Skip to main content

Browse the archive

https://github.com/wikimedia/operations-puppet
12 April 2024, 10:39:33 UTC
sort by:
Revision Author Date Message Commit Date
99019bf Guillaume Lederrey 08 February 2016, 10:44:39 UTC Ship Elasticsearch logs to logstash WIP - do not merge yet Introduce gelf4j as a log4j appender to send logs to logstash. Sending log to logstash is disabled by default and can be activated with $elasticsearch::send_logs_to_logstash. * deploy necessary packages (liblogstash-gelf-java, libjson-simple-java) * create symlinks from jars to /usr/share/elasticsearch/lib * add necessary configuration to /etc/elasticsearch/logging.yml Still to do (need some help): * understand where in hiera the logstash server should be configured * test in labs Bug: T109101 Change-Id: I683bcb75c91e6c2f4dd648517435a477ab23c2a7 11 February 2016, 09:51:45 UTC
24cc170 Subramanya Sastry 07 February 2016, 23:26:59 UTC ruthenium services: Few more tweaks * Now that we know a little bit more about the pattern of memory usage with node v4, use 8 workers on the parsoid service, one per testreduce client. * visual diffing testreduce clients post JSON to the testreduce server. * visual diffing requires the pngs/ directory to save images to. Change-Id: Ifb04c2b5b8acb2ffb835cf2af52472b7a786566c 07 February 2016, 23:37:27 UTC
6a2e3d9 Tim Landscheidt 07 February 2016, 20:32:57 UTC Tools: Fix "invalid byte sequence in UTF-8" in crontab backups Puppet is called with the option --show_diff by /usr/local/sbin/puppet-run. This process to show the applied difference only works when both old and new content is valid UTF-8 data, otherwise Puppet will fail with the error "invalid byte sequence in UTF-8". User-supplied crontabs can contain invalid UTF-8 sequences; this change thus marks them as their changes not be detailed during a Puppet run. Change-Id: Ide715b1c057ad18403751317c4746920ae983fc6 07 February 2016, 21:44:05 UTC
ab87987 Merlijn van Deen 07 February 2016, 17:39:12 UTC toollabs: sync jessie packages with precise/trusty Change-Id: I08d245518f5379ead8d3462ab155d1185a71d7e2 07 February 2016, 21:11:50 UTC
f5bf60c Merlijn van Deen 07 February 2016, 17:36:01 UTC toollabs: libbytes-random-secure-perl does not exist on precise Bug: T126168 Change-Id: I928062f9f0f52cf7b827331a5de7f3c49b2d4580 07 February 2016, 21:10:46 UTC
8ce3497 andrewbogott 07 February 2016, 03:16:59 UTC Fix ldap_user_name_attribute in keystone config Previously we were telling keystone that username and uid were the same. The 'name' should actually be the 'sn' which is the on-wiki name. Change-Id: I7e4d5d84998735957890cece9c73f1d3ac2d3acc 07 February 2016, 03:18:35 UTC
5039951 andrewbogott 06 February 2016, 23:41:15 UTC Specify wgOpenStackManagerProjectId in WikitechPrivateSettings.php Also adjust both wgOpenStackManagerProjectId and wgOpenStackManagerProjectName so they can be specified in heira. These settings are needed for OpenStackManager to make keystone queries. Bug: T115029 Change-Id: Ic8e9456be2170885fa6d7fa435b691a2ef47a981 07 February 2016, 02:54:30 UTC
845ffd1 YuviPanda 05 February 2016, 23:48:29 UTC labstore: Add error checking + logging info to create-dbusers Change-Id: I8a62c986490d2e22128e99a7acf640b333d8a84a 05 February 2016, 23:49:04 UTC
e10801b Ori Livneh 05 February 2016, 22:44:42 UTC Add 'Backend-Timing' response header on all Apaches Configure all Apaches to add a "Backend-Timing" response header with the format 'D=XXX t=XXX', where 't' is when the request was received and 'D' is how long it took to serve the request, both measured in microseconds. The microsecond resolution is impractical, but this is what Apache provides. Tested with Apache 2.2 and 2.4. Change-Id: Ib82b93d167be071cd3ae6ee577c1a235c7986719 05 February 2016, 23:32:20 UTC
cc7c597 Jcrespo 05 February 2016, 23:01:37 UTC Revert "Revert "Revert "Revert "Install Jessie on db1018"""" This reverts commit a376eed3c16eb2f4224ab9eef4669ef763266103. Change-Id: I371a9a68c366bec2633a60b211a36b4f3af9d1f1 05 February 2016, 23:02:14 UTC
