@crosbymichael @dmcgowan - Since the snapshotter runs in proc there is no need for a file lock. The issue here was that if the creation fails after the lock was taken the lock file was leaked to the file system. Once the lock file was present because of opening with flags os.O_EXCL|os.O_CREATE we were never able to obtain the lock again and would time out on 2nd->Nth retries.

Codecov Report

Merging #2682 into master will not change coverage.
The diff coverage is n/a.

@@          Coverage Diff           @@
##           master   #2682   +/-   ##
======================================
  Coverage      45%     45%           
======================================
  Files          92      92           
  Lines       10114   10114           
======================================
  Hits         4552    4552           
  Misses       4842    4842           
  Partials      720     720

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 3bc4ba2...e373126. Read the comment docs.

LGTM

LGTM

Since this is running inside of a transaction context there should already be a file level lock on the bolt database.