-
Bug
-
Resolution: Fixed
-
High (View bug fix roadmap)
-
7.5.0, 7.9.2, 7.7.4, 7.8.4, 7.10.2, 7.6.8, 7.11.2
-
7.05
-
16
-
Severity 2 - Major
-
31
-
-
Sometimes Jira development panel goes out of sync with FeCru. The behaviour was significantly improved in FeCru 4.5.1 (FE-6953). However, it is still happening and needs to be diagnosed/fixed in Jira.
Investigation
Fisheye/Crucible side
Investigation done by Fisheye/Crucible team revealed that Crucible correctly pushes events about new commits as well as correctly responds to Jira queries about particular issue key (team used reverse proxy to track communication between Jira and Crucible):
- Fisheye/Crucible sends update event about new commit or review state transitions to Jira http://localhost:8080/jira/rest/remote-event-consumer/latest/events/JiraReviewEvent:
{{ "sourceId": "486cef3d-38ab-3921-a562-50630743dd47", "sourceUrl": "http://localhost.charlesproxy.com:64669/fecru", "capabilities": ["JiraReviewEvent", "JiraIssueEvent", "atlassian-remote-event"], "initiator": { "emails": ["admin@admin.ad"], "username": "admin", "displayName": "admin", "resource": "http://localhost.charlesproxy.com:64669/fecru/user/admin", "avatar": "http://localhost.charlesproxy.com:64669/fecru/avatar/admin?s=48" }, "issueKeys": ["SIM-6"], "entity": { "id": "CR-3", "resource": "http://localhost.charlesproxy.com:64669/fecru/cru/CR-3" }, "state": "REVIEW"}
- Fisheye/Crucible sends a webhook to Jira http://localhost:8080/jira/rest/dev-status/1.0/webhooks/update-summary:
{{ "sourceId": "486cef3d-38ab-3921-a562-50630743dd47", "sourceUrl": "http://localhost.charlesproxy.com:64669/fecru", "capabilities": ["update-devtool-summary-request", "atlassian-remote-event"], "issueKeys": ["SIM-6"]}
- Jira requests update http://localhost:8060/fecru/rest/remote-link-aggregation/latest/aggregation?globalId=SIM-6:
{{ "targets": { "SIM-6": [{ "type": { "id": "repository" }, "count": 1, "objects": [{ "lastUpdated": "2018-08-28T13:59:10.000+0200", "count": 2, "commits": ["18ddb83188e5a3864911b54f8251132c4c7d0c86", "c7e1a9989b3cd83d99b3b3d2407e85ebd8a69082"], "type": "git" }] }, { "type": { "id": "review" }, "count": 3, "objects": [{ "lastUpdated": "2018-08-29T15:16:20.230+0200", "dueDate": "2018-09-04T13:30:00.000+0200", "state": "CLOSED" }, { "lastUpdated": "2018-08-29T15:50:56.718+0200", "dueDate": "2018-08-28T10:30:00.000+0200", "state": "CLOSED" }, { "lastUpdated": "2018-08-29T15:51:46.272+0200", "dueDate": "2018-07-22T13:11:23.000+0200", "state": "REVIEW" }] }, { "type": { "id": "branch" }, "count": 0, "objects": [] }] }}
- Jira also requests for commits etc.
Jira side
We can halt Jira update for couple minutes after communication was broken for some time. However after approximately 5 minutes Jira again asks for remote-link-aggregation and everything returns back to normal. After a night with interrupted connection, Jira stopped to update dev panel. It still was requesting Crucible for updates, but those were not appearing in Development panel. Changing state of related reviews (which produces an event) didn't help (data was not refreshed).
We suspect that it may be problem with timestamps in Jira cache.
Notes
1) If Jira rendered some data (even if they were stale), modal with 'more details' always shows correct data. It seems this modal always fetch data from Fisheye/Crucible directly and does not rely on any cached data.
2) This issue will possibly be there in multiple version of Jira, however, a good starting point will be when guaranteed delivery functionality was delivered (Jira Software 7.5.0).
- causes
-
-
- Closed
-
- mentioned in
-
Page Failed to load
-
Page Failed to load
-
Page Failed to load
-
Page Failed to load
-
Page Loading...
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
| Form Name | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
[JSWSERVER-16760] Jira development panel goes out of sync with fecru
Yes also happening for us from time to time with Jira 8.5.2 and Fisheye 4.8.0.
BTW This also happens from time to time between Bitbucket and Jira, so that only when a new commit against this Jira issue is performed the cache gets marked as outdated so that while the next view/access to this issue on Jira side the cache information is updated.
In our opinion a Action for a Jira issue is needed that, independent from the cache status, fetches the current commit data from all application link related SCM systems (e.g. bitbucket, fisheye, ..)
This is still an issue in Jira 8.5.2 and FeCru 4.8.0; If you have access please refer to the support tickets GHS-180063 and GHS-187121; Could this be looked into again? Thanks
Hi Erik,
I asked for a feature to clear the cache. That would be the final solution, in my opinion.
We had around 500k cache items in our database. I assume that that cache was just to big and caused the "out-of-sync". After removing all of those entries we fixed it. Of course, it must be rebuild afterwards. If users opening a issue with assigned commits, it will re-create the entry again.
Michael
We have seen similiar issues in integration between Jira and Github, any chance that this fix is applicable for that problem aswell?
@kbrazulewicz: actually they do ... see the Enterprise releases:
https://www.atlassian.com/blog/enterprise/introducing-enterprise-releases
@Krystian Brazulewicz if atlassian releases the FIX on the next JIRA version, that would be awesome and we would be thrilled.
ps: We pay a lot for Atlassian support and until this moment we didn't see any work done 
mparfianowicz as far as I can tell Jira Server does not release bugfix releases for past versions unless this is critical bug. According to Atlassian Server Bug Fix Policy this one here is at most "major".
This policy actually encourages customers to upgrade to most recent versions instead of staying on over 1 year old Jira 7.5.
gjoshi@atlassian.com did you really mean 7.13? Wasn't it already released (https://confluence.atlassian.com/jirasoftware/jira-software-7-13-x-release-notes-957981568.html)?
This issue still exists with Jira 8.5.3 and Fisheye 4.8.0.