commit 843ff506f7a1ea983bbfcb4897f4546d3e35fcee Author: Dan Callaghan Date: Thu Oct 4 15:58:52 2018 +1000 Automatic commit of release 26.0 commit baf41fe9d7c79e84ba01ef9ff9fb88b3fc6b958e Author: Dan Callaghan Date: Fri Sep 28 16:42:56 2018 +1000 Automatic commit of release 26.0rc1 commit b7ab9d7e0017970ee1f878e6022b335eeabd4701 Author: Dan Callaghan Date: Fri Sep 28 16:02:05 2018 +1000 release notes for 26.0 Change-Id: Ia47d5714cd61d9f13a7779f810a9e08e30fd324c commit 82779a6f62735e91b31b54014988b1bc32766151 Author: Dan Callaghan Date: Fri Sep 28 16:02:26 2018 +1000 docs: fix wrong heading level for 25.5 release notes Change-Id: I7343393436d34a6bfa86e20cdd0aa3e8b27ce79a commit 7239ee87fc9b1fbe07c994ef942d98801ddab333 Author: Dan Callaghan Date: Mon Sep 24 13:54:06 2018 +1000 expand device.fw_version column Bug: 1630884 Change-Id: I04a1d5ee2c9aeabe995b828deb20ffeaa67a880e commit 23579852806220d6a13bb57a5ccc031b258571e3 Author: Dan Callaghan Date: Mon Sep 24 13:01:47 2018 +1000 redirect TESTOUT.log requests to taskout.log For compatibility with user's scripts which were written assuming the Beah filename TESTOUT.log, to avoid having to update them (or add conditionals) when recipes are using Restraint instead of Beah. Bug: 1622805 Change-Id: I1fcb3385b077d7947976471f100734a49c9c58ee commit 8773dedf0a5c588fd82f350d4e21b6a57b70c87c Author: R贸man Joost Date: Wed Sep 5 16:36:22 2018 +1000 Make beaker-wizard recognize new CVE number scheme A syntax change has occurred with CVE ID numbers. The new syntax is in variable length starting with at least four digits: https://cve.mitre.org/cve/identifiers/syntaxchange.html This patch makes beaker-wizard compliant to accept these numbers. Bug: 1624909 Change-Id: I9c1773dcfc29b80cedcf02ffca1c0f94cd8cb82a commit ee4e03be745d84aff298d77d67984b28b5de22a4 Author: Jeffrey Bastian Date: Tue Sep 4 16:31:02 2018 -0500 verify /usr/libexec/platform-python is Python 3 Bug: 1625234 Change-Id: I473ad7dffc18b89ec58b9ac231f6cdae2bd589a8 Signed-off-by: Jeffrey Bastian commit 4146878b8ea90ef0d47efab892516e2344e39d71 Author: Dan Callaghan Date: Wed Aug 29 11:18:02 2018 +1000 add '-d 1' to yum install commands on RHEL3 and RHEL4 This avoids printing a lot of hashes which can take a long time if the serial console is slow. Bug: 1622753 Change-Id: I8b66b5b588b62565a9e53d99268f24403c3ee406 commit f4af3a1cfe28601f30d094569ce8283f7090c248 Author: Jacob McKenzie Date: Tue Sep 4 11:41:05 2018 +1000 Remove MD5 from anamon file upload Anamon was still populating md5 field in xmlrpc uploads, but beaker no longer uses this field. Change-Id: I0800269ef5d7aca35dfd4167f004e05711e7a38a commit a4372b3c7d88468c2e6ad9c3799d2c9f80395733 Author: Dan Callaghan Date: Wed Aug 29 10:30:22 2018 +1000 tests: ensure that beaker-init knows about the latest schema version Change-Id: I3cb59a65cccc46b39f8ec6611e21a1965a67cde3 commit 9f72fb830614869470f56c7d3ca847d2b180191b Author: Dan Callaghan Date: Fri Aug 24 14:08:18 2018 +1000 docs: use built-in extlinks extension instead of rolling our own Change-Id: I2f2cd20e4f3763afd5ef74830df7638f7116923c commit 880b7656111c2f65fc4fda6306a9243f1b14bd7f Author: Dan Callaghan Date: Wed Aug 8 17:02:00 2018 +1000 fix DECIMAL columns if their scale is wrong The default MySQL precision and scale for DECIMAL (or Numeric in Python land) is 10,0 but for unexplained reasons the Red Hat production instance of Beaker has these columns as DECIMAL(10,2) instead. This patch makes the precision and scale explicit for all such columns, and adds extra checks to test_database_migration to catch inconsistencies. A migration is added to fix any old Beaker instances which have this problem. Bug: 1600281 Change-Id: Ia7b94c40fca81adb2c0075f1caf384f9e1c64aa3 commit 9f16469b1a0536752ef36b1675cbda8d778cc27e Author: Dan Callaghan Date: Mon Jul 23 14:40:46 2018 +1000 tests: use /distribution/check-install for completeness In these test cases, the actual task name is incidental, it could be anything. However since /distribution/check-install is the way of the future, let's update them all to use the new name for completeness. Bug: 1188539 Change-Id: I88391273b58cb83bb9943fb3b72e97545feeed0a commit 257d963786d76a25dbabbb403d06203d34b0ddbd Author: Dan Callaghan Date: Fri Jul 20 17:12:58 2018 +1000 docs: replace all incidental references to /distribution/install The new /distribution/check-install is the way of the future, we want people to see that when they read examples and explanations in the docs. Bug: 1188539 Change-Id: Ia23df4fe1600224742242245f02bb43844b5bb07 commit 292f9efedfd29e91cea777b6ee07c6df7ac0ad87 Author: Dan Callaghan Date: Fri Jul 20 17:07:51 2018 +1000 docs: describe /distribution/check-install Also retain the description of /distribution/install as it has a long life ahead of it yet. Removed the suggestion that the install task uploads Anaconda logs, because it actually didn't really upload many of them. Anamon already does a much better job of this, and the Installation tab of the recipe page is generally the right place to look first. Bug: 1188539 Change-Id: Ia82c7780d75e2e7ed36ae27a69f01477703017ed commit 09ff3d9f66a830568f3580f0029e441d4f5527a8 Author: Dan Callaghan Date: Fri Jul 20 16:53:53 2018 +1000 STABLE tagging script: use /distribution/check-install It seems highly unlikely that anybody is still even using this Beaker facility, but might as well update it for completeness anyway. 馃し Bug: 1188539 Change-Id: Ie8043524ca0881b51cdef8f34656097ddac0043f commit ec2c85fc3947dad141b9d94f87adf24bd56e65a2 Author: Dan Callaghan Date: Tue Jul 24 14:23:08 2018 +1000 bkr workflows: use /distribution/check-install ... instead of /distribution/install, on Fedora 29 and up, and RHEL8 and up. On older distros we continue scheduling /distribution/install to avoid unexpectedly changing the task name that will appear in people's results. Bug: 1188539 Change-Id: I87870a21fc80aafe73cbf77802506c3407b0431a commit a8d94be2e6d243ecf73221090bbd0b41570ddfd9 Author: Dan Callaghan Date: Fri Jul 20 16:48:11 2018 +1000 inventory jobs: use /distribution/check-install ... instead of /distribution/install. Bug: 1188539 Change-Id: I98fb6314960e8d1bcc1aa189ecd1d56ba149bc1a commit bb557a00dc39c1e1f2372e3e449c1fc35bd1fff6 Author: Dan Callaghan Date: Fri Jul 20 15:05:31 2018 +1000 reserve workflow: use /distribution/check-install ... instead of /distribution/install, as the first task in the recipe. Bug: 1188539 Change-Id: Iffb52fe554b1b1f40b5711403288a3ae040e6393 commit 8f03655fb3fd141b24735d1008971aa60b39b66f Author: Dan Callaghan Date: Fri Aug 3 11:31:15 2018 +1000 tests: cover the "external watchdog script" functionality Change-Id: I421b074e17a3e02541e293eba6d9267da5b1c859 commit ce4abe9289c2b3524f7e754cb25483dd89ce356f Author: Dan Callaghan Date: Fri Aug 3 11:04:05 2018 +1000 tests: remove obsolete log rotation settings from LC config These settings were removed in commit b0b17428, when we switched to logging through syslog instead. Change-Id: I423b768f6dcd95cdbbdf962ee49728f26f9cfa10 commit e37b991c7b2a0939af22b86c4d1fe63ba20825e7 Author: Dan Callaghan Date: Thu Aug 2 17:37:17 2018 +1000 restructure beaker-watchdog using gevent, tidy up error handling This adjusts the beaker-watchdog daemon to closely match the structure of beaker-provision. The main loop is a polling loop which spawns and manages a set of greenlets. The actual aim of this patch is to clean up the error handling behaviour and shutdown behaviour, so that it doesn't crash mysteriously if an error occurs in any of the loops. Bug: 991269 Change-Id: I064d4253c9a933601c2f6da2b5a3296dc746ad07 commit 239959f100cb798f0736e37d901129f9776e4457 Author: Dan Callaghan Date: Thu Aug 2 14:22:04 2018 +1000 remove verbose traceback printing In theory, these detailed tracebacks showing all local variables in every stack frame, can help to debug unexpected problems in production. However in practice it turns out that they typically just make the logs totally unreadable (each verbose traceback is many screenfuls long) and they generally do not tell us anything that a normal traceback plus carefully worded log messages can't already tell us. Therefore to simplify things, let's remove the support and use regular logger.exception() calls instead. Change-Id: I362fc09ad23a2407ba10bd6eddc04158e1e152bb commit 563bdc20659dbcbbf5a72c337bb1fe1adf2c9119 Author: Dan Callaghan Date: Tue Aug 7 17:11:34 2018 +1000 fill in custom packages into %packages in sorted order This test was failing intermittently because the custom packages are being filled into the kickstart in database order, which is arbitrary. It would be nicer if we could put them into the kickstart in the same order they appear in the job XML, but we do not