VictorSTS's latest activity
-
VictorSTS replied to the thread Ceph squid OSD crash related to RocksDB ceph_assert(cut_off == p->length).Thanks for the heads up. Pretty sure most where created with Ceph Reef except a few that got recreated recently with Squid 19.2.3. I'm aware of that bug, but given that I don't use EC pools (Ceph bugreport mentions it seems to only happen on OSD... -
VictorSTS reacted to fstrankowski's post in the thread Ceph squid OSD crash related to RocksDB ceph_assert(cut_off == p->length) with
Like.
Best of luck to you * fingers crossed *. I had to rebuild the whole cluster in my clients case and fix ceph by manually restoring placement groups - which was a pain. -
VictorSTS replied to the thread Ceph squid OSD crash related to RocksDB ceph_assert(cut_off == p->length).@fstrankowski I'm fully aware about the risks of OSD being full and know how to deal with that, but in any case an OSD should break because of that ;) Definitely fragmentation has an impact on this and will watch it more closely from now on...
-
VictorSTS reacted to fstrankowski's post in the thread Ceph squid OSD crash related to RocksDB ceph_assert(cut_off == p->length) with Like.
Initially i'd like to raise concerns about the amount of available storage already beeing in use. By default CEPH doesnt allow more then 80% so you'd have to take precautions really soon while taking these concerns into consideration. I'd highly...
-
VictorSTS replied to the thread Ceph squid OSD crash related to RocksDB ceph_assert(cut_off == p->length).In short: if Ceph warns you about something, do something about it. Read the full bugreport and found this comment [1]: "This issue seems to mostly affect disks which were heavily fragmented.". Mine are and in fact I have some warnings related...
-
VictorSTS posted the thread Ceph squid OSD crash related to RocksDB ceph_assert(cut_off == p->length) in Proxmox VE: Installation and configuration.PVE8.4.14 + Ceph 19.2.3, 3 node cluster. All disks are PCIe NVMe. Different pools, some with zstd compression enabled. I'm seeing OSD crashing lately with the same failure. Journal shows that it is unable to properly run RocksDB with an assert...
-
VictorSTS reacted to Zappes's post in the thread I want nothing more than encrypted push sync with Like.
Sure, no problem: https://bugzilla.proxmox.com/show_bug.cgi?id=7251
-
VictorSTS replied to the thread I want nothing more than encrypted push sync.@Zappes please, link the Bugzilla you mention here [1]. I would love to be able pull (not push) encrypted syncs where the source is unencrypted for any reason and destination PBS must store encrypted backups. [1]...
-
For reference: https://bugzilla.proxmox.com/show_bug.cgi?id=6195
-
VictorSTS posted the thread Questions about PBS local sync in Proxmox Backup: Installation and configuration.Say I have a namespace called "PVE" where PVE cluster stores it's backups. In the same datastore, I have another namespace called "DELETED". When a VM is deleted from the PVE cluster, I move it's backups from namespace "PVE" to "DELETED" in order...
-
VictorSTS reacted to spirit's post in the thread 40 node prod cluster restarts when joining a new node or removing. with Like.
from last corosync release https://github.com/corosync/corosync/releases "A new option (totem.ip_dscp) is available to configure DSCP for traffic prioritization. Thanks to David Hanisch for this great improvement." could be interesting for...
-
VictorSTS replied to the thread Ceph 20.2 Tentacle Release Available as test preview and Ceph 18.2 Reef soon to be fully EOL.Just follow the docs: https://pve.proxmox.com/wiki/Ceph_Squid_to_Tentacle
-
VictorSTS replied to the thread Nested PVE (on PVE host) Kernel panic Host injected async #PF in kernel mode.Seeing this issue with nested PVE on PVE on EPYC 9124 CPU. Host has SWAP and KSM enabled, although it has plenty of free memory (100GB of 512GB). This cluster currently runs PVE8.4.14 and runs other workloads too besides nested PVE VMs, with...
-
VictorSTS replied to the thread Backup taken on a last day of a month as monthly backup.I've already discussed about it here [1]. When I really have the need to keep the last backup of the last day of the month, I use this on PBS: Create a namespace "MONTHLY" Create a daily sync job that runs "if day is 28 to 31 of each month at...
-
VictorSTS replied to the thread [noVNC] Endless loading spinner since Firefox 147.0.3.It works ok for me on v147.0.2 and v147.0.3, both on Linux, accesing noVNC from PVE9.0.x, PVE9.1.x and PVE8.x. Try it from a VM to discard some issue with cache/whatever on your current PC.
-
VictorSTS reacted to t.lamprecht's post in the thread Call for Evidence: Share your Proxmox experience on European Open Digital Ecosystems initiative (deadline: 3 February 2026) with Like.
Dear Proxmox-Community, we are asking for your support. The European Commission has opened a Call for Evidence on the initiative European Open Digital Ecosystems, an initiative that will support EU ambitions to secure technological sovereignty...
-
VictorSTS reacted to fba's post in the thread Network interface pinning inconsistencies: ISO installer vs pve-network-interface-pinning generate with Like.
This was meant as example for adding an interface after the pinning of the installer was used.
-
VictorSTS replied to the thread Backup migration between different namespaces.As stated above, it can be done with sync jobs + manual deletion from the source namespace. Currently, local sync jobs only allow to sync between different datastores (I'm still wondering why). You will have to add that PBS itself as a remote so...
-
VictorSTS replied to the thread Network interface pinning inconsistencies: ISO installer vs pve-network-interface-pinning generate.Thanks, but that's unrelated to the issue I described. Problem is that I end up with two .link files for nic0 because pve-network-interface-pinning doesn't recognize there's already a pinned name due to different .link file naming scheme. Both...
-
VictorSTS reacted to UdoB's post in the thread proxmox ceph performance with consumer grade samsung ssd with Like.
That's worst design possible. Beside the too cheap devices... Look at one node: when one OSD fails, the other one on the same node has to take over the data from the dead one. It can not be sent to another node because there are already copies...