From mboxrd@z Thu Jan 1 00:00:00 1970 From: Loic Dachary Subject: separating transient rados job errors Date: Thu, 11 Aug 2016 11:24:41 +0200 Message-ID: <57AC4459.6030108@dachary.org> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Return-path: Received: from relay2-d.mail.gandi.net ([217.70.183.194]:46319 "EHLO relay2-d.mail.gandi.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932271AbcHKJZP (ORCPT ); Thu, 11 Aug 2016 05:25:15 -0400 Sender: ceph-devel-owner@vger.kernel.org List-ID: To: Samuel Just Cc: Ceph Development Hi Sam, I propose that rados jobs known to experience occasional errors[1] are temporarily excluded from the nightlies while they are being fixed. And the nightlies could be green unless a new issue shows up. A wiki page listing the unreliable jobs[2] could be used to --filter-out the jobs it contains. When the problem is fixed, the job is removed from the wiki page and is no longer excluded from the nightlies. --filter-out <(curl -s http://tracker.ceph.com/projects/ceph/wiki/Rados-jewel.txt?key=$redmine_key | cut -f2- -d' ' | perl -n0e 'print join(",", split(/\S*\n/))') What do you think ? Cheers [1] http://tracker.ceph.com/issues/15785 fails 30% of the time [2] http://tracker.ceph.com/projects/ceph/wiki/Rados-jewel -- Loïc Dachary, Artisan Logiciel Libre