b90642d Alexandros Kosiaris 29 May 2015, 14:07:32 UTC Remove absented uuid-generator script It's has been absented for a long enough time already. Remove the resource Change-Id: I114b49ab630db45bd8b57b631ca81a5ae180ed12 05 February 2016, 22:13:40 UTC
6e221bc dzahn 05 February 2016, 22:02:42 UTC admin: add parsoid-test-admins to ruthenium Apply the new admin group parsoid-test-admins on host ruthenium via hiera. Bug:T124701 Change-Id: I0b686a5390392ce5830b4f4485bdcd06f0ad9b5f 05 February 2016, 22:05:27 UTC
a02da75 Brandon Black 05 February 2016, 21:58:46 UTC disable ipsec for cp3032, dead Change-Id: Id3140fcb75d6711d95d405f5e35b78e41062c098 05 February 2016, 21:58:46 UTC
c097257 RobH 27 January 2016, 00:20:30 UTC creation of parsoid-test-admins group This creates the parsoid regresstion testing admin group. While this adds the usergroup, it doesn't add any users to said group. Any users must be reviewed in the operations team meeting. As this doesn't add any users to these rights quite yet, it can sit unmerged until the ops meeting appends users in on 2016-02-01. This should be reviewed by members of the operations team, as it adds sudo rights to the cluster. Bug:T124701 Change-Id: I8a9d250349d88fa19388e4f5a1083a5951496ae1 05 February 2016, 21:53:40 UTC
25d9fc4 Merlijn van Deen 05 February 2016, 21:36:19 UTC scap: remove unnecessary unicode em-dash Change-Id: Id9b24db42969a13f115189547469cb58a574d277 05 February 2016, 21:36:19 UTC
b77ceae dzahn 05 February 2016, 20:03:06 UTC cygnus/technetium: install nmap, include standard cygnus and technetium are VMs for pentesting frack. Sets up a stub class to install the requested nmap. Includes standard on these instances. Bug:T126012 Bug:T118763 Change-Id: Iaff5d643680ecc04ecb0fed03d0ca3e81fa9283e 05 February 2016, 20:05:59 UTC
3c0c3c9 pt1979 05 February 2016, 18:13:41 UTC auto-install: add sarin to netboot Bug:T125752 Change-Id: I394d0be79fef68d9be8f62c04716f4e50bed56ba 05 February 2016, 19:56:20 UTC
8327aa1 dzahn 05 February 2016, 17:50:44 UTC dhcp: fix syntax error, missing bracket Change-Id: I08936dbca1038a051bdc10892d49ef537154366f 05 February 2016, 18:03:35 UTC
a945f06 Tyler Cipriani 05 February 2016, 17:55:12 UTC Reload keyholder-agent on keyholder-auth change Currently, if puppet adds a new key and a corresponding new keyholder-auth.d yaml file to go with it this requires an root to run: keyholder restart keyholder arm This change should ensure that only a `keyholder arm` is needed since it will reload the keyholder permissions if a permission file has changed. Bug: T125992 Change-Id: Ic9803696768fb8fa83d8e4ebad85119194c71c14 05 February 2016, 17:59:56 UTC
e654567 pt1979 04 February 2016, 18:00:20 UTC Add sarin to DHCP Bug:T125752 Change-Id: Ie4b0310770e0339a22e3141ddd01139c2d9706bb 05 February 2016, 17:41:36 UTC
abdfe2d Jan Zerebecki 05 February 2016, 15:53:14 UTC icinga contactgroups: add jzerebecki and irc-wikidata to wdqs Change-Id: I1ee72f782e1c4d84193eae2ac50928912f118e56 05 February 2016, 17:17:05 UTC
426256c addshore 05 February 2016, 17:05:40 UTC WDQS_Lag monitoring increase timespan Due to general lag in this data appearing in graphite and also due to the fact that the data is filled one host at a time it seems a 10 min scope does not cover enough data points. It would seem the requirements are 50% of data points be present. A 30 minuite scope should allow for that. Change-Id: I82ecd485c92a2b29f9928b9c018b4d3cab2c25e1 05 February 2016, 17:05:40 UTC
6ef010d addshore 05 February 2016, 13:16:21 UTC Add WDQS_Lag monitoring Change-Id: Idca39b59634e0e23cf2d1fe363a6ce1894ac6965 05 February 2016, 16:46:35 UTC
384eca3 Andrew Otto 05 February 2016, 16:17:59 UTC Reenable kafka1012 broker Bug: T125199 Change-Id: I74f896d5509aa9c7ed2c9e700161a8768b8e4585 05 February 2016, 16:17:59 UTC
f4baf3e Jaime Crespo 01 February 2016, 15:33:03 UTC Testing db jessie installer problems on db2030 This will clarify if it is host-dependent or installer-updates- dependent. Bug: T125256 Change-Id: Ib8a883e7a15ee9a3846e372e086a1319276451df 05 February 2016, 15:37:16 UTC
65dfde1 Brandon Black 05 February 2016, 15:29:37 UTC temporarily de-sensitize ipsec icinga checks Change-Id: I2b66bd0eb515622b7c870cf9d853dbce999c19a9 05 February 2016, 15:29:49 UTC
