tree 6fbcae91b714e14846b97cf25a80a3c02e7702f2
parent 7a8b421f0bc2e78f75b178239f9b001f113a293e
author Adithya Chakilam <achakila@codeaurora.org> 1633387482 -0500
committer Adithya Chakilam <achakila@codeaurora.org> 1634687900 -0500

Parallelize Schema 130 and 131

Both these schemas iterate through all repositories and perform updates.
Since this is a slow operation on NFS, parallelizing them would be
efficient. Default number of migration threads is equal to number of
processors or can be manually configured through
cache.projects.loadThreads property.

This reduced runtime for migration 130 on a gerrit site with ~16k repos
on NFS from ~570secs to ~140secs on a machine with 24 processors and
schema 131 runtime reduced from ~150secs to ~35secs on same setup.

Also, refactor Schema 106 to use common methods created to avoid
boilerplate code.

Change-Id: If26c90f659b3b1d537e45034f04d9f6b7825e537