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
mds: don't maintain bloom filters in standby replay #12133
Conversation
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.
Couple questions/changes.
fs=0 | ||
for name in a b c d e f g h i j k l m n o p | ||
do | ||
sleep 5 # wait for daemons to make it into FSMap |
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.
This appears to be sleeping 5 seconds for each fs. I don't think that was meant to happen?
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.
Fixed
CInode *in = dn->get_linkage()->get_inode(); | ||
|
||
if (dn->get_linkage()->is_primary()) { | ||
in->get_nested_dirfrags(subdirs); |
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.
I must be misremembering how EExport works. Why do we need to go down one level to pull out the child CDir bloom filters, but not all the way to the leaves?
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.
remove_blooms calls itself, it does go down to the leaves.
I think it's better to avoid adding bloom filter while standby replay |
The daemons will get assigned to roles in the order they start otherwise. The trick is to let them all go into the FSMap first and then increase max_mds. Signed-off-by: John Spray <john.spray@redhat.com>
...and thereby avoid needing to clear them when replaying an EExport Fixes: http://tracker.ceph.com/issues/16924 Signed-off-by: John Spray <john.spray@redhat.com>
@ukernel good point, I agree. I've updated this. |
Reviewed-by: Greg Farnum gfarnum@redhat.com |
Fixes: http://tracker.ceph.com/issues/16924