[SERVER-13972] connections_opened.js failure Created: 04/Apr/14 Updated: 11/Mar/15 Resolved: 30/Apr/14 |
|
| Status: | Closed |
| Project: | Core Server |
| Component/s: | Testing Infrastructure |
| Affects Version/s: | None |
| Fix Version/s: | 2.6.2, 2.7.0 |
| Type: | Bug | Priority: | Minor - P4 |
| Reporter: | Randolph Tan | Assignee: | Spencer Brody (Inactive) |
| Resolution: | Done | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Attachments: |
|
||||||||||||||||||||
| Issue Links: |
|
||||||||||||||||||||
| Operating System: | ALL | ||||||||||||||||||||
| Backport Completed: | |||||||||||||||||||||
| Participants: | |||||||||||||||||||||
| Linked BF Score: | 0 | ||||||||||||||||||||
| Description |
|
http://buildlogs.mongodb.org/mci_0.9_windows_64_2k8_debug/builds/43780/test/noPassthroughWithMongod_2/connections_opened.js
http://buildlogs.mongodb.org/mci_0.9_windows_64_debug/builds/42074/test/noPassthroughWithMongod_0/connections_opened.js
|
| Comments |
| Comment by Githook User [ 16/May/14 ] | |||
|
Author: {u'username': u'stbrody', u'name': u'Spencer T Brody', u'email': u'spencer@mongodb.com'}Message: | |||
| Comment by Eric Milkie [ 16/May/14 ] | |||
|
It is okay to backport. Unfortunately, that commit has no associated server ticket so it will be hard to track the backport process. | |||
| Comment by Siyuan Zhou [ 16/May/14 ] | |||
|
It seems this issue happened on v2.6 branch. https://mci.10gen.com/ui/task/mongodb_mongo_v2.6_windows_64_2k8_debug_f076d57d427e7fde945303aaec32ba594de5872c_14_05_16_00_22_40_noPassthroughWithMongod_windows_64_2k8_debug spencer - is it appropriate to backport this change to 2.6? | |||
| Comment by Spencer Brody (Inactive) [ 30/Apr/14 ] | |||
|
That last commit should have fixed this. I'm going to resolve this issue until/unless this happens again. | |||
| Comment by Spencer Brody (Inactive) [ 29/Apr/14 ] | |||
|
https://github.com/mongodb/mongo/commit/b64939f3ac16843525adf6f0e893a78409ddaf16 | |||
| Comment by Eric Milkie [ 28/Apr/14 ] | |||
|
If a server is overloaded and you'd like to be able to log in with the shell to save it, you probably would like to have a connect timeout longer than 5 seconds. I think it should be raised regardless of pending refactoring, as long as it is easy to change in the shell code itself. | |||
| Comment by Spencer Brody (Inactive) [ 28/Apr/14 ] | |||
| Comment by J Rassi [ 26/Apr/14 ] | |||
|
this test is spawning hundreds of mongo shell processes on the same machine as the server, and in five seconds expects all of them to complete their connect() calls (and then release the socket mutex to the respective connectbg threads). this seems reasonable to expect for the release builds, but not (at all!) for the debug builds. after reflecting, i think this test should be disabled on the debug buildvariants. i don't think we should shoehorn connect timeout logic into the network subsystem right now, and that we should just wait to address the underlying issue until we give that subsystem a proper refactor. | |||
| Comment by Eric Milkie [ 25/Apr/14 ] | |||
|
Yes. | |||
| Comment by Spencer Brody (Inactive) [ 25/Apr/14 ] | |||
This verifies that the failure is a connection timeout. Not clear why the connection attempt is taking so long though... milkie - do you think we should increase the default connection timeout from 5 seconds? | |||
| Comment by Spencer Brody (Inactive) [ 23/Apr/14 ] | |||
|
Hopefully https://github.com/mongodb/mongo/commit/bd1f40ab79656b14874bf3c496ad93fcffa89413 should reveal if this is a simple timeout. | |||
| Comment by Randolph Tan [ 07/Apr/14 ] | |||