as we're often migrate VMs between nodes and even clusters we have sometimes orphaned VM backups on our PBS hosts.
Did I miss something, or isnt it possible to prune by date?
We do weekly backups of VM an the VM gets migrated to another PVE cluster, so the backup of it is then on...
our PBS backups from PVE (both latest versions) are failing sometimes with "interrupted by signal" - but there is no further info in the logs, only this:
INFO: creating Proxmox Backup Server archive 'vm/5297/2021-02-25T22:30:00Z'
ERROR: Backup of VM 5297 failed - interrupted by signal...
we're currently testing PBS with PVE. Backups works so far, but when restoring VMs with XFS the VM is unbootable (see attached image) and always needs a xfs_repair in rescue. This is reproducible with different VMs and backups.
With ext4 its fine and without errors.
The backup setting...
we're currently in the process of trying to build proxmox-backup-client for CentOS 7 to evaluate and test the pbs beta.
We've worked through some build strategies and are currently trying a dynamically linked binary (glibc, libacl, libfuse3 etc.) w/ statically linked OpenSSL 1.1 which -...
with latest PVE 5.1 we have sometimes VMs (KVM) going down and need to be started manually during backup process. It appeas random across our VMs and hosts. We use NFS storage and "snapshot" mode with LZO compression.
Any idea whats wrong or how to fix this?
we've built a helper tool for PVE clusters called vman which comes with useful functions which PVE is missing, for example:
* auto balance VMs in a PVE cluster depending on used and free memory
* flushing an entire PVE node (i.e. for updates/reboots)
* flushing only HA VMs (for updates)...
We observed a strange issue, when using scheduled backups with vzdump (from local ssd storage, snapshot, lzo).
It seems when the NFS storage is under heavy network load or the uplink to the NFS node is "full", vzdump blocks the running VM to death.
We can see lots of "[sda] abort" messages...
have you ever tried to setup PVE through a Java emulated KVM from a Linux or Mac system? :-)
It can be real pain because you might have no chance to create the "@" sign, but this is required when installing.
We currently use as a workaround to copy (!) the "@" sign from the error message...
just wondering why the PVE Wiki mentions the unicast limit explicitly with "do not use it with more that 4 cluster nodes"?
An example in the corosync git repo (see here) has 16 nodes with unicast, can't believe they are wrong there?
Is there any proof that this 4 node limit does really...
is disk throttling (kvm) also throttling the backup speed (mode: snapshot)? We're seeing extremely slow backups from VMs where iops_rd and iops_wr is set to 100 each and we have no other explanation for this behavior.
The whole backups takes almot 24hours. Other lager VMs only...
we're using "pvesh" to gather different metrics from our VMs (for monitoring purposes) and it works really good.
One important thing whats missing and would be very useful is IOPS from each VM here or in any other way?
Because we're able to limit IOPS within the panel it seems the value of...
we just updated a node with "apt-get dist-upgrade" from a 3 node cluster as usual and it did reboot here:
How to avoid that?
After reboot with "dpkg --configure -a" all packages were proper installed.
we've got a 3 node PVE 4 (latest tools/updates) cluster with DRBD 9.
When cloning a VM, the state is always "Outd" on the 3rd node as follows:
102:vm-123-disk-1/0 Connected(3*) Seco(node1,node3)/Prim(node2) UpTo(node3)/Outd(node1)/UpTo(node2)
On the other 2 nodes everything is ok (UpTo)...
we've got a new HP DL180 with 2x240 GB SSD Cache (RAID1, HP SmartCache on Smart Array P840 Controller) and 10 spinning drives (RAID10).
But, when activated, pveperf is incredible slow:
CPU BOGOMIPS: 115186.44
HD SIZE: 24.48 GB...
we've got a problem with one of our proxmox nodes. ksoftirqd/0 eats here sometimes 100% CPU on Core 0 (system has dual xeon E5-2620v2, 256 GB RAM, RAID6) and network is then slow /w packet loss.
We already bootet with kernel options nohz=off and pcie_aspm=off but it did not help. Overall...