[SERVER-2813] Temporary MapReduce Collections Won't Drop Created: 22/Mar/11  Updated: 12/Jul/16  Resolved: 29/Mar/11

Status: Closed
Project: Core Server
Component/s: None
Affects Version/s: 1.7.5, 1.8.0
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: Kirk Morales Assignee: Unassigned
Resolution: Done Votes: 0
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Ubuntu 10.04 Server 64-bit, sharding


Operating System: Linux
Participants:

 Description   

Attempting to remove empty, temporary Map Reduce collections fails (tried deleting collections in a variety of ways, none of which work):

db.runCommand(

{drop: 'tmp.mr.foo_12980807563244151_25433_inc'}

)

{ "errmsg" : "ns not found", "ok" : 0 }

Output from command:

Mon Mar 21 17:23:49 [conn326] query: test.$cmd

{ drop: "tmp.mr.foo_12980807563244151_25433_inc" }

Mon Mar 21 17:23:49 [conn326] run command test.$cmd

{ drop: "tmp.mr.foo_12980807563244151_25433_inc" }

Mon Mar 21 17:23:49 [conn326] CMD: drop test.tmp.mr.foo_12980807563244151_25433_inc
Mon Mar 21 17:23:49 [conn326] query test.$cmd ntoreturn:1 command:

{ drop: "tmp.mr.foo_12980807563244151_25433_inc" }

reslen:78 0ms

I've tried dropping these collections through mongos as well as by logging in directly through the masters of each replica set.



 Comments   
Comment by Eliot Horowitz (Inactive) [ 29/Mar/11 ]

The repair should only be needed after an unclean shutdown.

Comment by Kirk Morales [ 29/Mar/11 ]

Just ran a repair running v1.8.0 on the same database and the temporary collections are gone!

Do you recommend running repair somewhat frequently then to take care of this should it happen again? E.g., should we run it once a day to clear out namespace issues and to compact indexes (especially since we're constantly deleting a lot of data from collections)?

Comment by Eliot Horowitz (Inactive) [ 29/Mar/11 ]

Did you do a repair after the unclean shutdown?

Comment by Kirk Morales [ 29/Mar/11 ]

There was an unclean shutdown. We were running 1.7.5 at the time and trying to setup sharding. I /think/ what happened was we were trying to write into the local mongod while reading from a mongos port. We started getting socket errors everywhere.

Comment by Eliot Horowitz (Inactive) [ 29/Mar/11 ]

Is it possible the database was corrupted?
Was there an unclean shutdown?

Generated at Thu Feb 08 03:01:15 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.