[SERVER-5606] Mongos sharding filled with "update will be retried b/c sharding config info is stale" errors Created: 14/Apr/12  Updated: 06/Dec/22  Resolved: 22/Feb/18

Status: Closed
Project: Core Server
Component/s: Logging, Sharding
Affects Version/s: 2.0.2
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: joe piscitella Assignee: [DO NOT USE] Backlog - Sharding Team
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

ubuntu 11.04


Assigned Teams:
Sharding
Operating System: ALL
Participants:

 Description   

I believe this bug was suppose to be fixed in 2.0.2 and there are notes it is fixed in 2.3 (appears there is a fix (https://jira.mongodb.org/browse/SERVER-4387) but what do we do until then?

mongos logs are full of stale config error message.
we are running 2.0.2 on all machines, 4 application servers with mongos and a 6 machine shard/replica set with 3 shards and 2 member replica set with arbiter. Tried flush router command on all mongos instances but still no luck.

Sat Apr 14 06:52:59 [conn120] update will be retried b/c sharding config info is stale,
Sat Apr 14 06:52:59 [conn121] update will be retried b/c sharding config info is stale,
Sat Apr 14 06:52:59 [conn121] update will be retried b/c sharding config info is stale,
Sat Apr 14 06:52:59 [conn120] update will be retried b/c sharding config info is stale,
Sat Apr 14 06:52:59 [conn120] update will be retried b/c sharding config info is stale,
Sat Apr 14 06:52:59 [conn120] update will be retried b/c sharding config info is stale,
Sat Apr 14 06:53:01 [conn121] update will be retried b/c sharding config info is stale,



 Comments   
Comment by Kevin J. Rice [ 20/Sep/12 ]

It would be nice to know if this is a warning of a situation that should be addressed, perhaps with application-level code, networking people, or sysadmins...

Comment by Tad Marshall [ 15/Apr/12 ]

Th fix you referenced is in 2.0.3 and in 2.0.4 which you could download, but I don't think they will silence these warnings. It is normal in an active sharding cluster for mongoses to discover that their config is stale when they try to do an update, it is the usual method for finding out that they need to update their sharding config. The logging is too verbose, as Eliot said. It may be helpful when problems show up to see these messages, but in normal operation they just tell us that the system is doing what it is supposed to do.

Comment by Eliot Horowitz (Inactive) [ 15/Apr/12 ]

There should be more to the message than that per message, can you send the whole thing?
Either way - its not a bug, just overly verbose messaging.

Generated at Thu Feb 08 03:09:23 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.