[sheepdog] [PATCH 1/2] collie: parallelize parse_vdi() with work queue

Hitoshi Mitake mitake.hitoshi at gmail.com
Mon Aug 19 10:17:43 CEST 2013


At Mon,  5 Aug 2013 20:29:32 +0900,
Hitoshi Mitake wrote:
> 
> Current collie issues request for gathering VDI information in a
> sequential manner (parse_vdi()). This way is not scalable when a
> number of VDIs becomes larger.
> 
> This patch parallelize parse_vdi() with work queue. Some collie
> commands which call parse_vdi() can enjoy performance improvement.
> 
> The below is an sample of collie vdi list. The test is done on 16
> nodes cluster which has 3000 VDIs.
> 
> Before:
> $ time sh -c "collie/collie vdi list -a 10.68.13.1 > /dev/null"
> sh -c "collie/collie vdi list -a 10.68.13.1 > /dev/null"  8.81s user 0.24s system 70% cpu 12.876 total
> 
> After:
> % time sh -c "collie/collie vdi list -a 10.68.13.1 > /dev/null"
> sh -c "collie/collie vdi list -a 10.68.13.1 > /dev/null"  14.35s user 2.02s system 209% cpu 7.816 total
> 
> Signed-off-by: Hitoshi Mitake <mitake.hitoshi at lab.ntt.co.jp>

I've refined this series and will post the new version later. Please
ignore this one.

Thanks,
Hitoshi



More information about the sheepdog mailing list