Haf develop crashes on CI
Issue relates to haf_block_explorer and reputation_tracker - in these apps this bug occures:
Changes done in:
- reptracker 4d80d257
- haf haf@83f7fe9b
Either one of these changes, or possibly both, are causing a sync "crash" in the app. While the sync stage appears green the logs stops without processing full 5m blocks.
From my investigation, I discovered that reverting to the previous version of the health checker resolves the issue. The relevant changes can be found here: !58 (commits)
The first commit, bump haf, updates HAF to a nearly latest development version (haf!566 (merged) with my additional commit changing openapi-rewrite script, not related to this issue). This newer version of HAF appears to break sync without generating an error message:
reptracker-backend-block-processing-1 | NOTICE: HAF instance is ready. Exiting wait loop.
reptracker-backend-block-processing-1 | NOTICE: Reptracker is attempting to process a block range: <1650001, 1660000>
reptracker-backend-block-processing-1 | NOTICE: Reptracker processed block range: <1650001, 1660000> successfully in 0.019531 s
reptracker-backend-block-processing-1 |
reptracker-backend-block-processing-1 | NOTICE: HAF instance is ready. Exiting wait loop.The second commit, test, reverts the health checker to an older version that does not utilize last_active_at. Implementing this change results in successful synchronization.
in balance_tracker this bug doesn't occur most likely because btracker's develop still uses the old version of heath checker:
hafbe:
haf_block_explorer!250 (closed) (rc9 haf, succesfull sync)
haf_block_explorer!251 (merged) (the same changes, develop haf)
The synchronization was successful; however, the logs indicate that both the Reptracker and hafbe didn't reach 5m blocks.