Search results

I agree with this assumption. The one should at least be warn before and upgrade. I'm facing the same issue with 50+ OSDs and have no idea how to sort it out I don't have another cluster to play with and found not much info how correctly destroy all OSDs on single node, wipe all disks (as well...

Check your multipath.conf file. Seems one more “}” bracket is missing at the end

Resolved, thanks

After a successful upgrade from PVE 5 to PVE 6 with Ceph the warning message "Legacy BlueStore stats reporting detected on ..." appears on Ceph monitoring panel Have I missed something during an upgrade or it's an expected behavior? Thanks in advance

Single connection will be always limited to the speed of single interface. LACP bond increase total throughput (read as sum of all connections).

Mine configs: root@pve2:~# cat /etc/network/interfaces # network interface settings; autogenerated # Please do NOT modify this file directly, unless you know what # you're doing. # # If you want to manage parts of the network configuration manually, # please utilize the 'source' or...

I'm facing almost the same issue with couple of setups after an upgrade to 5.4. Could you show network config and lspci output. Probably we could find out something in common

There is no dedicated net. But switch is not loaded (according to SNMP stats). And once again: everything was fine before an upgrade

Below how omping result looks like now: root@pve2:~# omping -c 600 -i 1 -q pve2 pve3 pve4A pve3 : waiting for response msg pve4A : waiting for response msg pve4A : joined (S,G) = (*, 232.43.211.234), pinging pve3 : joined (S,G) = (*, 232.43.211.234), pinging pve3 : given amount of query...

omping test now shows 60% drop(( it was not the case with 5.3 (I performed that tests on all cluster setups)

This morning I restarted corosync on all the nodes again. Cluster was forking for couple of minutes and than hanged May 15 09:40:10 pve1 systemd[1]: Starting Corosync Cluster Engine... May 15 09:40:10 pve1 corosync[24728]: [MAIN ] Corosync Cluster Engine ('2.4.4-dirty'): started and ready...

On another cluster I'm facing different issue but again after an upgrade to 5.4 Could you please take a look into: https://forum.proxmox.com/threads/proxmox-cluster-broke-at-upgrade.54182/#post-250102 I'm fully confident that my network switches are configured inline with PVE docs IGMP snooping...

Here is a log from the node that became a part of cluster and than left May 14 10:26:08 pve4A corosync[20900]: notice [MAIN ] Completed service synchronization, ready to provide service. May 14 10:26:08 pve4A corosync[20900]: [CPG ] downlist left_list: 0 received May 14 10:26:08 pve4A...

On the node that hangs I see root@pve2:~# systemctl status pve-cluster ● pve-cluster.service - The Proxmox VE cluster filesystem Loaded: loaded (/lib/systemd/system/pve-cluster.service; enabled; vendor preset: enabled) Active: active (running) since Mon 2019-05-13 21:49:07 MSK; 16min ago...

I'm facing exactly the same issue. Unfortunately reinstalling is not possible

unmounting "/mnt/pve/cephfs-backup" works (temporary) for all nodes but one. root@pve-node3:~# umount /mnt/pve/cephfs-backup umount: /mnt/pve/cephfs-backup: target is busy (In some cases useful info about processes that use the device is found by lsof(8) or fuser(1).)...

root@pve-node2:~# cat /etc/pve/storage.cfg dir: local disable path /var/lib/vz content images maxfiles 0 shared 0 zfspool: local-zfs disable pool rpool/data blocksize 8k content images nodes...

Yes, nothing has changed. No ideas so far( What has been checked - 3 older kernels (one - that has been used in similar environment, the only difference on that setup - IPoIB instead of 10Gbe on ring#0 - anyway ring#0 is working in both setups). All the nodes we rebooted (VMs were migrated)...

I'm facing the same issue with the latest 5.4 syslog May 7 14:44:27 pve-node4 pvestatd[2269]: A filesystem is already mounted on /mnt/pve/cephfs-backup May 7 14:44:27 pve-node4 pvestatd[2269]: mount error: exit code 16 May 7 14:44:37 pve-node4 pvestatd[2269]: A filesystem is already mounted...

[mon.HOSTNAME] debug mon = 0/5 correct?