4ee6d81 Giuseppe Lavagetto 05 February 2016, 15:23:49 UTC scap: correct clone directives Also correct the erroneous documentation of git::clone Change-Id: Ie6705ea3b4f0c36675bbb67acb916c54b7ce7623 05 February 2016, 15:25:22 UTC
eb4fefc Giuseppe Lavagetto 05 February 2016, 14:59:30 UTC scap: clone repositories for l10update Change-Id: Ifabea760d91691dacd1f6292e58ec5a7bb4c8b09 05 February 2016, 15:04:56 UTC
ab9ca6e Marko Obrovac 05 February 2016, 02:30:41 UTC Revert "Revert "RESTBase: enable metrics batching"" This reverts commit a842841bbf6b02b22c50bda70a3a8d3e55827607. The bug we initially witnessed in 79b5c511b3920e9174302aabf081b4d347a94e36 has since been fixed, so now it's safe(r) to enable metrics batching for RESTBase. Bug: T121231 Change-Id: I3d505061b3058837c183d8ce47c4a3c06980a63d 05 February 2016, 14:54:30 UTC
35e8ca0 Ariel T. Glenn 05 February 2016, 13:25:58 UTC dumps: workaround for pgrep check if script already running pgrep test works differently depending on how script is run, with or without explicit /bin/bash ahead of the script name. Now we use -f (full) and all instances are returned including the current one we just started up, but we then exclude processes of the same pgroup as the current one. Change-Id: I3cfefa816113f93ded8e7b2311865564b076409c 05 February 2016, 13:25:58 UTC
bcf11b0 elukey 05 February 2016, 11:58:09 UTC Re-Add mw1228.eqiad.wmnet to the api_appserver pool Bug: T122005 Change-Id: I0ac208b2616a3804b43236a426fbc6c41210d85f 05 February 2016, 11:58:09 UTC
2c452ee elukey 05 February 2016, 10:27:15 UTC Add mw1228.eqiad back to the DSH list. Bug:T122005 Change-Id: I3da9b7a2b26c494d38ecdffa2731eb8f4c0b34c0 Signed-off-by: elukey <ltoscano@wikimedia.org> 05 February 2016, 10:52:17 UTC
23ac866 dzahn 05 February 2016, 04:35:27 UTC grafana::dashboard, $ensure first, present by default Change-Id: Ia85f566c2419c69575f938c04771d15a71f73bec 05 February 2016, 04:37:12 UTC
ecc9159 Tim Starling 02 February 2016, 23:18:06 UTC For all apache access logs, use the WMF cache log format For all apache access logs, use a tab-separated log format similar to https://wikitech.wikimedia.org/wiki/Cache_log_format except that the first two fields (hostname and sequence number) are omitted. Two fields are added to the end: * %u, the authenticated username, for the benefit of lists.wikimedia.org and anything else using HTTP authentication. * %{c}a, the IP address resolved by mod_remoteip, which is available in Apache 2.4+. This requires extra configuration to provide useful information, and note that the proxy path (%a and X-Forwarded-For) is still useful information for incident analysis. Apache's default log formats are annoying -- they don't even contain a full URL for pasting into a web browser. vhost_combined might almost allow you to put one together, except that it uses %v instead of the actual host header, which might not have any relationship to the matched ServerAlias. Execution time is essential information for performance debugging and incident response, and you often don't know you need it until it is too late. The date format is ridiculous, and putting double quotes or square brackets around space-separated fields is remarkably unhelpful if you need to analyse the logs using simple tools like awk or regexes. I am replacing it with the "wmf" format everywhere because I never want to see a stupid "combined" log again. I haven't reviewed all these modules to make sure they are using the apache module (which creates the log format in default.conf via apache::init), I just checked a few randomly. I did copy the LogFormat into the appserver apache2.conf, but I don't think it is used. If any server is missing the LogFormat directive, the effect is to treat the "wmf" format as a literal string, i.e. all log lines will just be "wmf". Change-Id: I0435ef658f0e8bef1507cd57f5c316115531ce93 05 February 2016, 01:50:14 UTC
7594bcd Brandon Black 05 February 2016, 00:25:40 UTC check_systemd_unit_state: fix silly syntax bug Change-Id: I7ee25e07f54af728f1cd2b27422c60aa51431e75 05 February 2016, 00:25:40 UTC
2efa193 Brandon Black 05 February 2016, 00:03:12 UTC traffic-pool: s/true/yes/ for consistency Change-Id: Icc35edf0ca02f608bc2194f0d4677d3dc1157d00 05 February 2016, 00:18:03 UTC
575ab51 Brandon Black 04 February 2016, 23:50:32 UTC check_systemd_unit_state: support oneshot/RemainAfterExit Change-Id: Ic8420a3c97551bd8361bede8a561620697bdf565 05 February 2016, 00:18:01 UTC
8a78abf Brandon Black 05 February 2016, 00:14:25 UTC check_systemd_unit_state: pep8 cleanup Change-Id: I04bdb70f1fe793740dbc9a008cc58a064004c452 05 February 2016, 00:15:41 UTC
5ecfd8c Marko Obrovac 04 February 2016, 23:49:19 UTC RESTBase: Change to v0.10.x style config As of v1.1.1, service-runner supports setting the application's base path, which ensures that relative paths in the config and its included files are correctly interpreted. Change-Id: I0c0f99221cf856014d334e16c7c76f57007a23e6 04 February 2016, 23:52:09 UTC
a15824b Marko Obrovac 04 February 2016, 23:06:28 UTC RESTBase: Labs: Use relative module path Without a prefixing dot, service-runner tries to load an npm module with the given name, instead of looking for the file in the base path. Change-Id: I8f68441dc60bddea4ded7950419f01f3c7f9a8ae 04 February 2016, 23:48:48 UTC
744967b Brandon Black 04 February 2016, 17:20:54 UTC add traffic-pool systemd service unit Change-Id: I4a4ace604ba88c039cc367c09ea9eaeb020686fd 04 February 2016, 23:30:00 UTC
69a1b88 jdlrobson 04 February 2016, 00:04:13 UTC Make mobile-beta an available platform The mobile beta site varies drastically from the stable site so let's track navigation timing results in a separate bucket Change-Id: Idb452b6081d392261a40054f159c4995d887d58a 04 February 2016, 23:15:29 UTC
d64cfcd Subramanya Sastry 04 February 2016, 22:52:33 UTC parsoid-rt-client: parsoidConfig reqd even if using a global parsoid svc Change-Id: I3adf7f3c7937e93072833804ff5276dbc64be4c1 04 February 2016, 23:04:02 UTC
7fbc293 Ori Livneh 04 February 2016, 22:38:36 UTC Write redis's pid file to /var/lib/redis by default. - We configure redis to daemonize. This is not great, but tricky to change, because redis's configuration and its Upstart job configuration have to be changed in a particular order to avoid Upstart losing track of redis's pid. - We configure redis to write its pid file in /var/run/redis/. - Upstart setuid / setgid redis, so it can't create /var/run/redis. - We could drop the setuid / setgid directives from the job file and use start-stop-daemon to run redis and drop privileges instead, but that means changing "expect fork" to "expect daemon", which is tricky -- see the first point above. - We could have a separate Upstart job that is configured to run before any redis instance jobs and that creates /var/run/redis. But this is ugly. - We can make redis write its pid file to /var/lib/redis, which persists across reboots. This is the approach I went with. If a stale pidfile is present, redis will clobber it, which is the desired behavior. Change-Id: I5b3c0964fcbb2343dca99494915bcf88e630e64b 04 February 2016, 22:53:28 UTC
51fc169 dzahn 28 January 2016, 03:27:40 UTC grafana: fix top-scope var without namespace Add $ensure to resource parameters. ./modules/grafana/manifests/dashboard.pp WARNING: top-scope variable being used without an explicit namespace on line 42 Change-Id: I0eb977bfa4ef1110128b59edefe5545ab14bd2b6 04 February 2016, 22:41:12 UTC
17b29f4 Marko Obrovac 04 February 2016, 22:34:43 UTC RESTBase: Labs: Use RB 0.10.x config style The new version of service-runner allows us to specify the application's base path, which ensures that relative-path strings are interpreted correctly by the application and do not rely on the CWD which changes from one environment to the other. Change-Id: Ia934e64c2b8e450b205047a52fdcae6a272e8f88 04 February 2016, 22:38:30 UTC
68f4072 Subramanya Sastry 03 February 2016, 23:58:50 UTC ruthenium: Some more tweaks to parsoid + visualdiffing services * Added packages used to build uprightdiff to the visualdiff manifest. * Added missing env vars to the parsoid-vd-client systemd script. * Added the nodejs-legacy package to the parsoid-tesing role since the phantomjs binary installed by the phantom node module requires 'node' (whereas debian installs nodejs by default). Change-Id: Ifeea9cc1fa68dffb130815a6ea9283aee87aa1c9 04 February 2016, 22:27:44 UTC
