I’m currently running an interactive job on Bamboo (cpu023, shared-cpu) and the scratch partition is slow enough as to be unusable. Operations like ls take ~1-2s to perform, and a file transfer to local /tmp/ is proceeding at about 4-5 MB/s maximum speed.
Is there an issue with BeeGFS for that file system?
Thanks!
Berk
I was about to report the same. I did time ls and
real 0m38.679s
user 0m0.003s
sys 0m0.016s
Hi, @Berk.Gercek and @Lucille.Delisle1
I think the issue was due to high load on the storage as for now it seems ok to me:
I copied a 12GB file from scratch to /tmp in 23 sec which is not exceptional but not bad:)
(bamboo)-[sagon@cpu001 scratch]$ time cp mybigfile /tmp
real 0m23.624s
user 0m0.013s
sys 0m6.796s
(bamboo)-[sagon@cpu001 scratch]$ ls -lah mybigfile
-rw-r--r-- 1 sagon unige 12G Jun 27 09:14 mybigfile
Can you confirm it is working for you as expected? Unfortunately it is hard to diagnose such performance issue because it is highly dependent on the users workload.
Hi @Yann.Sagon ,
I confirm that this morning everything is fine.
Thanks