I have just installed Pydio Cells 1.0.4 on my Debian 8 VPS.
I noticed high cpu and memory usage (After sometime my server gets laggy). Furthermore there is more than one
home/pydio/cells start proccess.
Is this normal? If so can someone explain to me why there are so many processes.
we recommend that you run cells with 4 gb of ram and 2 core cpu,
it can run with less but if you do some demanding tasks your hardware will be used at full power.
here’s a requirement page if you need more information.
About the multiple process they all are services, you can see that each task has a different name, and if you do
./cells list each task will match a service.
is it planned to reduce cpu/ram consumption ?
So far I was using pydio server on a double core with 2gb of ram, I had almost 3Tb of data and a few users, and never saw any ressources limitation on my server (I have plenty of others services)
With pydio cell, on the same exact server, no data, no user, clean install (last version without php and a mariadb base), it takes 1gb of ram and it takes permanently between 50 and 80% of global cpu…). After 2 days I had to shut it down (initially I thought it was only a temporary ressource consumption, but as I have a monitoring in place, I can tell it was permanent).
So even, good old bloated php seems lighter. (and I know that golang should be better for ressource consumption but something went wrong there.).
I know cells is a lot better in term of architecture and surely a lot of other stuff but plain install shouldn’t consumme that much from the start, it feels like a regression to me so far.
High CPU and High Memory usage.
I have the same problem - high CPU and high memory usage.
There is definitely a problem with the code base of Pydio Cells.
I have a fresh install on Debian 9, with latest PHP, MariaDB and Cells.
CPU is 20% at idle and memory uses 1.5GB without any users or data. Fresh install.
All CPU usage is coming from Pydio:
the team was on holidays,
we are currently looking and working on the memory consumption.
edit: we actually already patched once which reduced the memory consumption
could you tell me which pydio cells version are you using?
Yes, it is 1.2.4 – the latest version.
PHP and MariaDB are also the latest.
I’m also seeing a high CPU usage on a new (and idle) Cells deployment.
My server specs :
- CPU = Intel i5-2300 (4 x 2.80GHz)
- RAM = 16 Gb DDR3
- Storage = 2 Tb SATA
- OS = Debian 9.7 (up to date 2019-02-15)
- MariaDB = 10.3.12-MariaDB
- Pydio = 1.2.5 (Docker)
- Docker = 18.09.2
Here some graphs (vertical redline = Cells deployment) :
First observations :
- + 15% CPU usage (on idle)
- + 250 processes
- + 10 000 CPU interrupts (LOC) /s
CPU interrupts is probably the source of the problem, this is a ~700% increase !
I hope this can help.
LOC/s will increase with the number of processes as this is the task/thread switching The more (active) processes are running the more task switches will occur.
Yes I now that … but I think something is wrong with the volume on an idle instance here.
Hi all - check the latest HEAD
You’ll love it.
I didn’t test yet but it’s good to hear that !