New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
FileStore: support multiple ondisk finish and apply finisher #6486
Conversation
@liewegas , please reivew it, thanks. |
throttle_ops(g_ceph_context, "filestore_ops",g_conf->filestore_queue_max_ops), | ||
throttle_bytes(g_ceph_context, "filestore_bytes",g_conf->filestore_queue_max_bytes), | ||
op_finisher(g_ceph_context), | ||
ondisk_finisher_num(g_conf->filestore_ondisk_finisher_threads), | ||
apply_finisher_num(g_conf->filestore_apply_finisher_threads), |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
prefix these with m_ and put them next to all the other tunables, a few lines down?
Cosmetic issues aside this looks good to me! |
ondisk finisher and apply finisher wouble be bottleneck in fast ssd. Because apply finisher should free memory which may be slow. Signed-off-by: Haomai Wang <haomai@xsky.com> Signed-off-by: Xinze Chi <xinze@xsky.com>
5e76eea
to
db85bdd
Compare
@XinzeChi @liewegas @yuyuyu101 I don't know if it is better if implement member control in Finisher(default 1, will not affect other caller) ? |
@XinzeChi , do you have any performance number to show ondisk finisher and apply finisher is bottleneck for ssd |
Sorry, misread something. |
@XinzeChi This passed testing but I just realized it was with the default of 1 instance. Updating the teuthology ceph.conf to set this to 3 and I'll run it through again. |
osd: FileStore: support multiple ondisk finish and apply finishers Reviewed-by: Sage Weil <sage@redhat.com> Reviewed-by: Samuel Just <sjust@redhat.com>
ondisk finisher and apply finisher wouble be bottleneck in fast
ssd. Because apply finisher should free memory which may be slow.
Signed-off-by: Haomai Wang haomai@xsky.com
Signed-off-by: Xinze Chi xinze@xsky.com