JIRA corruption

[Daiji]

I'm unable to check on the status of a JIRA issue I filed some time
ago because of JIRA corruption.

See:
http://jira.openqa.org/browse/SRC-517

The stack trace starts with

java.lang.RuntimeException: Found more than one user with name
'aksonov'; ids [GenericEntity:OSUser][passwordHash,null][name,aksonov]
[id,23599], [GenericEntity:OSUser][passwordHash,null][name,aksonov][id,
23598]
at com.opensymphony.user.provider.ofbiz.OFBizAbstractProvider.findUser
(OFBizAbstractProvider.java:159)

Can someone with appropriate privileges look into this? I'd hate to
have to re-file the JIRA from scratch... the bug was titled "Add
support for running Selenium HTML IDE files as tests that hit the
Selenium RC server" and had a contribution I created that would
basically act as a Selenese driver. I wanted to check on its state
and also ensure that it would work properly under the latest Selenium
RC release.

Thanks,

- Daiji

[Daiji]

I'm wondering about not having received a response in a week to this
post. Is it that:
1) I'm sending to the wrong place? (If so, please advise the correct
place I should post)
2) No one cares?
3) No one noticed?
4) No one knows what do do about this?
5) It's low priority? (If so, I guess I should re-post the JIRA and
save its description just in case this happens again?)

I hope I'm not annoying anyone by re-posting; but I was curious why it
seemed that every other post that followed this got some sort of
response, but none for this one.

Thanks,

- Daiji

[Santiago Suarez Ordoñez]

Actually, I think this is not very important. There's a move going on at the moment, the idea is to end up using google code's bug tracker tool, so probably no one wants to put much effort on it.

Patrick, any thoughts?

Saludos
Santi

[Simon Stewart]

Umm... we need the database to be valid before I can do the move :)

Simon

[Daiji]

Ok, so I've re-filed the bug as http://jira.openqa.org/browse/SRC-753
with some details on what I've done with the code since then. FYI
the original was assigned to Patrick L.

- Daiji

On Sep 7, 5:08 am, Simon Stewart <simon.m.stew...@gmail.com> wrote:
> Umm... we need the database to be valid before I can do the move :)
>
> Simon
>
> On Sat, Sep 5, 2009 at 2:34 PM, Santiago Suarez Ordoñez
>
>
>
> <santi...@gmail.com> wrote:
> > Actually, I think this is not very important. There's a move going on at the
> > moment, the idea is to end up using google code's bug tracker tool, so
> > probably no one wants to put much effort on it.
>
> > Patrick, any thoughts?
>
> > Saludos
> > Santi
>

[Patrick Lightbody]

Daji,
Quick question: did you have to create a new username to get JIRA
working for you, or did it suddenly just start behaving for you?

Patrick

[Daiji]

Hi Patrick!

I actually never had a problem with JIRA not working. I only had a
problem accessing that one specific bug. It's possible to get the XML
content of the bug if you don't go directly to the bug (i.e. obtain it
from the list view). It's also possible to get the attachments of the
bug if you do some hacked-up URL manipulation using the internal db id
(I figured out how to do this by cross-comparing against other bugs
with attachments). I suspect it's really just because of that one
commenter (aksonov) apparently having two userids somehow (either in
the userid field or the name field... you can prob just query the JIRA
db). I suspect that if you deleted that specific userid, the problem
would go away, but I'd probably fiddle with a test JIRA installation
to verify that before really doing that. It also might be possible to
resolve this if you're really careful by manipulating the db directly
to merge the two aksonov user accounts.

- Daiji

On Sep 16, 12:27 am, Patrick Lightbody <patr...@lightbody.net> wrote:
> Daji,
> Quick question: did you have to create a new username to get JIRA
> working for you, or did it suddenly just start behaving for you?
>
> Patrick
>
>
>

> On Tue, Sep 15, 2009 at 7:29 PM, Daiji <dtakam...@gmail.com> wrote:
>
> > Ok, so I've re-filed the bug ashttp://jira.openqa.org/browse/SRC-753

[Daiji]

Actually, now that I look at that stack trace more closely, it's
specifically from identifying the issue's watchers (it really should
just look up the count, instead of identifying individuals, imo).

Btw, if you need someone to help figure out how to reproduce this
problem in a fresh-install JIRA database and/or how to resolve this
sort of db corruption, let me know. But I imagine Atlassian should
offer to help you guys out here.

- Daiji

java.lang.RuntimeException: Found more than one user with name
'aksonov'; ids [GenericEntity:OSUser][passwordHash,null][name,aksonov]
[id,23599], [GenericEntity:OSUser][passwordHash,null][name,aksonov][id,
23598]
at com.opensymphony.user.provider.ofbiz.OFBizAbstractProvider.findUser
(OFBizAbstractProvider.java:159)

at
com.opensymphony.user.provider.ofbiz.OFBizCredentialsProvider.handles
(OFBizCredentialsProvider.java:114)
at org.openqa.sass.osuser.SassOfbizProfileProvider.getPropertySet
(SassOfbizProfileProvider.java:24)
at com.opensymphony.user.Entity.getPropertySet(Entity.java:87)
at com.opensymphony.user.User.getFullName(User.java:83)
at com.atlassian.core.user.BestNameComparator.compare
(BestNameComparator.java:35)
at java.util.Arrays.mergeSort(Arrays.java:1284)
at java.util.Arrays.sort(Arrays.java:1223)
at java.util.Collections.sort(Collections.java:159)
at
com.atlassian.jira.issue.watchers.DefaultWatcherManager.getCurrentWatchList
(DefaultWatcherManager.java:59)
at
com.atlassian.jira.web.component.issuesummary.IssueSummaryWebComponent.getHtml
(IssueSummaryWebComponent.java:63)
at org.apache.jsp.decorators.issuesummary_jsp._jspService
(issuesummary_jsp.java:545)

[Patrick Lightbody]

Thanks for the offer for help. The OpenQA infrastructure (Confluence,
Clearspace, and JIRA) have all slowly gotten worse and worse. I think
it all stems from our unified LDAP/SSO login system.

Simon is working on migrating the JIRA bugs for Selenium to Google
Code. These days I'm inclined to agree that the less we have to host
and manage ourselves, the more time we can spend on writing code for
the actual _project_. So even though I love JIRA, I'm OK with using
Google Code going forward :)

2009/9/16 Daiji <dtak...@gmail.com>: