| This page is primarily of interest to SRB developers. The link below takes you to the SRB status page for the automatic build and test runs which operate continuously (via a Tinderbox system). These builds being done from our CVS source tree (pre-release) (copied shortly after the hour) on four test hosts (two Linux, an OS X, and a Solaris).
The name of the build is SRB2_0_0rel which actually means the current SRB source (when we set up CVS, we named our tree poorly).
If one of the four hosts isn't showing up on the web display at all, it generally means the tinderbox client script has stopped. We sometimes do this to examine a problem.
These are configured to run various autotest options, one is 'autotest.sh 0 1' which does an install (install.pl) and some simple tests. Sometimes, based on a configuration file, the scripts run the full 'autotest.sh', which does a more comprehensive set of tests and takes 9.5 hours or so.
We have one host (one of the Linux hosts) configured as a normal tinderbox system. If there is a failure on it, it will stop and remove everything, and try again. So for that host, if it is red, please do not checkin new changes until we resolve and fix what is wrong. Also, after you commit changes, you should check that the builds are still working OK. The cvs repository is copied to the QA hosts two minutes after each hour, so after you commit, QA runs that start after that should have your updates. The time on the QA hosts can be a few minutes off.
Other QA hosts are configured for the tinderbox script to exit if there is a failure so that we can login and check into the problem. If they fail, the display will show as a long yellow block, with a short Time value but with a much earlier start time. After a day or so these will disappear so if you see fewer than 4 hosts that's most likely what happened.
All the hosts are configured to run various tests at different times of the day. We can also configure them to use older versions out of CVS.
System currently unavailable