[sheepdog] sheepdog + libvirt problems

Vasiliy Tolstov v.tolstov at selfip.ru
Tue Jul 21 12:24:05 CEST 2015


Hi again. I'm stress test libvirt with sheepdog storage backend, and
have errors while try to start 20-30 domains in the same time.
As i see libvirt refresh pool each time when in creates volume, so
dog node info -r run each time, in case of 20-30 vm in the same time
libvirt says about:
internal error: pool 'sheepdog' has asynchronous jobs running.
so - may be add timestamp to output to determine on which time info is valid ?
or how to speedup pool refresh?

root at cn33:~# time dog node info -r
0 2707202199552 10276126720 2696926072832 0%
1 2707202207744 9408106496 2697794101248 0%
2 2694368509952 10993508352 2683375001600 0%
3 2694368501760 11215777792 2683152723968 0%
4 2694274146304 10708291584 2683565854720 0%
5 2694368509952 9773056000 2684595453952 0%
6 2694368509952 10863489024 2683505020928 0%
7 2694368432128 13157613568 2681210818560 0%
Total 21580521017344 86395969536 21494125047808 0% 343597383680

real    0m0.298s
user    0m0.000s
sys     0m0.012s



-- 
Vasiliy Tolstov,
e-mail: v.tolstov at selfip.ru


More information about the sheepdog mailing list