Multiple threads of single process creating load on system

I am using a shared host after running pydio i am getting multiple process threads of below given process even i i haven’t used the pydio-cell yet just configured it no files has been added.

/app/cells start --fork --config local --registry grpc://:8000 --broker grpc://:8003 ^pydio.grpc.username

Please any one can help ?

I am using linuxserver.IO default image no custom changes has been made i want to limit the process or stop them.

Well …

Pydio Cells is a server built using a micro-service architecture… You might have a quick look here to have a first overview:

but TL;DR you have a main process and a few forks that run even if the server is idle.

With v4 we have made put some more work so that the full footprint of the app is minimal typically when idle (and we have measured some very promising performances): even if main process and forks are there they shouldn’t consume so much resources but then if it still too much for you, the only solution is to turn-off the service, but it won’t work so well then :frowning:

I concur — Pydio Cells does, indeed, take a toll on any server, even when absolutely idle. On my Ubuntu Linux 22.04 box, with 8 cores and way too much free memory, all Cells processes and sub-processes will take up close to a GByte of RAM — but perhaps 70% of that is shared memory — which means that, although big, Cells does not take so much memory as it seems. Additionally, when viewed with top or a similar tool, it’s almost certainly on the top ten running applications — always! — but it’s actually not consuming a large fraction of the available CPU cycles. It’s just always running and doing ‘something’, that’s why it’s always listed at the top… in other words: it’s rare for most Cells processes to be idle for long periods of time (which would mean they’d get scheduled out of memory by the OS). A few will be always set to running, or ready to be run, by the system scheduler. That’s why it seems that Cells is taking so many resources.

Oh… and this comment is after visually observing what is happening on the system with a few tools — and running a release candidate for the 4.x branch. The latest version of the 3.x branch did seem to consume slightly more, mostly because there would be many more subprocesses listed on the ‘constantly ready to run’ state. Indeed, 4.x seems to have hardly more that 3-4 subprocesses on the top ten list, if it’s mostly doing nothing (i.e. no user is actively logged in and using it). There are, of course, scheduled jobs running every now and then, thus Cells ‘never sleeps’, so to speak.

I’d totally agree that the 4.x branch looks very, very promising; long are gone the days where a single process would consume 100% resources of a single core (which would be too bad if that’s the only one you had — something that isn’t that uncommon with virtual servers)…

This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.