continuum-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Wendy Smoak (JIRA)" <j...@codehaus.org>
Subject [jira] Reopened: (CONTINUUM-1823) Fewer notifications when Subversion server is unreachable
Date Tue, 10 Mar 2009 21:38:13 GMT

     [ http://jira.codehaus.org/browse/CONTINUUM-1823?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Wendy Smoak reopened CONTINUUM-1823:
------------------------------------


It looks like CONTINUUM-634 would address the "for every project" part of the complaint, but
not the "every hour" part, so I don't think this is a complete duplicate.

As mentioned on the dev list, after the transient state changes (CONTINUUM-1864, CONTINUUM-1829,)
I'm still seeing a notification for every scheduled build using the 'bad-scm' example project.
 I can't take svn down, so I'm using a project with an intentionally incorrect svn url to
simulate it.  Another way to test this is to disable the subversion user account after adding
the project so that you get an authorization error when it tries to build.

http://www.nabble.com/Repeated-%27PREPARE-BUILD-ERROR%27-notifications-td22416935.html


> Fewer notifications when Subversion server is unreachable 
> ----------------------------------------------------------
>
>                 Key: CONTINUUM-1823
>                 URL: http://jira.codehaus.org/browse/CONTINUUM-1823
>             Project: Continuum
>          Issue Type: Improvement
>            Reporter: Erik Drolshammer
>
> By default Continuum sends notifications when the state of a project changes, so two
consecutive build failures will result in only one email. However, when the build failure
is caused by an unreachable Subversion server a notification is sent every hour, for every
project (See output below). This resulted in approx. 1500 emails in my ci-folder after the
weekend. Which, imho, is a bit unfortunate. 
> Would it be possible to adopt the principle advocated by TestNG; that this is caused
by a _single_ failure and all other failures are derived failures which are given the status
SKIPPED in TestNG terminology? 
> In other words, I propose the following extension: 
> Create a list of SCM-repositories known to Continuum. 
> Run a simple availability test on each of these and set status "OK" or "Unavailable".
 
> Projects that use an unavailable repository can thus be SKIPPED until the repository
is available again. 
> The repository test can be run e.g. once an hour (preferable configurable) and should
only send notifications when the state changes. 
> Further, it would be nice if it was possible to set up a list of notifiers separate from
the project notifiers like it is for projects now. It would typically be interesting for system
administrators to subscribe to email notifications for unavailable SCM repo. (A cheap surveillance
solution.) 
> Build Error:
> ****************************************************************************
> Provider message: The svn command failed.
> Command output: -------------------------------------------------------------------------------
> svn: PROPFIND request failed on '/svn/projectName/trunk/'
> svn: PROPFIND of '/svn/projectName/trunk/': could not connect to server (https://10.0.0.10)


-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Mime
View raw message