Latest activity
-
fstrankowski replied to the thread Filesystem choice for consumer SSD.The only reason coming to my mind would be an absurdly low amount of RAM. Anything else? -
UdoB reacted to Onslow's post in the thread Replication of a VM between two independent Proxmox clusters (LVM storage) with
Like.
@katti Depending on your requirements on a how long the services can be unavailable and on how much fresh data can be lost, a workaround could be frequent backups to PBS. And restoring to the other cluster when needed). Thanks to "dirty bitmaps"... -
fstrankowski replied to the thread Einführung von Einschränkungen / Kennzeichnungspflicht von KI-Beiträgen.Genau dabei helfen Diskussionen wie diese hier. Bleiben wir mal bei Proxmox dann können wir festlegen, dass die Modelle zu einem maßgeblichen Anteil durch eben dieses Forum hier befüttert werden. Daher ist es nicht nur in unserem, sondern...
-
KKingneutron replied to the thread Filesystem choice for consumer SSD.I would plan on replacing the consumer-level drives long-term. Definitely keep an eye on the wearout indicator. For speed you could try lvm-thin (allows snapshots) or XFS (faster than ext4) - and if you need snapshots on that, backup to PBS and...
-
Aalejandro85 posted the thread Need help - Active Snapshoot Notification in Proxmox VE: Installation and configuration.Hi team, I'm looking for a solution to send notifications via email (or any other method) when, for example, a VM has an active snapshot larger than 1GB. Is this possible to achieve? I'm trying to starting do something like this with help of...
-
fstrankowski reacted to *alx*'s post in the thread Einführung von Einschränkungen / Kennzeichnungspflicht von KI-Beiträgen with Like.
Wenn, dann sollte man die KI Beiträge am Anfang kennzeichnen. Dann kann ich mir das Lesen auch sparen. Wenn ich eine Antwort von einer KI haben will, dann befrage ich selber eine. Und wenn sich jemand nicht die Mühe macht seine Beiträge selber zu...
-
UdoB replied to the thread Einführung von Einschränkungen / Kennzeichnungspflicht von KI-Beiträgen."KI" (es gibt bisher keine künstliche Intelligenz! Es gibt nur Sprachmodelle.) ist gekommen um zu bleiben. Eine Kennzeichnungspflicht erscheint plausibel, ist aber weder mittel- noch langfristig durchsetzbar und vor allem nicht wirklich...
-
spirit replied to the thread Network Optimization for High-Volume UDP Traffic in PVE.250566pps is quite low, I mean , you should reach 1~2mpps for any each packetsize. I remember to reach easily 7~9gbit with 1core/thread with standard 1500mtu. (with epyc v3 3,5ghz and cpu forced to max frequency) -
UdoB reacted to fstrankowski's post in the thread Einführung von Einschränkungen / Kennzeichnungspflicht von KI-Beiträgen with Like.
Ich bitte um eine Diskussion/RFC zu o.g. Thematik. Mittlerweile nehmen die Beiträge welche erkenn- und nachweisbar zu nahezu ausschließlich KI-Content bestehen massiv überhand. Besteht die Möglichkeit dies einzuschränken und/oder eine...
-
UdoB replied to the thread Filesystem choice for consumer SSD.While I am definitely a ZFS fan there might be situations where LVM (-thin) is the recommended choice. Sadly this might be that rare occasion...
-
LLSX3 replied to the thread Restore mount local disks after reinstall PBS Server.I did a bit more research, with the following result: the corresponding file was not found in the directory.
-
-
5I'm trying a new project: I need to pin say 64g of ram to a vm. how can we configure a vm with interleaved memory on a fixed numa node (pinning the cpu cores in that numa node to a vm)? I know i need to use qemu and kvm config, disabled ksm...
-
UdoB reacted to BobhWasatch's post in the thread HA configuration for node migration/restart only with Like.
No, it still doesn't make sense. Proxmox is somehow supposed to know if this was an "intended" shutdown rather than a VM crash, and do something different in that case. And you want "on demand" machines to have high availability? What the heck...
-
OOnslow replied to the thread Replication of a VM between two independent Proxmox clusters (LVM storage).@katti Depending on your requirements on a how long the services can be unavailable and on how much fresh data can be lost, a workaround could be frequent backups to PBS. And restoring to the other cluster when needed). Thanks to "dirty bitmaps"...
-
UdoB reacted to jokergermany's post in the thread Migrating from mdadm to zfs despite using different disk sizes? with Like.
yeah i will only do */15 * * * * /usr/bin/syncoid -r tank/kritische_daten backup/kritische_daten > /dev/null
-
Aalexskysilk reacted to bbgeek17's post in the thread Storage for small clusters, any good solutions? with Like.
@alexskysilk You're right to call that out. "Very small environments" was a poor choice of words on my part. The distinction I was trying to make isn't about node count, it's about business value. We have customers running three node clusters...
-
Hhawxxer replied to the thread CPU Arguments cuts passthroughed NVMe performance by 4 times.Well I guess it is not solved, it works some boots, other boots its not working, without any changes to the config. Someone else has some ideas? :I
-
bbgeek17 replied to the thread Storage for small clusters, any good solutions?.@alexskysilk You're right to call that out. "Very small environments" was a poor choice of words on my part. The distinction I was trying to make isn't about node count, it's about business value. We have customers running three node clusters... -
bbgeek17 reacted to pveuser's post in the thread Storage for small clusters, any good solutions? with Like.
We're definitely at cross purposes here. Highly available storage <> highly available applications. And striving to have an infrastructure that delivers both is not a bad thing at all.
-
Kkmentzelos replied to the thread Storage for small clusters, any good solutions?.I’m exploring a similar setup using DRBD (simple configuration, not LINSTOR). I’ve developed a set of scripts to automate storage management in small Proxmox clusters: Provision per-VM storage with DRBD + LVM Resize storage layers safely (LV →...
