! But when require a lot access and deal with big files, here's came the devil! The VM just get IO error and die before 2 or 3 days... On both physical servers, not hight load. I check with HTOP and TOP as well. iostat doesn't show nothing wrong. In side the VM, a lot I/O error from time to time... IO goes high endeed, but is so expect because I am using imapsync to sync mails to old server to Zimbra Mail Server... But I do not expect the VM die with IO errors! It's so frustrating... :( 2015-10-26 11:04 GMT-02:00 Dmitry Petuhov <mityapetu...@gmail.com>: What's virtual HDD config? Which controller, which cache mode? I suppose it's bad idea to run KVM machines via NFS: it may produce big enough delays under high loads, which may look like timeouts on client side. If you want some network storage, iSCSI can be better chiose. 26.10.2015 12:48, Gilberto Nunes пишет: Admin or whatever is your name... I have more than 10 years deal with Unix Linux and Windows. I know what I done. To the others: Yes! Should be straightforward any way... Proxmox server is a PowerEdge R430 with 32 GB of memory. Storage is the same server. Both with SAS hard driver. Between this servers there's a cable in order to provide a gigaethernet link. In second server, I have ubuntu 15.04 installed with DRBD and OCFS mounted in /data FS. In the same server, I have NFS installed and server FS to Proxmox machine. In proxmox machine I have nothing, except a VM with Ubuntu 14.04 installed, where Zimbra Mail Server was deploy... Inside bothe physical servers, everything is ok... NO error in disc and everything is running smoothly. But, INSIDE THE VM HOSTED WITH PROXMOX, many IO errors!... This make FS corrputed in some point that make Zimbra crash! BTW, I will return Zimbra to a physical machine right now and deploy lab env for test purpose Best regards 2015-10-25 22:38 GMT-02:00 ad...@extremeshok.com <ad...@extremeshok.com>: Your nfs settings. Hire you people with the knowledge or up skill your knowledge. Sent from my iPhone > On 26 Oct 2015, at 1:33 AM, Gilberto Nunes <gilberto.nune...@gmail.com> wrote: > > Well friends... > > I really try hard to work with PVE, but is a pain in the ass... > Nothing seems to work.. > I deploy Ubuntu with NFS storage connected through direct cable ( 1 gb ) and beside follow all docs available in the wiki and internet, one single VM continue to crash over and over again... > > So I realise that is time to say good bye to Proxmox... > > Live long and prosper... > > > > > -- > > Gilberto Ferreira > +55 (47) 9676-7530 > Skype: gilberto.nunes36 > > _______________________________________________ > pve-user mailing list > pve-user@pve.proxmox.com > http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user _______________________________________________ pve-user mailing list pve-user@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user -- Gilberto Ferreira +55 (47) 9676-7530 Skype: gilberto.nunes36 _______________________________________________ pve-user mailing list pve-user@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user _______________________________________________ pve-user mailing list pve-user@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user -- Gilberto Ferreira +55 (47) 9676-7530 Skype: gilberto.nunes36 -- Gilberto Ferreira +55 (47) 9676-7530 Skype: gilberto.nunes36 _______________________________________________ pve-user mailing list pve-user@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user _______________________________________________ pve-user mailing list pve-user@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user -- Gilberto Ferreira +55 (47) 9676-7530 Skype: gilberto.nunes36 -- Gilberto Ferreira +55 (47) 9676-7530 Skype: gilberto.nunes36 _______________________________________________ pve-user mailing list pve-user@pve.proxmox.com http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user

_______________________________________________
pve-user mailing list
pve-user@pve.proxmox.com
http://pve.proxmox.com/cgi-bin/mailman/listinfo/pve-user