Mailing List Archives
Authenticated access
|
|
|
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [HTCondor-users] zswap for jobs
- Date: Wed, 8 Apr 2026 14:04:03 +0200
- From: Emily Kooistra <a66@xxxxxxxxx>
- Subject: Re: [HTCondor-users] zswap for jobs
Hi Bert,
Thanks for the info!
On 4/8/26 12:36, Bert DeKnuydt wrote:
Does anyone have operational experience with running zswap on EPs?
Yes. I've been running zswap for years, on almost everything, including
EPs. Here, most jobs use little or no
swap at all, but some just 'explode', often only temporarily. I've
always had the (admittedly: subjective) feeling
that EPs tend to survive such events (as in remaining responsive) much
better using zswap. In fact, I see little
disadvantage in enabling it (as compared to zram).
Thats good to hear.
I was wondering what the compression rates people are getting for this?
No hard numbers; I doubt for real data it's more than 2.
That was also my expectation.
In my opinion, that's the advantage of zswap: the kernel manages the
lot, there's no tuning needed (balancing between
zwap and disk-cache).
Yea that was also my impression, altho beeing able to limit the total
amount of memory a job can store in zswap would be beneficial, given
this right now is unlimited. (Or well up to the system max), by having
condor also set memory.zswap.max based on a classadd expresion, similary
to the other cgroup limits.
Emily