Fork CKAN core for the solr-down-200-code fix
This following bug in CKAN core has been a pain for us given our solr periodically restarts a few times a week.
- https://github.com/ckan/ckan/issues/8002
Now it is fixed and merged into CKAN master branch, we want to cherry-pick this change now.
- https://github.com/ckan/ckan/pull/8094
Sketch
- Cherry-pick this change into
ckan-2.10.4 - Verify the fix
- Create a
ckan-2-10-4-fixbranch on GSA CKAN - point ckan package in the catalog repo to the forked branch in the requirement.in
The change will be included in CKAN 2.10.5. Waiting for it.
@FuhuXia, should we move this from blocked to in progress, or are we still blocked on CKAN for adding a fix to our fork?
Let us fix it using our fork. Don't want to wait any longer.
Catalog and inventory are on patched CKAN core. No more need to clear cloudfront cache on solr restart.
obviously we missed some edge cases that solr down still returns 200 code and it is cached by cloudfront.
reopen.
Fix deployed and verified. Getting 500 and 502 code during SOLR restarting.