[SERVER-5909] renameCollection op is blocking, and killOp does not kill Created: 23/May/12  Updated: 09/Jul/16  Resolved: 18/Sep/12

Status: Closed
Project: Core Server
Component/s: Concurrency, Performance, Stability, Tools
Affects Version/s: 2.0.4
Fix Version/s: None

Type: Bug Priority: Major - P3
Reporter: RL Assignee: Andy Schwerin
Resolution: Done Votes: 0
Labels: killop, renameCollection
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

1) SMP Thu Mar 22 08:00:08 UTC 2012 x86_64 x86_64 x86_64 GNU/Linux
2) A 5 node replica set environment.
3) 99% idle IO stat
4) 70GB collection


Issue Links:
Related
related to SERVER-7059 renamedatabase blocks Closed
Operating System: Linux
Participants:

 Description   

During the execution of:
> db.runCommand(

{ renameCollection: "mydb.oldname", to: "mydb.newname" }

What happened:
Suddenly all the operations started to stack up. Collection in question was over 70GB.

What was expected:
It was expected that the operation would yield. Or at the very least such behavior should be documented...

Related:
db.killOp() - did NOT kill the operation
db.currentOP() - reported the OP as killed (killed:true), yet all the other ops never got executed. Additionally all the new ops continued to fill up the queue.



 Comments   
Comment by Eliot Horowitz (Inactive) [ 18/Sep/12 ]

This is as expected.
Its impossible to be yield as we rename an entire db.
SHould be doc'ed though.

Generated at Thu Feb 08 03:10:13 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.