> On a similar note, it seems like bug 1195415 is among the leading > causes of regression failures now. This manifests not as failures > in run-tests.sh but as core files found afterward. There's little > to be gained by setting V-1 for an unrelated bug we already know > about. Perhaps we should modify the Jenkins regression scriptlet > so that it collects those cores *silently* until we find and fix > that bug (which I'll be working on shortly BTW). I've submitted http://review.gluster.org/#/c/10157/ as another alternative. Here's the commit message: = This started as a way to identify which test created new core files, = since that's a critical piece of debugging information that's missing = when we only check for cores at the end of a run. It also exits = *immediately* either on bad status or discovery of a new core file. = This allows the run to be retried or aborted quickly, to reduce the = latency of all jobs in the regression-test pipeline. Reviews would be most welcome. We're currently setting a record for the lowest regression-test success rate ever. Let's get that pipeline moving again. _______________________________________________ Gluster-devel mailing list Gluster-devel@xxxxxxxxxxx http://www.gluster.org/mailman/listinfo/gluster-devel
