-
Type: Bug
-
Resolution: Cannot Reproduce
-
Priority: Critical - P2
-
None
-
Affects Version/s: None
-
Component/s: None
-
None
i'm connecting to a sharded cluster, without setting any of the w, wtimeout or j parameters. I've tried running an update command which alters a single document in the following ways:
from pymongo.write_concern import WriteConcern dc1 = db.col.with_options(write_concern=WriteConcern(w=0)) db.col.update({"_id":"x"},{..}, w=0) dc1.update({"_id":"x"},{..}, w=0) dc1.update_one({"_id":"x"},{..})
but in the logs, i can see the exact command running with
writeConcern: { w: 1 }
The update is an increase on a field, and whenever there is a lot of activity and many workers try to increase that field they tangle eachother, and the command takes a lot to execute. I've assumed it was because of the writeconcern, and i've been assuming that disabling the write concern would make it faster, but it doesn't work.
- is related to
-
SERVER-14003 2.6 mongos does not support overlapped writes at w:0
- Closed