Search results

Proxmox VE kann mit vielen unterschiedlichen Storages umgehen. Am besten schaust mal in die Doku. :) https://pve.proxmox.com/pve-docs/chapter-pvesm.html Qcow2 ist ein Image Format un muss bei der Erstellung oder bei Migration/Move Disk spezifisch ausgewählt werden. Aber performanter ist wohl...

Gewagte Aussage. :) Ext4 ist ein Filesystem das keine Snapshots kann. Hier muss entweder Qcow2 Image Dateien verwendet werden. Oder LVM(-thin) als Storage. Du kannst auch auf dem LVM ein LV erstellen und ein Dateisystem darauf einrichten. Dann sind auch beide Optionen möglich.

pct help, will list all the commands and options for the tool. With eg. pct help listsnapshots and gather the name of the snapshot in question. Then you can run the delsnapshot. And with man pct you can get the help page for it as well.

See pct help delsnapshot. :)

Ceph will not work without the MONs. You will need to provide more space to the partition where the store.db is located.

Run that command with -force if the snapshot doesn't exist anymore. root@pve1:~# pct delsnapshot <vmid> <snap> -force

My test VM starts with the above settings. Or do you mean the OS doesn't boot?

Well, then you may look into the openipmi and query the values from there.

I tested this on my cluster and a VM started without issue. What's the pveversion -v of your node? And the qm config <vmid>?

The autoscaler uses a factor of 3x by wich it will warn about an over-/under provisioning of PGs in a pool. The factor is caused by fill level of the pool. https://docs.ceph.com/en/octopus/rados/operations/placement-groups/#viewing-pg-scaling-recommendations

On of these pools has more PGs than the autoscaler thinks is necessary. Hence the warning. This is regardless of the actual amount of PGs on an OSD. Though 225 PGs is already a very high number.

Yup, there seems to be no module for that one in the kernel. The ipmi interface (openipmi) should still provide some additional data. But Fujitsu also has its own software and plugins, maybe these may help. https://download.ts.fujitsu.com/prim_supportcd/SVSSoftware/ Otherwise you can disable...

Did you run sensors-detect? I am missing the fan speeds in the output. Also you can check the data through IPMI, Fujitsu usually has a good load of sensor data there as well.

Yes, but usage is not performance. To directly check perf counters, run ceph daemon osd.<ID> perf dump or deploy a monitoring through the MGR.

Please update to the latest Proxmox VE and post a new thread if the issue persists.

Does a manual listing on the cli work? You will need to specify the MONs, user and key by hand.

/var/log/syslog, dmesg, /var/log/ceph/*.log

Yes, but your cluster is Nautilus as well. Hence, why to look in the logs (any / all of them), to see what might be logged that can explain it. After restart of the MONs, they are listing on both ports, assuming the msgrv2 protocol has been activated. Is there a firewall in between?

Can you restart on of the OSDs and check the state and log of it? No. That is not how it works. :)

The easiest is to destroy & create the OSD again. You can do it a host at a time. But if you want to minimize the impact, then best reweight the OSD first to 0. This will cause Ceph to distribute the data on that OSD to the others on that node. Once it is empty a destroy & create will just...