fix: remove dead restart_failed_sync from custody backfill #8553
Conversation
|
|
|
|
| // Check if we need to restart custody backfill sync due to a cgc change. | ||
| if self.restart_if_required() { | ||
| return Ok(ProcessResult::Successful); | ||
| } |
There was a problem hiding this comment.
what is the reasoning for deleting this?
|
we probably want to keep |
eserilev
added
waiting-on-author
| // inform the backfill sync that a new synced peer has joined us. | ||
| if new_state.is_synced() { | ||
| self.backfill_sync.fully_synced_peer_joined(); | ||
| self.custody_backfill_sync.fully_synced_peer_joined(); |
There was a problem hiding this comment.
i don't think we want to remove this, i think we want to actually get restart_failed_sync to work appropriately
ANtutov
force-pushed
the
fix/custody-backfill-dead-restart-flag
branch
from
bf01ec8 to
a84a1cb
Compare
|
Also squashed tp reduce noise in commits |
|
please read my comment above. We probably want to keep restard_failed_sync and fix its usage instead of deleting it |
Done. I kept restart_failed_sync and wired it up like in BackFillSync, added a failed_sync flag so custody backfill only restarts after a failure once a new fully-synced peer joins. |
2 participants
Make custody backfill restart semantics match BackFillSync. Instead of immediately restarting after a failure, track failed attempts with a new failed_sync flag and only allow restart once a new fully-synced peer joins and sets restart_failed_sync. This prevents tight restart loops on the same bad peers and uses fully_synced_peer_joined from SyncManager to drive intentional retries.