[SERVER-9231] sharding/gridfs.js failing Created: 03/Apr/13  Updated: 11/Jul/16  Resolved: 08/Jul/13

Status: Closed
Project: Core Server
Component/s: Testing Infrastructure
Affects Version/s: None
Fix Version/s: 2.5.1

Type: Bug Priority: Major - P3
Reporter: Ian Whalen (Inactive) Assignee: Mathias Stearn
Resolution: Done Votes: 0
Labels: buildbot
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified

Attachments: Text File gridfs.txt    
Issue Links:
Duplicate
Related
related to SERVER-10039 Sorting on chunks not applied in differ Closed
Backwards Compatibility: Fully Compatible
Operating System: ALL
Participants:
Linked BF Score: 0

 Description   

http://buildlogs.mongodb.org/Linux%2064-bit%20DEBUG/builds/2039/test/sharding/gridfs.js

fileObj: {
	"_id" : ObjectId("515c83346f0ec0572698de13"),
	"filename" : "mongod",
	"chunkSize" : 262144,
	"uploadDate" : ISODate("2013-04-03T19:30:19.963Z"),
	"md5" : "63776150205c0a8ea74832560e3e29a4",
	"length" : 208803033
}
filemd5 output: {
	"md5state" : BinData(0,"AAAgAQAAAAC9IOle7Rm24RV8I9G9vHxmBwAAAAAAAABw6E9O6H8AACcAAAAAAAAAkOhPTuh/AAAIvzhi6H8AAAQAAAAAAAAAJwAAAAAAAADg6E9O6H8AAA=="),
	"numChunks" : 9,
	"md5" : "326ae24f1c65a01eae699bb56e28a9dc",
	"ok" : 1
}
assert: ["63776150205c0a8ea74832560e3e29a4"] != ["326ae24f1c65a01eae699bb56e28a9dc"] are not equal : undefined
Error: Printing Stack Trace
    at printStackTrace (src/mongo/shell/utils.js:37:15)
    at doassert (src/mongo/shell/assert.js:6:5)
    at Function.assert.eq (src/mongo/shell/assert.js:32:5)
    at testGridFS (/home/yellow/buildslave/Linux_64bit_DEBUG/mongo/jstests/sharding/gridfs.js:33:12)
    at /home/yellow/buildslave/Linux_64bit_DEBUG/mongo/jstests/sharding/gridfs.js:59:1
Wed Apr  3 15:30:20.048 JavaScript execution failed: ["63776150205c0a8ea74832560e3e29a4"] != ["326ae24f1c65a01eae699bb56e28a9dc"] are not equal : undefined at src/mongo/shell/assert.js:L7
failed to load: /home/yellow/buildslave/Linux_64bit_DEBUG/mongo/jstests/sharding/gridfs.js



 Comments   
Comment by Mathias Stearn [ 08/Jul/13 ]

Resolving. The dbtemprelease issue appears to be unrelated to this issue even though it failed in the same test. It is still open as SERVER-9539.

Comment by Greg Studer [ 08/Jul/13 ]

I think SERVER-10039 fixes the issue below, looks like we were refreshing with the latest collection version but still had a mismatch.

>Sharding-related - not sure why we didn't refresh our versions when we got a repeated error here...
http://buildlogs.mongodb.org/Linux%2064-bit%20debug%20dur%20off/builds/2321/test/sharding/gridfs.js

Throwing this back to @mathias to close if he thinks the second part of this is related, since I'm not sure I fully understand the dbtemprelease stuff he's talking about.

Comment by Mathias Stearn [ 07/Jun/13 ]

Sorta. The issue is that dbtemprelease's destructor is implicitly noexcept in c++11 so throwing from it results in a call to terminate(). While that does need to be fixed, I think that the exception being thrown is a SCE so solving this ticket would also cause that one to stop failing. I reopened that ticket because even once that test stops failing we still need to add noexcept(false) to the dbtemprelease destructor to consider that ticket closed.

Comment by Greg Studer [ 07/Jun/13 ]

These look like different failures:

Sharding-related - not sure why we didn't refresh our versions when we got a repeated error here...

http://buildlogs.mongodb.org/Linux%2064-bit%20debug%20dur%20off/builds/2321/test/sharding/gridfs.js

This just looks something locking-related in the MD5Cmd exploded with a sigabort, taking down the server.

http://buildbot-special.10gen.com/builders/Nightly%20OS%20X%2064-bit%2010.7%20C%2B%2B11%20DEBUG%20Subscription/builds/91/steps/shell_3/logs/stdio

Comment by Eric Milkie [ 23/May/13 ]

Another failure! With a different signature. This is from the C++11 builder.
http://buildbot-special.10gen.com/builders/Nightly%20OS%20X%2064-bit%2010.7%20C%2B%2B11%20DEBUG%20Subscription/builds/91/steps/shell_3/logs/stdio

Comment by Eric Milkie [ 23/May/13 ]

Good news! Failed again, and we got the extra debugging info now:
http://buildlogs.mongodb.org/Linux%2064-bit%20debug%20dur%20off/builds/2321/test/sharding/gridfs.js

Comment by auto [ 08/May/13 ]

Author:

{u'date': u'2013-05-08T17:05:10Z', u'name': u'Mathias Stearn', u'email': u'mathias@10gen.com'}

Message: Extra debugging info for SERVER-9231
Branch: master
https://github.com/mongodb/mongo/commit/fbadbba833f283a846d3f9ac03f765ef9d0b9b97

Comment by Eric Milkie [ 25/Apr/13 ]

I rescind my comment – we now believe something else is going wrong, either with the test or with the filemd5 command or with sharding itself.

Comment by Eric Milkie [ 25/Apr/13 ]

Ok then you should definitely make deletes synchronous with the balancer (via the config) and then wait for it to be finished before you do hash comparisons.

Comment by Mathias Stearn [ 25/Apr/13 ]

Part of the point of the test is to ensure that it works correctly during migrations. Turning off the balancer would defeat that.

Comment by Eric Milkie [ 03/Apr/13 ]

Should the balancer be turned off during this test? Or wait for it to finish before doing md5 checks?
The test failed because we were doing some migration deletes and it caused the hashes to be different.

Comment by Ian Whalen (Inactive) [ 03/Apr/13 ]

possibly related to/dupe of SERVER-9229

Generated at Thu Feb 08 03:19:46 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.