Loading...

XML

Word

Printable

JSON

Type: Bug
Resolution: Won't Fix
Priority: Major - P3
Fix Version/s: None
Affects Version/s: None
Component/s: None
Labels:
None

Assigned Teams:

Replication
Operating System:
ALL
Confidence Status:
None
Work Order:
3
CAR Domain/s:
None

Aha! Reference:
None
Tracking Level:
None
Risk Status:
None
Exec Notes:
None
Goal Name(s):
None
Goal Link:
None

During oplog replay in initial sync we don't advance the oldest timestamp. This can pin a lot of data in the cache. Resulting symptoms include

slow initial sync due to resulting cache churn
growth of lookaside (cache overflow) table and corresponding file WiredTigerLAS.wt
in extreme cases cache-full hangs were observed. This was believed fixed with ~~SERVER-33191~~, but we should be alert for new occurrences of this issue.

is caused by

WT-4016 Measure and improve lookaside performance with stable_timestamp set

Closed

related to

SERVER-33191 Cache-full hangs on 3.6

Closed

SERVER-34900 initial sync uses different batch limits from steady state replication

Closed

SERVER-34938 Secondary slowdown or hang due to content pinned in cache by single oplog batch

Closed

SERVER-34942 Stuck with cache full during oplog replay in initial sync

Closed

SERVER-36238 replica set startup fails in wt_cache_full.js, initial_sync_wt_cache_full.js, recovery_wt_cache_full.js when journaling is disabled

Closed

(1 related to)

Assignee:: [DO NOT USE] Backlog - Replication Team
Reporter:: Bruce Lucas (Inactive)
Participants:: [DO NOT USE] Backlog - Replication Team, Bruce Lucas, Lingzhi Deng
Votes:: 0 Vote for this issue
Watchers:: 22 Start watching this issue

Created:: Aug 07 2018 04:50:54 PM UTC
Updated:: Oct 02 2023 06:26:29 PM UTC
Resolved:: Oct 02 2023 06:26:29 PM UTC