Skip to content

Avoids listing the sorted logs dir multiple times during log recovery.#4874

Merged
keith-turner merged 1 commit into
apache:3.1from
keith-turner:accumulo-4505-recovery-paths
Sep 13, 2024
Merged

Avoids listing the sorted logs dir multiple times during log recovery.#4874
keith-turner merged 1 commit into
apache:3.1from
keith-turner:accumulo-4505-recovery-paths

Conversation

@keith-turner
Copy link
Copy Markdown
Contributor

The log recovery code would list the sorted walog files 3 to 4 times during recovery. These changes modify the code to only list the files once. Also the listing is cached for a short period of time to improve the case of multiple tablet referencing the same walogs. This along with #4873 should result in much less traffic to the namenode when an entire accumulo cluster shutsdown and needs to recover.

Noticed this while working on #4873. While working on that was generally trying to understand where the tsever was spending time when loading a tablet w/ walogs. Noticed these directory listing showing up a little bit in timing logging I added. Just a few milliseconds, but it seemed unnecessary to repeatedly list the same dir over and over. Suspect the impact is worse at scale.

The log recovery code would list the sorted walog files multiple times
during recovery.  These changes modify the code to only list the files
once.  Also the listing is cached for a short period of time to improve
the case of multiple tablet referencing the same walogs.  This along
with apache#4873 should result in much less traffic to the namenode when an
entire accumulo cluster shutsdown and needs to recover.
@keith-turner keith-turner added this to the 3.1.0 milestone Sep 11, 2024
@keith-turner keith-turner merged commit 0c66dda into apache:3.1 Sep 13, 2024
@keith-turner keith-turner deleted the accumulo-4505-recovery-paths branch September 13, 2024 18:00
@ctubbsii ctubbsii modified the milestones: 3.1.0, 4.0.0 Mar 13, 2025
amcdonaldccri pushed a commit to amcdonaldccri/accumulo that referenced this pull request May 26, 2026
… log recovery. (apache#4874)

The log recovery code would list the sorted walog files multiple times
during recovery.  These changes modify the code to only list the files
once.  Also the listing is cached for a short period of time to improve
the case of multiple tablet referencing the same walogs.  This along
with apache#4873 should result in much less traffic to the namenode when an
entire accumulo cluster shutsdown and needs to recover.

{"fundingSource": "41201", "team": "FED.ICGSA.OPS.MOE", "fshGit": "dummy-lo", "fshDocker": "sha256:20cf0045"}
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants