gurubert's latest activity
-
gurubert replied to the thread latency on ceph ssd.Slow networking also adds latency. -
see this old blog: https://www.sebastien-han.fr/blog/2014/10/10/ceph-how-to-test-if-your-ssd-is-suitable-as-a-journal-device/ The main reason is sync write, without plp, the ssd need to write directly to a full nand. for example, you need to...
-
It's just that these are SSD consumer drives without PLP. These are slow if you force them to write the data (sync writes). You'll find many threads about this on the forum. Maybe you could buy small enterprise SSDs with PLP and put the WAL+DB...
-
don't use consumer ssd with ceph or zfs. you need PLP to handle sync write
-
I went through the list. I see two problems with your SSDs: - All the SSDs have no Power-Loss-Protection (PLP). Without PLP all writes have to be completed to report it as finished. This slows Ceph writes down. With PLP the drive could report...
-
gurubert replied to the thread Ceph in 2026 HDD.For erasure coding with k=8 and m=3 you need at least 12 nodes.
-
Those were hacks and only accelerated specific workloads or patterns. Cache misses still gave the worst case scenario performance and VM and databases both are workloads that should be calculated accordingly (unlike filer NAS). Ceph/Linux has a...
-
The main bottleneck with hard drives is that they’re relatively slow. So ~100 IOPS peak and maybe 5-10MB/s in random read write scenarios. You need to read/write 100TB/day -> 1200MB/sec -> you need 300 disks just to keep up multiplied by the...
-
Hello, I wonder, what's the rationale behind a tool like this. What problem is it trying to solve?
-
We have a 24 node cluster, some on 10G, soon all will be on 2x100G. No problems, even having 14 nodes on a very busy 2x10G segment, with Ceph and no dedicated fabric. Things are way more robust than people like to make it sound. There are some...
-
gurubert reacted to Johannes S's post in the thread Clarification about Ceph suitability for a 2-node storage architecture with a witness/quorum node with
Like.
Don't. Ceph needs at least three nodes unless you change the defaults ( and doing this is a bad idea, see: https://forum.proxmox.com/threads/ceph-pool-size-is-2-1-really-a-bad-idea.68939/ ) The recommended requirements in the documentation are... -
gurubert reacted to MarkusKo's post in the thread Clarification about Ceph suitability for a 2-node storage architecture with a witness/quorum node with Like.
No, Ceph needs at least 3 nodes, for production 5 nodes or more (odd number) are recommended: PVE docs
-
gurubert reacted to UdoB's post in the thread Clarification about Ceph suitability for a 2-node storage architecture with a witness/quorum node with Like.
Two years ago I tried to use Ceph in my homelab. My experience: https://forum.proxmox.com/threads/fabu-can-i-use-ceph-in-a-_very_-small-cluster.159671/ I am not a Ceph user now. But I would not start with fewer than five active nodes. Just my...
-
Have you already purchased the motherboards? If not price out single cpu systems with plenty of memory. The savings over dual cpu ( assuming you were pricing that out) may leave cash to afford nvme’s or data a center grade ssd
-
It depends on your performance expectations, but ceph is quite capable to deal with this kind of mixed hardware- as long as you make sure that you have EQUIVALENT CAPACITY PER OSD TYPE on each osd server. In your case, assuming you want all 6...
-
Hi AllYourSmurf if you are able to do so, you should distribute your nvmes more equally across, the nvme nodes or even better across all nodes. With the current distribution only 3TB of your NVME-Storages will be usable (with the default 3-2...
-
gurubert reacted to IsThisThingOn's post in the thread [Community project] Proximo — an open-source, least-privilege MCP/API layer for managing PVE with an AI agent (feedback wanted) with Like.
You brought it up, not me. Just a friendly advice, based on my personal experience. Nothing more, nothing less. Feel free to ignore it. But back to topic. Let us not put the cart before the horse. Why should I even want to manage PVE with an AI...
-
gurubert reacted to meyergru's post in the thread [Community project] Proximo — an open-source, least-privilege MCP/API layer for managing PVE with an AI agent (feedback wanted) with Like.
@broadway: Let’s take a step back from the technical scaffolding you are proposing. Hash-chains, dry-runs, and scoped RBAC tokens are standard practices for traditional API automation. They do not, however, mitigate the core issue we are...
-
gurubert reacted to meyergru's post in the thread [Community project] Proximo — an open-source, least-privilege MCP/API layer for managing PVE with an AI agent (feedback wanted) with Like.
If the token only needs read-only privileges, the agent cannot actually do anything harmful, that is correct. On the other hand - it cannot do anything at all.
-
gurubert reacted to IsThisThingOn's post in the thread [Community project] Proximo — an open-source, least-privilege MCP/API layer for managing PVE with an AI agent (feedback wanted) with Like.
Not trying to be a dick, but I would never run software maintained by a single person with mental health issues. But back to topic. Let us not put the cart before the horse. Why should I even want to manage PVE with an AI agent? Proxmox is...