From: Marco Gaiarin <gaio@lilliput.linux.it>
To: dorsy via pve-user <pve-user@lists.proxmox.com>
Cc: pve-user@lists.proxmox.com
Subject: Re: [PVE-User] A less aggressive OOM?
Date: Sun, 13 Jul 2025 16:28:35 +0200 [thread overview]
Message-ID: <hdsckl-ucq3.ln1@leia.lilliput.linux.it> (raw)
In-Reply-To: <mailman.1274.1752159599.395.pve-user@lists.proxmox.com>; from SmartGate on Sun, Jul 13, 2025 at 16:36:01PM +0200
Mandi! dorsy via pve-user
In chel di` si favelave...
Thanks to all, particulary to Victor for the wonderful analisys, that lead
me to learn a bit better OOM dump...
>> if OOM kicks in because half of the ram is being used for
>> caches/buffers, i would blame OOMkiller or ZFS for tha. The problem
>> should be resolved at zfs or memory management level.
> Absolutely no!
> You are responsible for giving ZFS the limits. As even described in the
> proxmox documentation here:
> https://pve.proxmox.com/wiki/ZFS_on_Linux#sysadmin_zfs_limit_memory_usage
I'm a bit in the side of Roland on this. ARC is a (indeed, complex)
buffer/cache, so seems reasonably that, if i need to sacrifice something, it
is better to sacrifice cache than VM.
Aniway, if i understood well, default ZFS was to have ARC at 50% of the RAM;
after PVE 8.1, PVE modify the default to 10% (for new installation); there's
also a 'rule of thumb' to setup ARC, so 10% is somewhat a 'starting point'.
In some server i can setup easily swap (i have a disk for an L2ARC, so i can
simply detach, partition a bit and reattach as L2ARC and swap).
Clearly, i'll set swappiness at 1, to be used only when strictly needed.
Thanks to all!
--
_______________________________________________
pve-user mailing list
pve-user@lists.proxmox.com
https://lists.proxmox.com/cgi-bin/mailman/listinfo/pve-user
next prev parent reply other threads:[~2025-07-13 14:39 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-07-07 9:26 Marco Gaiarin
2025-07-07 21:39 ` Victor Rodriguez
2025-07-08 16:31 ` Marco Gaiarin
2025-07-10 8:56 ` Victor Rodriguez
2025-07-10 9:08 ` Roland via pve-user
2025-07-10 14:49 ` dorsy via pve-user
2025-07-10 16:11 ` Roland via pve-user
[not found] ` <98ace9cf-a47f-40cd-8796-6bec3558ebb0@web.de>
2025-07-10 16:15 ` dorsy via pve-user
2025-07-13 14:28 ` Marco Gaiarin [this message]
2025-07-08 12:05 ` Roland via pve-user
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=hdsckl-ucq3.ln1@leia.lilliput.linux.it \
--to=gaio@lilliput.linux.it \
--cc=pve-user@lists.proxmox.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.