persist the ordering so the best we can do instead is obtain a consistent ordering by sorting them on the way out. This should fix the test, but it should also be nicer for humans too. The packages will not appear in the kickstart in a totally random-looking order. Change-Id: Ia3d957232219c81d0272406643e4178a79bac87f commit 10856277c595664e1844710cdd3a9f3a58e370d7 Author: R贸man Joost Date: Thu Jul 19 16:08:26 2018 +1000 Do not pre-fill task requirements in the kickstart Because we're switching to Restraint as the default test harness for Fedora 29, RHEL8 and higher, we can stop populating packages in the kickstart. Restraint will install all task dependencies before it executes the tasks. Bug: 1589614 Change-Id: I9cbce2137955d3f1a83130199753f2758cfd39c4 commit 010597cbd9f57575c3ecdd4c0cc61d2b3566bca8 Author: Dan Callaghan Date: Wed Jul 25 17:07:52 2018 +1000 show lists in sorted order on the task page These fields are lists of string identifiers. They come back from the server in arbitrary database order. Let's sort them alphabetically so they look nice for the humans. This also helps the tests too. I found this because an unrelated patch was causing a test failure, like this: ====================================================================== FAIL: test_exclusive_releases (bkr.inttest.server.selenium.test_tasks.TestSubmitTask) ---------------------------------------------------------------------- Traceback (most recent call last): File "/usr/lib/python2.7/site-packages/bkr/inttest/server/selenium/test_tasks.py", line 269, in test_exclusive_releases 'RedHatEnterpriseLinux6\n' AssertionError: u'RedHatEnterpriseLinux7\nRedHatEnterpriseLinux6' != 'RedHatEnterpriseLinux6\nRedHatEnterpriseLinux7' It was because my patch had caused RedHatEnterpriseLinux7 to be inserted into the database earlier than RedHatEnterpriseLinux6, thus causing the server to start returning the values in a different order. Change-Id: I7e9c906c73420bf2f1feebdf057a9b92b9b9fe63 commit ea0711c0bad6bd21c93e4921700a869245ee0d67 Author: Dan Callaghan Date: Mon Jun 4 17:25:58 2018 +1000 replace job dirty_version/clean_version columns with a simple is_dirty flag Bug: 1297603 Change-Id: Ic1ee5d3d4f17a201fe583a0d49e2c574a305fcf2 commit 7eac8d96e9dd81d7e08164761c84a2ab61b250d0 Author: R贸man Joost Date: Fri Jul 6 14:39:10 2018 +1000 Make restraint the default harness Bug: 1589610 Change-Id: Ia4e86c7af75b44d3b45127722f99b2affc6f1c61 commit 2c348907aa958c3ac77bee55b15ee8189c6fa0c6 Author: R贸man Joost Date: Tue Jul 10 16:00:22 2018 +1000 Use ks_meta variable to disable default harness repository In order for users to prevent Beaker from including the default harness repository, this patch introduces a new ks_meta variable 'no_default_harness_repo'. When this variable is set, the default Beaker harness repository is not configured in the kickstart. Bug: 1599136 Change-Id: I8a910d851a850b8169620c93c44dab8571fcd4c9 commit 2e91025941e3fb2e14e3e8a685f1873c2b7dbb4b Author: R贸man Joost Date: Wed Jun 13 11:00:43 2018 +1000 Remove reference to stack overflow The idea of posting questions for Beaker on Stack Overflow and therefore building up a repository of Questions/Answers which can be found by Google never took off. Change-Id: Ic704bf22795e54b82741087ca52c1f41561ba6a8 commit 9a5c1f052b9d5798f7e16ad334c2f9a964bdfb91 Author: R贸man Joost Date: Wed Jun 13 10:58:52 2018 +1000 Add link to PDF version of our user guide Change-Id: I1880d5d133c05035a1f63216a4c05e82a7d7c747 commit 415ffce0d6b0f3b3250157eef2c3f7803498162b Author: Matt Tyson Date: Tue Mar 13 07:10:25 2018 -0400 bkr workflow-reserve: reserve a system Fix broken doc markup. Options such as --machine, --arch, --distro are options that belong to the bkr client itself. The proper cross referencing needs to be set up so Sphinx can generate links back to the bkr client docs, instead of looking for the options in the current doc. Change-Id: I702d4fc3b8d06d7dc54df9c731fec58f4dfee6c3 commit 00b158e1c3ce75aa7e5b199a8d334215884040ee Author: Matt Tyson Date: Tue Mar 13 05:56:29 2018 -0400 uncomment whats-new/next for develop Change-Id: I0e7e6092d64f168f1f18176b89349635fad5aa39 commit 65d5b4aacbd08c0904bcfd5ec12c16d1d8ebcc76 Author: Matt Tyson Date: Tue Feb 20 23:04:55 2018 -0500 bkr workflow-reserve: reserve a system Bug: 1165960 Change-Id: I0826647a69ee705094f0cc0d10b64b22169b0786