-
Type: Bug
-
Resolution: Unresolved
-
Priority: Major - P3
-
None
-
Affects Version/s: None
-
Component/s: None
-
8
-
StorEng - Defined Pipeline
This ticket comes from WT-8003, and is describing a bug within the mongos test REP-2152 and the genny workload test. The other test comes from a branch:
I have another Genny workload from the sharded stress testing project that can reproduce this issue (see details in PERF-2665). It's a bit slow to run, but I ran it twice on Atlas with v5.0.4 using sys-perf-atlas and had this error happen both times. Hoping this helps with the investigation, let me know if you have any questions!
The tests are going through the $sample stage and is not being able to find non-duplicate keys and eventually fails out.
To describe the context on why this issue happens is due how the random algorithm works:
- If disk records are > 1000, find a visible record from page disk.
- Look to find a visible record from the insert list
- If disk records are < 200 or the leaf page is clean, try to find a visible page disk.
- Last step, compute a random record between 0 - 250 entries from the page and perform bt_curnext and bt_curprev until we find a visible record.
It is most likely that the last step is being reached and causing the random skew to happen. The developer will need to investigate the reason behind why the randomness is skewing. This ticket will need to investigate the reason why and if it potentially duplicates WT-11533, or WT-11534
- related to
-
WT-8003 Fix frequent duplicate keys returned by random cursor in resharding test
- Closed