7e3e96f Subramanya Sastry 03 February 2016, 21:31:00 UTC parsoid-rt-client: Have testreduce clients use global parsoid service * Instead of spinning up a private 1-worker parsoid service per testreduce client, have the clients talk to a global 4-worker parsoid service. This eliminates the testreduce failures we have encountered with node 4.2 that doesn't restart failed workers on the same port as it was running on. * Also using only 4 workers for the 8 testreduce clients to stress the parsoid service a bit more than usual to test memory behavior on node 4.2 * Removed VCAP_APP_PORT setting used by the long-dead appfog support. Replaced with PORT=8142 which is where we want this service running. Change-Id: I0febb9e26d10b0cd324e404e294053f760318b22 04 February 2016, 22:22:59 UTC
c62c102 Sébastien Santoro 02 February 2016, 22:48:24 UTC RESTBase and Labs DNS configuration for ady.wikipedia Bug: T125501 Change-Id: I4cda48113ab4d34b978026f04a86ebf54240c690 04 February 2016, 19:45:16 UTC
4936758 Ariel T. Glenn 04 February 2016, 18:25:33 UTC dumps: make dumps cron script maintainable, dryrun option plus clean up various test conditions that were broken lesson to be learned: test after every single tweak Change-Id: I8e924ad219a7aaf75fecc19f634b407be985eea2 04 February 2016, 18:49:11 UTC
407426c Brandon Black 04 February 2016, 17:50:09 UTC fix output typo in "depool" Change-Id: I7a2390be9b429675a1dcccdd4d34d1f824da8e77 04 February 2016, 18:31:22 UTC
1341508 aude 29 January 2016, 13:17:30 UTC Update WikidataBuildResources git source (github -> gerrit) Bug: T111173 Change-Id: Iff725deaa0d2d5c64d3a80ff6402e07a3ff71ee9 04 February 2016, 17:54:01 UTC
2270cb3 Guillaume Lederrey 04 February 2016, 14:02:44 UTC Adding monitoring of some ElasticSearch thread pools Thread pools suggested by @dcausse have been added to monitoring. We also monitor rejections for all thread pools. Modification seem trivial, but have not been tested locally. Please review carefully. Bug: T125782 Change-Id: I02e4c3fd16c2def3b20835f20ca79c0823936f12 04 February 2016, 17:37:52 UTC
5593d75 Eric Evans 25 January 2016, 19:50:45 UTC [production]: match restbase config to current Cassandra cluster Depends-On: If738e7401a60e47ca45b875ed3abefa40dcebe11 Bug: T123869 Change-Id: Ie3a0b06c8a2164f967718624e8c33a8ade7b7d01 04 February 2016, 15:14:17 UTC
73e9bd1 Eric Evans 02 February 2016, 17:41:57 UTC Change default consistency to localOne It has been suggested that reducing read consistency to ONE on the AQS cluster might help reduce some load until more disk IO can be provisioned. Other than setting consistency on a request-by-request basis, the only "knob" provided is `defaultConsistency`, which impacts both reads and writes. However, in the context of AQS, this should be OK as RESTBase is only used for reads anyway. Bug: T124947 Change-Id: I2147afe9a8145200c6217d6b5c32cc35902679c4 04 February 2016, 15:00:37 UTC
1ca2386 Filippo Giunchedi 03 February 2016, 15:32:03 UTC uwsgi: create /run/uwsgi, including at boot graphite1001's uwsgi failed to start after a reboot because of missing /run/uwsgi Change-Id: I8aa70e5f626cf5b3c237bd2418ed48ffcf1ca342 04 February 2016, 11:21:30 UTC
8ffe97f Alexandros Kosiaris 02 February 2016, 12:37:37 UTC etherpad: Be pedantic about defaultPadText Put the missing dot at the end of the last sentence Change-Id: I427319c6de7ebd4edb6968ec07664ada29c77bf9 04 February 2016, 11:19:31 UTC
c3c3562 Alexandros Kosiaris 03 February 2016, 12:20:30 UTC otrs: redirect iodine to the test database iodine is the old OTRS server. It hosts an old version of OTRS software and hence should not have access to the production database in order to avoid any kind of problem. Change-Id: Ide8abb20d43d7122445adcb4c1485691b46a37c8 04 February 2016, 11:19:10 UTC
f01cd33 Alexandros Kosiaris 04 February 2016, 11:14:44 UTC otrs: PostMaster.pl is deprecated, replace it with Console.pl otrs.PostMaster.pl is deprecated in 5.0.x so replace it with the one we get in the notice in the log files. Double quote it to avoid problems Change-Id: Id77961459aed30f10f970242b12ef850ccfd3075 04 February 2016, 11:14:44 UTC
