MDS OpTracker #1809
Merged
MDS OpTracker #1809
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
I'm not sure why we ever had this instead of just doing things in the subclass constructor, and the semantics around it don't make much sense for anything else (we called mark_event first anyway). Signed-off-by: Greg Farnum <greg@inktank.com>
Use this instead of direct access to the Message underneath when dumping the TrackedOp. Signed-off-by: Greg Farnum <greg@inktank.com>
Instead of relying on the message's get_recv_stamp, take a timestamp when the TrackedOp is constructed. Rename get_received_at() -> get_initiated(). Signed-off-by: Greg Farnum <greg@inktank.com>
…ns clean up Right now, the OpRequest uses it to clean up Message payload data. Signed-off-by: Greg Farnum <greg@inktank.com>
Give it to the OpRequest (currently, the only TrackedOp implementation). Signed-off-by: Greg Farnum <greg@inktank.com>
Further parameterize the template to allow passing in an arbitrary parameter, and move all the Message-based event marking from there into OpRequest. Signed-off-by: Greg Farnum <greg@inktank.com>
This ensures we always have an event for state_string(). Signed-off-by: Greg Farnum <greg@inktank.com>
Signed-off-by: Greg Farnum <greg@inktank.com>
We were passing the causative MDS (as an int), but pushing down the actual Message will help us as we set up an OpTracker. Signed-off-by: Greg Farnum <greg@inktank.com>
Signed-off-by: Greg Farnum <greg@inktank.com>
…Requests We now have a single constructor and one path to build MDRequests with. Signed-off-by: Greg Farnum <greg@inktank.com>
Signed-off-by: Greg Farnum <greg@inktank.com>
Signed-off-by: Greg Farnum <greg@inktank.com>
Signed-off-by: Greg Farnum <greg@inktank.com>
Signed-off-by: Greg Farnum <greg@inktank.com>
Signed-off-by: Greg Farnum <greg@inktank.com>
Conflicts: src/mds/Locker.cc src/osd/OpRequest.cc src/osd/OpRequest.h Signed-off-by: Greg Farnum <greg@inktank.com>
|
This looks good! Reviewed-by:, and fix or ignore my nits at your option. |
Signed-off-by: Greg Farnum <greg@inktank.com>
gregsfortytwo
added a commit
that referenced
this pull request
May 16, 2014
MDS OpTracker Reviewed-by: Sage Weil <sage@inktank.com>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This branch makes some changes to the TrackedOp infrastructure to allow Ops not based on Messages, then puts an OpTracker into the MDS and uses it to track all MDRequests through their lifetime. The tracking is not highly individualized, but does cover all of the common points and is specific enough to let you correctly identify repeating journal points.
This branch was unittest clean prior the master merge (unfortunately master had a minor issue at the merge point) and passes local testing against several workunits. I haven't run it through a whole suite but it's pretty low-risk at that level.