poolpOrg
huge memory swap usage
Describe the bug Runing plakar to backup an huge directory ( over 4Tb ), very slow to create ( stopped 4 days after running ), and huge memory swap consumption
To Reproduce Steps to reproduce the behavior:
- Run plakar at backup with excludes files & directories ( 4Tb of data )
- Let's it runing
- Monitoring memory
- Swap increase , server is overload
Expected behavior Why using a lot of swap memory ?
Desktop (please complete the following information):
- Debian 12 / bookworm , uptodate
- Plakar Version v1.0.3
- shell ssh ;)
Hello, do you mind sharing the number of cores and the amounts of memory that you have ?
I opened an #1542 to discuss the solution to this problem but am curious
Hello, 16 cpu cores , with 32Gb for ram soft raid ( 2 ssd for / with 512Gb ) soft raid ( 4 hdd for data & backup with 39Tb ) vendor_id : AuthenticAMD cpu family : 23 model : 113 model name : AMD Ryzen 7 PRO 3700 8-Core Processor
What is the source of the backup, and the destination storage ? With 1.0.3, 16 cores and 32gb should be plenty fine. Is it a slow ramp up, or did it saturate RAM quickly forcing the system to swap to disk?
Source and backup are on the same raid space , datas are daily updated via rsync in a first "spare" directory, and I try to make a backup of this daily spare . It's a slow ramp up with memory swap saturation and after disk swap saturation. The last run take 8 hours and I need to kill the process before ( cpu usage > 100, and ram saturation ).
I've tryied on another backup ( databases, env 500Gb ), the backup run fine