64c7251 Jaime Crespo 03 February 2016, 09:36:10 UTC Add echo tables to the list of private tables Bug: T125591 Change-Id: Iaad18d570c4386bacfe5f5e9fc473eb6bb82a4c1 04 February 2016, 09:39:42 UTC
90edd7f Ariel T. Glenn 04 February 2016, 09:33:36 UTC dumps: skip labtestwiki for addschanged and pagetitles dumps can't be dumped from snapshot hosts Change-Id: I4bb49b4c28c21f6b00190b25d37693c77bc0b915 04 February 2016, 09:33:36 UTC
5e3f7ed YuviPanda 04 February 2016, 01:57:09 UTC tools: Replace php5-mysql with php5-mysqlnd Claims to be a fully compatible but faster replacement, so let's give it a shot. Provides functions unavailable to php5-mysql Bug: T125758 Change-Id: I9646ddf3667f74d90c3f079602378f4f1a1b5e55 04 February 2016, 01:58:16 UTC
dccc1a1 dzahn 28 January 2016, 03:08:13 UTC ldap: fix top-scope var without namespace ./modules/ldap/manifests/role/config.pp WARNING: top-scope variable being used without an explicit namespace on line 5 WARNING: top-scope variable being used without an explicit namespace on line 10 WARNING: top-scope variable being used without an explicit namespace on line 11 ./modules/ldap/manifests/role/client.pp WARNING: top-scope variable being used without an explicit namespace on line 28 Change-Id: I8fbb141a27ab0873d6cfcf87f6fb93436e016d71 04 February 2016, 00:45:29 UTC
6fa9c94 dzahn 28 January 2016, 03:06:16 UTC mediawiki/jobrunner: fix top-scope var without namespace ./modules/mediawiki/manifests/jobrunner.pp WARNING: top-scope variable being used without an explicit namespace on line 27 Change-Id: I8b82c2218b0103acdc7358d126864f5159868df5 03 February 2016, 23:33:53 UTC
1f5cb06 dzahn 28 January 2016, 03:26:14 UTC varnish: fix top-scope var without namespace ./modules/varnish/manifests/common/directors.pp WARNING: top-scope variable being used without an explicit namespace on line 37 Change-Id: Ibfc3b406665c1fd82652f20b236ff6a80d06f099 03 February 2016, 23:29:29 UTC
44ba34e dzahn 28 January 2016, 03:03:02 UTC dataset: fix top-scope var without namespace ./modules/dataset/manifests/nfs.pp WARNING: top-scope variable being used without an explicit namespace on line 25 ./modules/dataset/manifests/cron/rsync/peers.pp WARNING: top-scope variable being used without an explicit namespace on line 12 The second one looks like a mistake, you meant $enabled here, not $absent, right? Change-Id: I43b22198c16602a14bcbb553821aff6474b53f2f 03 February 2016, 23:12:38 UTC
47b50c7 dzahn 28 January 2016, 03:17:51 UTC osm: fix top-scope var without namespace, rm cruft ./modules/osm/manifests/planet_import.pp WARNING: top-scope variable being used without an explicit namespace on line 25 removing the entire line that is not used anymore Change-Id: Ib941673a231f1de31a2f39346c3a15c6287f2f90 03 February 2016, 23:07:59 UTC
3b57429 dzahn 28 January 2016, 03:45:03 UTC ipsec: fix top-scope var without namespace ./manifests/role/ipsec.pp WARNING: top-scope variable being used without an explicit namespace on line 36 Change-Id: I2460bd3b028e7669b879a2331ae3e5d8a78ebac6 03 February 2016, 22:59:59 UTC
de70fa7 YuviPanda 30 January 2016, 00:45:40 UTC toollabs: Do not hardcode Host header Allows for more flexibility this way Change-Id: Ie1a2e18b16051b454d8eef730f5a7966ada8f98e 03 February 2016, 19:32:16 UTC
71e6d3c andrewbogott 03 February 2016, 04:16:07 UTC Fixes to the glance image backup cron Specified more details about the cron, in hopes of getting the ensure => absent to actually do something. Change-Id: Ie8b09e844c71417e30b2d0dc37bcda60e1d16e7e 03 February 2016, 19:21:24 UTC
fa78c1e Chad 02 February 2016, 23:20:30 UTC Revert "tin: disable l10nupdate until we figure out if it works with HHVM" This reverts commit 444e8aa7b46014dd3097c83e3c7e279aa67703a6. Change-Id: Idb29ea99b0ed5ed2b56436279511d5e9d6e669fa 03 February 2016, 18:53:04 UTC
9d7f939 dzahn 28 January 2016, 03:29:11 UTC openstack: fix typo, "spandby-server" for glance if $spandby_server ... <.. that's a standby server Change-Id: I27e81631f44d3837e1c4d133b5377419d82a73a5 03 February 2016, 18:25:19 UTC
d808824 Alexandros Kosiaris 02 February 2016, 15:02:37 UTC otrs: Route OTRS email to mendelevium Route OTRS inbound email to mendelevium instead of iodine Bug: T74109 Change-Id: I2ce04581d9dae0bffb13983a1fc74009d2c3c7bc 03 February 2016, 17:52:42 UTC
22f96b5 Faidon Liambotis 01 October 2015, 03:10:29 UTC otrs: disable SessionCheckRemoteIP Disable remote IP checks for sessions. We use Cookies over TLS so we don't need this extra check. Having it enabled is actually hurtful, as it logs out users that e.g. use IPv6 with privacy extensions enabled. This is insecure before OTRS 4.0, because the session token is getting leaked on the address bar after login and users may inadvertently share their session by sharing the URL with someone else (upstream OTRS bug 8008). Bug: T87217 Change-Id: I3286690a74df4ec80c9e6263326cb14578eb2a53 03 February 2016, 17:44:36 UTC
197df53 Alexandros Kosiaris 02 February 2016, 12:31:07 UTC misc-web: Route ticket.wikimedia.org to mendelevium Since mendelevium will be powering up ticket.wikimedia.org, have varnish pass connections to it uncached to mendelevium Bug: 74109 Change-Id: I94718cd38b5002e5b9e6a540febedd3610f4fc56 03 February 2016, 17:37:33 UTC
4d36c75 Alexandros Kosiaris 02 February 2016, 12:25:38 UTC otrs: OTRS search slow, increase between_bytes_timeout OTRS search can be seriously slow, taking well over 40s to return results. This has been reported over the years in T17210 which refers to http://bugs.otrs.org/show_bug.cgi?id=3724 which has been closed an INVALID. It is quite possible it will never be fixed, so increase the varnish timeout to 60s so at least it does not terminate the connection fully breaking the behavior, causing a regression. Bug: 74109 Change-Id: I1b7b3180dd760984fa7420b068c0f4d254cfd3f1 03 February 2016, 17:36:46 UTC
dc35a85 elukey 03 February 2016, 17:21:26 UTC Add Subbu to a temporary parsoid-root group. Bug: T125166 Change-Id: I188d729b55020227c059ded1a0b325563ac12afd 03 February 2016, 17:29:54 UTC
a842841 Marko Obrovac 03 February 2016, 17:27:51 UTC Revert "RESTBase: enable metrics batching" This reverts commit 79b5c511b3920e9174302aabf081b4d347a94e36. Change-Id: I208df45e2537d2f19a7532b0f37f897bf6f9ce53 03 February 2016, 17:28:10 UTC
3b843c5 Alexandros Kosiaris 03 February 2016, 17:20:50 UTC otrs: Collapse all VirtualHosts into one Remove the now unneeded virtualhosts and keep just the main one that is HTTP only since the caching proxy clusters take care of the rest Change-Id: I4b0b200de99451bdcc2f6065bcb109771e2b61f7 03 February 2016, 17:22:42 UTC
79b5c51 Marko Obrovac 02 February 2016, 16:37:02 UTC RESTBase: enable metrics batching RESTBase sends a lot of metrics to graphite. In order to minimise the amount of requests done (each of which is rather small in payload), batch consecutive events together. Change-Id: I9fe06567cfbc2e5ac17de9c6e644a3fd8246d265 03 February 2016, 17:17:32 UTC
033c6b5 Marko Obrovac 29 January 2016, 23:30:09 UTC MobileApps: Change RESTBase URI The Mobile Content Service has changed the way they include the RESTBase URI to a more flexible scheme that allows them to use both RESTBase's internal URI, as well as the generic https://{domain}/api/rest_v1 external URI. This patch conforms to the new standard. Note: to be deployed in conjunction with I6d703b627fc22a0bfe7c27b32efa54ef0e10b5cf . Change-Id: I4523d97598ef10d2075bd5ec9f60ba8bf5acfef4 Depends-On: I6d703b627fc22a0bfe7c27b32efa54ef0e10b5cf Bug: T125252 03 February 2016, 17:04:57 UTC
cdd563b Brandon Black 29 January 2016, 10:35:55 UTC eqiad: remove last cache_mobile frontend Needs to wait on analytics readiness for T122651 Bug: T122651 Bug: T109286 Change-Id: Ic41d810fa0007d310feb723d255a020367556c32 03 February 2016, 16:44:33 UTC
73fb868 Andrew Otto 03 February 2016, 16:15:59 UTC Remove old oozie logs in /var/log/oozie after they are rotated Change-Id: I20af0383cc0bfa39b21daf6424073d526a6ec27c 03 February 2016, 16:15:59 UTC
680c3ad Andrew Otto 03 February 2016, 16:11:35 UTC Remove webrequest_mobile from list of webrequest topics that camus imports Bug: T122651 Change-Id: Iad6c6ed738ee3538c0a044ffbf956a8fa04d64e1 03 February 2016, 16:11:52 UTC
ef4b6a3 Jaime Crespo 03 February 2016, 15:51:10 UTC Reconfigure db1060 (depooled) Change-Id: I6c902fdd4fc71a9545a0e9a10d7e12ea1aef3936 References: T120122 T125215 03 February 2016, 15:54:32 UTC
b054fed Alexandros Kosiaris 03 February 2016, 15:34:58 UTC conftool: Remove tilerator from conftool data tilerator is NOT an LVS service. The tilerator entries in conftool are just leftover garbage. Removing them Change-Id: I46a648d870d25d6185a40c896a53b75d1d508cf7 03 February 2016, 15:34:58 UTC
1512d9e Jaime Crespo 03 February 2016, 13:44:13 UTC Reconfigure db1054 (depooled) Change-Id: Ie546c53f74e1ed38331d1cea330414fa38302c4d References: T120122 T125215 03 February 2016, 13:44:13 UTC
f0ec128 Jaime Crespo 03 February 2016, 11:47:32 UTC Reconfigure db1067 (depooled) Change-Id: I21a402e14a1da53137864d8b68baca586107a764 References: T120122 T125215 03 February 2016, 11:47:32 UTC
22c129b Ariel T. Glenn 03 February 2016, 11:35:13 UTC fix variable name in dump cron job for hugewikis Change-Id: I10e286fcf9feb67b5a0b1effa91842ebc5f86aa0 03 February 2016, 11:35:13 UTC
b0f9526 dzahn 28 January 2016, 03:33:56 UTC salt: fix top-scope var without namespace ./modules/role/manifests/salt/minions.pp WARNING: t top-scope variable being used without an explicit namespace on line 30 Change-Id: Ia00738e449c80e9680e8f7e6ca24d2987266d7bc 03 February 2016, 10:21:46 UTC
b9a56e7 lokal-profil 04 January 2016, 22:43:49 UTC Support bzip2 compression format Depends on: Ieb517b5ad677abaa541274d69155855b95787f12 to /operations/dumps/dcat Bug: T118397 Change-Id: I8eb0077d60afbcc87663e3bdd41c5f04d959d330 03 February 2016, 10:14:20 UTC
5729627 Ariel T. Glenn 03 February 2016, 09:53:28 UTC dumps: add labtestwiki to list of wikis we don't dump for permissions reasons Change-Id: Ie1cfcc215a1c51a68b945051eefc7698423b5788 03 February 2016, 09:54:30 UTC
5f16e57 Alexandros Kosiaris 03 February 2016, 09:01:02 UTC Update mendelevium's OTRS config Set the correct values of m2-master and database Bug: T74109 Change-Id: Iefc1ef5db8701efe8484141122b7dfc8b87bf33d 03 February 2016, 09:04:19 UTC
341eee8 Alexandros Kosiaris 03 February 2016, 08:50:16 UTC Remove mendelevium's OTRS config override Bug: T74109 Change-Id: Idbbf8a436897ceb27e0fd4ebe37a1455ceadd8d4 03 February 2016, 08:55:40 UTC
fbcdf47 Giuseppe Lavagetto 03 February 2016, 08:41:02 UTC scap::master: fix commit hook reference Change-Id: I5f89b635e9d208d49492b110cd104185e0cd29bb 03 February 2016, 08:41:02 UTC
5964d83 Chad Horohoe 02 February 2016, 18:24:18 UTC deploy master: recurse submodules on clone Otherwise we end up weirdly half-setup Change-Id: I340f05125aa040eb8c027bf75ea78560bc1ca7b1 03 February 2016, 08:33:06 UTC
a68f04d Giuseppe Lavagetto 02 February 2016, 15:20:18 UTC scap::master: add commit-msg hook It was unpuppetized, it should be. Change-Id: I3976a1f9df3e93ef8df279ab29dfbad162bd9a27 03 February 2016, 08:31:08 UTC
ab183dc YuviPanda 03 February 2016, 07:34:18 UTC tools: Add kube-system service account to abac Allows it to read everything. Useful for setting up DNS Change-Id: Id69ae1f4dff81eaa26a93b22fe93f2a28604a213 03 February 2016, 07:37:08 UTC
da99c7c Tim Landscheidt 27 January 2016, 02:39:44 UTC Tools: Deploy root web automatically again Change-Id: Ia464364afd07a1ca8d2cb3c0c4aed7fe35bac91b 03 February 2016, 03:09:13 UTC
8651d48 Tim Landscheidt 30 January 2016, 04:50:43 UTC shinken: Indent and align generated configuration Currently, the configuration is generated à la: | define host { | use generic-host | host_name tools-mail | address 10.68.16.27 | hostgroups tools,role::labs::tools::mailrelay | contact_groups tools | notes tools | } This change indents the directives and aligns their values by their longest key: | define host { | use generic-host | host_name tools-mail | address 10.68.16.27 | hostgroups tools,role::labs::tools::mailrelay | contact_groups tools | notes tools | } I was unable to implement the alignment with jinja2, so I resorted to standard Python. Change-Id: Ieef7cc54c1d659e7a88794c9240993944a838af4 03 February 2016, 03:07:07 UTC
back to top