same for me, removing the HAProxy config let me access the web gui again. Weird behaviour. Why this is happening now after the Uprade to 4?
Proxmox Backup Server 4.0 released!
- Thread starter t.lamprecht
- Start date
I am Running Proxmox Backup 3.4. I want to install PVE 9 on a new server. Will I be able to restore Backups coming from a PVE8.4 environment using Backup 3.4? I would like to stay with 3.4 for while but also want to have a PVE 9 test environment.
Yes, PVE 9.0 is compatible with PBS 3.4, and restoring backups created on 8.4 works as well, see also from the initial announcement:
Hey there.
First of all, great release of PBS4 and also the new S3 feature. I was able to connect it successfully to an AWS S3 bucket and currently it works smooth for me.
However, I don't get it running on Synology C2 storage, which is also an S3 compatble storage and much cheaper than AWS S3. Trying to connect the S2 storage, I face the following issues:
Any idea or are there any plans to implement Synology C2.
Kind regards,
Dirk
First of all, great release of PBS4 and also the new S3 feature. I was able to connect it successfully to an AWS S3 bucket and currently it works smooth for me.
However, I don't get it running on Synology C2 storage, which is also an S3 compatble storage and much cheaper than AWS S3. Trying to connect the S2 storage, I face the following issues:
Code:
<?xml version='1.0' encoding='UTF-8'?>
<Error><Code>NotImplemented</Code><Message>Conditional object PUTs are not supported.</Message><RequestId>tx3145a330181342cf80f12-006894a8da</RequestId></Error>
<?xml version='1.0' encoding='UTF-8'?>
<Error><Code>NotImplemented</Code><Message>Conditional object PUTs are not supported.</Message><RequestId>txdeb0db15f8ef4f9dad49d-006894a8db</RequestId></Error>
<?xml version='1.0' encoding='UTF-8'?>
<Error><Code>NotImplemented</Code><Message>Conditional object PUTs are not supported.</Message><RequestId>tx313cf7a43b8f4fcaaaba4-006894a8db</RequestId></Error>
TASK ERROR: access time safety check failed: failed to upload chunk to s3 backend: chunk upload failed: unexpected status code 501 Not Implemented
Error: task failed (status access time safety check failed: failed to upload chunk to s3 backend: chunk upload failed: unexpected status code 501 Not Implemented)Any idea or are there any plans to implement Synology C2.
Kind regards,
Dirk
Please try with the provider quirks set to
Skip If-None-Match header, which you can find in the advanced settings of the S3 endpoint create/edit window.That should fix this provider specific limitation.
I’m experencing too the web ui crash after upgrading to 4. My pbs is a VM virtualized in pve behind traefik reverse proxy.
Looking up-thread to the posts on ha-proxy, it looks like there might be an issue if your proxy is doing the equivalent of whatever the "check" flag does in ha-proxy.
See: https://forum.proxmox.com/threads/proxmox-backup-server-4-0-released.169306/post-789268
I'm not familiar with ha-proxy or Traeffik, but maybe try to figure out what that failing ha-proxy config option does and see if you can map it to your Traeffik config?
This should be fixed with proxmox-backup-server 4.0.12-1 which was just uploaded to pbs-test repo (some CDN nodes are still syncing, so might take a few more minutes). At least it fixes one reproducer here and in any case an issue that could happen on connection acceptance and seemingly was even much harder to hit in PBS 3.
That version also contains a fix for OIDC realm login and the HttpOnly cookies authentication flow.
There is no fixed timeline, but as the changes between the last version and this are small, they only include three targeted fixes, and as we got already quite a bit of positive feedback validating these fixes, we fast tracked the package updates and moved these changes now already.
short question:
in the PBS 3to4 upgrade wiki it is stated that minimum free space on "/" should be 10GB.
Can I upgrade with only 6GB free space or will I run into problems...?
in the PBS 3to4 upgrade wiki it is stated that minimum free space on "/" should be 10GB.
Can I upgrade with only 6GB free space or will I run into problems...?
It might (!) but we did not really test that, might work and depends a bit on what (additional) packages you installed, so cannot tell you anything for certain for this being untested and also potentially setup specific.
Thanks for the quick reply!
No extra packages are installed, just PBS 3.x
However, PBS is installed as a vm on a Synology NAS. But that runs just fine.
(tempted to do a vm snapshot, but there are lots of GB's to "snapshot" and not much NAS diskspace left).
Just got the update on the community repo. PBS is now working on tailscale!
Thanks
I just pulled in the 4.0.12-1 update too for my LXC container instance
Can confirm HAProxy with check directive is no longer causing the proxmox-backup-proxy to reset connections now.
All working as expected now.
Can confirm HAProxy with check directive is no longer causing the proxmox-backup-proxy to reset connections now.
All working as expected now.
Hey Chris.
Thanks for getting back to me. It work by setting the quirks option.
Thanks again!
Kind regards,
Dirk
I upgraded only my PBS bare metal server to 4 only a couple of days ago (current 4.0.12). Followed the guide, upgrade went fine with the exception of getting a blank web GUI page trying to login, luckily SSH was working. I had to rem out everything having to do with enterprise repos, then apt update again to get hundreds of MB of updates that it didn't get and apply. Odd, but whatever...
So now, PBS is taking DOUBLE the memory usage at idle after a nightly backup occurs. It would stay at around 46% usage, now I sit at 93%. You can see before the updated was almost at 50%, reboot, backup runs and sits at 93%. Reboot, it clears until another backup runs. The backups are all doing as they ahve done, and my Datastore is, and has been ZFS on PBS. Nothing has changed in my environment with the exception of updating PBS. Top doesn't really give me anything, the two top services are the proxmox-backup-proxy and -api services.
So now, PBS is taking DOUBLE the memory usage at idle after a nightly backup occurs. It would stay at around 46% usage, now I sit at 93%. You can see before the updated was almost at 50%, reboot, backup runs and sits at 93%. Reboot, it clears until another backup runs. The backups are all doing as they ahve done, and my Datastore is, and has been ZFS on PBS. Nothing has changed in my environment with the exception of updating PBS. Top doesn't really give me anything, the two top services are the proxmox-backup-proxy and -api services.
Attachments
Last edited:
And why is this a problem? Not used RAM is wasted. But you can limit the ZFS ARC Cache:
https://pbs.proxmox.com/docs/sysadmin.html#limit-zfs-memory-usage
I'm not necessarily saying it's a problem, but it's an irregular issue from what's been normally running, then to update and seeing noticeable differences is worth mentioning, that's why I posted it. That can potentially also crash other PBS users out there if they are running out of memory because of this.
Even in the documentation you linked:
Last edited: