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
7c1ba8a Andrew Otto 08 July 2016, 15:32:45 UTC Update eventlogging kafka consumer args to match with python-confluent-kafka consumer Bug: T133779 Change-Id: Ia49006ea2430313ca89a3099c5348a1e4372189a 08 July 2016, 18:46:10 UTC
29929a9 YuviPanda 08 July 2016, 14:42:26 UTC dynamicproxy: Remove redundant redundantproxy.conf file Teehee Change-Id: I6e60558c01d032f1cbb5c14db32251d3bbc7a244 08 July 2016, 14:43:10 UTC
fc6f443 elukey 08 July 2016, 13:14:56 UTC Remove Websocket upgrade request from Varnishkafka webrequest config Websockets upgrade usually lead to long lived requests that trigger VSL API timeouts. Varnishkafka does not have a nice support for these use cases, moreover we haven't decided yet if weberequest logs will need to take them into account or not. At the moment these requests get logged incorrectly and with partial data (due to the VSL timeout) so it makes sense to filter them out to remove noise from Analytics data. I have been testing this setting for days on cp300[89].esams and it dramatically reduce the amount of timeouts registered. Bug: T136314 Change-Id: I77ec0e5b3b9a4b988e3056d640c3d09212dcb97b 08 July 2016, 14:20:02 UTC
5270f9d andrewbogott 07 July 2016, 01:28:36 UTC Increase conntrack limits for nova compute nodes. Bug: T139598 Change-Id: Id84c67064ec78f9d2c197ada2b33c44895b49d91 08 July 2016, 13:54:26 UTC
bfc761e Giuseppe Lavagetto 08 July 2016, 13:12:06 UTC puppetmaster::passenger: use sslcert::dhparam on jessie Since we're using a DH parameters file on jessie, we need to add it or apache wouldn't be able to start. Bug: T98173 Change-Id: Ib4674f8a7f85f6049c4ce6af4a8e20f0a6bea90d 08 July 2016, 13:18:59 UTC
6ca6bca Yuvipanda 08 July 2016, 13:05:16 UTC Revert "tools: Move static site to use http2 as well" Bug: T139743 Bug: T134383 This reverts commit 1c91c78af41cbec34121a735788f4024514de788. Change-Id: I267dee86d4c1276c1308728ed95eb9fe72a679ec 08 July 2016, 13:05:16 UTC
1c91c78 YuviPanda 08 July 2016, 12:34:17 UTC tools: Move static site to use http2 as well spdy is deprecated in newer nginx versions Bug: T134383 Change-Id: I2f441319d4d7e29bb3264b1e20aced5c14dd02e9 08 July 2016, 12:36:10 UTC
c7bc8a2 YuviPanda 08 July 2016, 12:01:47 UTC dynamicproxy: s/spdy/http2/ for domainproxy as well spdy is deprecated in new nginx versions Bug: T134383 Change-Id: Ia4425b4ec4a52d6ffe53d416ae23befdff418118 08 July 2016, 12:11:27 UTC
6d87d1b YuviPanda 07 July 2016, 17:56:23 UTC dynamicproxy: Use http2 rather than spdy spdy is deprecated. Bug: T134383 Change-Id: Ida1e1679bb4ed687b52482f3ac5fff9ef29040cc 08 July 2016, 12:00:50 UTC
eed836e YuviPanda 07 July 2016, 17:45:42 UTC dynamicproxy: do not override nginx.conf No reason to, and it causes issues elsewhere when you upgrade nginx packages and they have breaking changes in their nginx.conf. Bug: T134383 Change-Id: I1de30ed6e8b9f6543753f3bcb582ce724548f1fb 08 July 2016, 12:00:28 UTC
9d558eb Giuseppe Lavagetto 08 July 2016, 09:59:45 UTC puppetmaster: switch rhodium to install jessie Bug: T98173 Change-Id: Ifd772bd7bbe0a91c962ec4fe834805c65118e12f 08 July 2016, 10:07:36 UTC
2534f26 Moritz Muehlenhoff 04 July 2016, 07:59:07 UTC Install fonts-sil-lateef on scalers Bug: T138136 Change-Id: Ifa6c3155dcd0ce8d06a439226d828596cbe115dc 08 July 2016, 10:01:09 UTC
07d8690 Giuseppe Lavagetto 08 July 2016, 07:03:17 UTC nutcracker: lower verbosity on the maintenance hosts Change-Id: Ic71226fb3f8b50a3fec14e8716eb943d17291b55 08 July 2016, 07:03:17 UTC
29188f7 Antoine Musso 31 May 2016, 12:59:28 UTC zuul: enhance logging Newer Zuul version log SMTP/Gerrit interactions under the 'connection' logging bucket. That is similiar to the 'gerrit' one, log it at INFO level. Add debug logging for paramiko in Zuul to help catch issues with Zuul/Gerrit ssh connection. Send it to zuul.log and zuul-error.log as well in case it logs interesting problems. Stop Gearman debug, WARNING+ is good enough. Change-Id: I5db3f755c6be6ca67b697991d220205fa757c556 07 July 2016, 23:31:57 UTC
572a7fa Antoine Musso 24 June 2016, 11:20:20 UTC contint: Android SDK deps on all slaves The Android SDK is installed automatically by the Jenkins Android Emulator plugin. It depends on system packages to be installed, which were only made available on Trusty slaves. As we are migrating the jobs to Jessie nodes: * move Android dependencies to contint::packages::androidsdk * Get it installed on both Jessie and Trusty (excluding Precise) Bug: T138506 Change-Id: I402ca9e67e03d545ecca2de8db9094210c08e991 07 July 2016, 23:21:12 UTC
ff7a040 Merlijn van Deen 14 February 2016, 21:12:13 UTC toollabs: install inkscape on exec nodes precise: inkscape (0.48.3.1) trusty: inkscape (0.48.4) jessie: inkscape (0.48.5) Bug: T126933 Change-Id: I2b2d460c7c37f1ab13e379d17dfe00f723f4d3ca 07 July 2016, 22:50:35 UTC
39d5aea Moritz Muehlenhoff 10 June 2016, 11:47:28 UTC Enable base::firewall for labtestmetal2001 Enable base::firewall from the start, additional ports can be opened up as setup progresses. Change-Id: I7d3633dd30059803ae5cbaed0ad2177999a9b628 07 July 2016, 22:35:39 UTC
be4c7b5 Chad Horohoe 01 July 2016, 18:40:29 UTC Gerrit: move nasty ssh key to crons class, only user Change-Id: I40de2bfa1f42c7bcbb4198003bcb172e32bbc078 07 July 2016, 21:50:15 UTC
70fd51d Eric Evans 07 July 2016, 21:04:30 UTC Enable instance 1009-c Bug: T139362 Change-Id: I94f0ee1b3ccaa1ca8caee43c3ace3a6365ec9c41 07 July 2016, 21:08:04 UTC
1b2749d Chad Horohoe 07 July 2016, 18:31:04 UTC Gerrit: Puppetize the known_hosts file for replication Change-Id: I72b35eac63597036982c9a5bbd7b3b748d1d4d81 07 July 2016, 20:54:58 UTC
daeacfe Chad Horohoe 07 July 2016, 17:46:39 UTC Phabricator: Minor nit, move things that aren't templates out of templates Change-Id: I32033101af2bd4521fceca4dd22aa66785082e0e 07 July 2016, 20:34:14 UTC
d3308b5 Chad Horohoe 07 July 2016, 20:31:27 UTC Gerrit: Block a really bad person Change-Id: If5149f364f772f06b2679ae083085137c54cd6aa 07 July 2016, 20:31:51 UTC
92682f1 Eric Evans 07 July 2016, 18:32:11 UTC Upgrade codfw nodes to Cassandra 2.2 Bug: T126629 Change-Id: I960498f2bee9d8c4ca9e80a024dbacdc56dfa04f 07 July 2016, 18:32:11 UTC
8ab1661 Moritz Muehlenhoff 07 July 2016, 15:10:21 UTC xenon: Use DOMAIN_NETWORKS Only accepts traffic from hosts in production. Use DOMAIN_NETWORKS to allow the use of role::xenon in labs. Change-Id: I90ace12a5462ecd82509f0b3392fcd325628b8cb 07 July 2016, 17:18:35 UTC
c66c5e9 Chad Horohoe 07 July 2016, 16:38:04 UTC Phabricator: Don't run dumps or mail scripts from non-primary host phab2001 is a backup node right now, it shouldn't be doing the dumps and mail scripts. The former is wasteful and the latter can result in duplicate mails. Change-Id: Ib9815fe6a8d2e9e2978d284cf64f9bee03cea90f 07 July 2016, 16:38:48 UTC
f235d2c Eric Evans 07 July 2016, 16:25:30 UTC Upgrade remaining rack 'd' Cassandra nodes to 2.2.6 Bug: T126629 Change-Id: I03b1e356ef2257263e63fad147de475a793e69b2 07 July 2016, 16:31:13 UTC
c3a2635 Chad Horohoe 07 July 2016, 15:17:18 UTC Phab: properly disable crons for maintenance Follows up Iab5e0fb5. Commenting a cron doesn't remove it from the system. These should not be running at the moment but they still are Bug: T138460 Change-Id: Ide926d45b246a2f2c44d0db69ec3f0b928d4c41c 07 July 2016, 16:24:07 UTC
f87da4d Andrew Otto 07 July 2016, 15:03:09 UTC Include hive::client role in hadoop::worker role to get hive-site.xml and other deps Change-Id: Ia8a59b8b0143111d4a59cae168776f8561ebd510 07 July 2016, 15:05:56 UTC
65a8b69 Emanuele Rocca 07 July 2016, 10:20:53 UTC package_builder: add WMF lintian vendor profile Lintian's behavior can be customized through the use of vendor profiles. Introduce a new profile called wikimedia for WMF-specific customizations and provide a data file for the changes-file check adding jessie-wikimedia and others to the list of known dists. In the future, additional customizations can be added by placing files under modules/package_builder/files/lintian-wikimedia/. See Lintian User's Manual for the details of how vendor profiles and vendor-specific data files work. This commit also sets the newly created wikimedia vendor profile as the default by adding `LINTIAN_PROFILE = wikimedia` to /etc/lintianrc. Ref: https://lintian.debian.org/manual/index.html Change-Id: I739c761daac0c773a45836179f2f53dec47af676 07 July 2016, 14:51:21 UTC
c475660 Andrew Otto 07 July 2016, 14:32:15 UTC Include analytics_cluster::hive::client role on analytics1030 to see if this fixes HiveSpark in cluster mode Change-Id: I784b26885f0dad2ef435d1924bb7d83754f7be93 07 July 2016, 14:49:17 UTC
e955bc9 elukey 07 July 2016, 13:58:40 UTC Add analytics to the AQS monitoring contact group. Change-Id: Ib7b4f365c73d1574fd0d38db854a7826207e16d6 07 July 2016, 14:31:10 UTC
1ae8a73 Giuseppe Lavagetto 07 July 2016, 13:50:24 UTC etcd::backup: fix scripts when there are no logs to remove Change-Id: I77a87c2e8d00c06fc74fcd5fe2b494d2a91c77cb 07 July 2016, 13:50:24 UTC
22f7fc3 andrewbogott 06 July 2016, 17:25:12 UTC Fix c/p bugs in dhcpd config for new labvirt servers Change-Id: Iafb98a83066b6b4a61c60de4c9077408cbab88ca 06 July 2016, 17:25:37 UTC
efd1182 Jcrespo 07 July 2016, 11:50:23 UTC Fix special case when the mariadb server is not a slave When the server has not replication active, an array is not returned. Move the sanitization down, only when the status is correctly returned. Change-Id: I287aa7f35dacdcd73c0ce4ff7667fc8a3037299b 07 July 2016, 11:56:56 UTC
37d2a93 YuviPanda 05 July 2016, 17:55:34 UTC tools: Upgrade kubernetes to v1.3.0 Change-Id: I9e6fb60c8ff40f30ec21a30f786a5dd33df2d476 07 July 2016, 11:55:58 UTC
d4c3840 YuviPanda 07 July 2016, 11:51:07 UTC tools: Make toolschecker return FAIL when it fails NOT OK looks like it passes our Icinga check for 'OK', but is thwarted by the 503 vs 200. This is still clearer. Change-Id: I822b1242514b87d0068d3d75253e24ddb737f6c1 07 July 2016, 11:54:12 UTC
f21ff1f YuviPanda 07 July 2016, 11:19:33 UTC tools: Add icinga check for kubernetes webservice Bug: T131929 Change-Id: I798bae8d4fb170d0515a4d114cdd3cc1f5caeac1 07 July 2016, 11:21:14 UTC
9074e70 Jcrespo 07 July 2016, 11:16:43 UTC Sanitize SQL errors printed to icinga and IRC Do now show the query that caused a MySQL replication error. Bug: T122457 Change-Id: I235704aa1b2243066c1fae65dcb6c93a1454f5be 07 July 2016, 11:18:31 UTC
1330c39 YuviPanda 07 July 2016, 11:15:55 UTC tools: Fix k8s webservice backend check Bug: T131929 Change-Id: I5c2265eb023da8779c3d41cc3737cc69c3c72ccb 07 July 2016, 11:16:48 UTC
cc6b63e Jcrespo 07 July 2016, 11:05:33 UTC Update mariadb submodule (change mariadb alert location) Change-Id: I7224fa2cadf75a019cce6dd497fe2bfd1a0f051e 07 July 2016, 11:07:13 UTC
0adcd3a Daniel Zahn 01 July 2016, 14:48:00 UTC icinga: move check_mariadb plugin into module should go with Change-Id: I8854f6b9a3349fe That fixes the very last "puppet urls without modules" warning, globally. Change-Id: I43416d92da4d452dab5586de25304cbe520193fb 07 July 2016, 10:58:50 UTC
a402d66 Alexandros Kosiaris 07 July 2016, 10:03:42 UTC Introduce network::constants::frack_networks Fundraising is as far as networking goes in its own realm. Add a $frack_networks variable and ferm macro to allow setting firewall rules in a more finegrained manner for fundraising services accessing other realms services Change-Id: I3a6327e46d94801aaf0fae6a7ff17b33c1cc7d4a 07 July 2016, 10:32:46 UTC
d2c9a5c Moritz Muehlenhoff 07 July 2016, 10:22:51 UTC Add fonts-taml-tscu font to scalers Bug: T117919 Change-Id: Idc9d1b0a80822d6f8d44e41e6d0f32e8e53e026c 07 July 2016, 10:22:51 UTC
07afb4e Guillaume Lederrey 06 July 2016, 21:37:29 UTC Correct scoping issues in role::osm::master On labsdb1006, augeas fails as slave IP address is empty. Variables are referenced as top scope, when in this case they are defined in node scope. Moving those variables to hiera is cleaner as it removes this unusual node scope variables. Change-Id: Ide5770b0bff3f48636b5719056b8d275599c7aa5 07 July 2016, 09:56:05 UTC
8de4bae Elukey 07 July 2016, 09:39:20 UTC Revert "Revert "Raise the Hadoop HDFS datanode heapsize to 2GB."" Restore the original commit because we ruled it out from the weird heap size decrease observed for the Hadoop namenodes. It turned out that we were deleting tons of files from HDFS at the same time. This reverts commit 13ad76089a73671243b88c281cde04b639712c4b. Change-Id: I3b1b35c795a479117da6c2350c1ffcd8784a7f19 07 July 2016, 09:39:33 UTC
2735f67 Moritz Muehlenhoff 06 July 2016, 12:04:18 UTC dumps: Restrict to PRODUCTION_NETWORKS Only accessed from hosts in the production networks. Change-Id: I09ce8f370c9a0f786963846127d9716b950f2026 07 July 2016, 09:08:09 UTC
fcbfb93 Daniel Zahn 07 July 2016, 02:08:43 UTC remove erbium, gadolinium from hiera These remnants should have been removed along with I4e1dd35f7f4ea5. Thanks to Krenair for pointing them out. Bug:T123029 Change-Id: Ibd51517478081d0d39cac8ecaa93309f7cecd609 07 July 2016, 02:15:30 UTC
db835a4 Chad Horohoe 06 July 2016, 23:20:39 UTC Gerrit: A few minor tweaks to rsync replication - Only need the bare hostname, not the rsync:// protocol - Put the contents in /srv/gerrit since we're copying the whole directory. Otherwise we're in /srv/gerrit/git/git Bug: T125018 Change-Id: Ida5f66466b47c9fd766f13b658fb29e891a53524 06 July 2016, 23:21:42 UTC
c076608 Chad Horohoe 01 July 2016, 18:58:58 UTC Gerrit: Setup rsync between old and new machines Setup rsyncd on the new server, lead, to copy from the old server, ytterbium, and push gerrit data. Note how i changed the actual rsync command in the cron, switched the order of source and destination, since we are pushing. Bug:T125018 Change-Id: Id9d3020a3be0f848e9a39b878664758a5a2b6cfd 06 July 2016, 22:50:52 UTC
6c444ab Daniel Zahn 05 July 2016, 20:19:10 UTC admin: replace ssh key for andyrussg Replacing the SSH key for user andyrussg as requested on T139213. New key from P3333 pasted by https://phabricator.wikimedia.org/p/AndyRussG/ linked to https://www.mediawiki.org/wiki/User:AGreen_%28WMF%29 confirmed via: https://office.wikimedia.org/w/index.php?title=User:AGreen_%28WMF%29&oldid=191337 Bug:T139213 Change-Id: Id2a735d570e4cf23a8479195ce569db9a9e3ce5b 06 July 2016, 22:37:32 UTC
25b5763 Eric Evans 06 July 2016, 20:21:58 UTC Upgrade restbase1009 to Cassandra 2.2.6 Bug: T126629 Change-Id: Ic15fff2bf69bd03103c8948e46cc1527f6e3896b 06 July 2016, 20:21:58 UTC
ab5158d Guillaume Lederrey 29 June 2016, 11:46:27 UTC Postgresql - allow multiple entries for the same user in pg_hba.conf There are cases where we need multiple entries for the same user. For example multiple slaves for replication are using the same user from different IPs. This change changes the contract of postgresql::user. Bug: T138092 Change-Id: I54c7acac16e359579365d0f2d748c2197f5bba70 06 July 2016, 19:41:42 UTC
972de0a Andrew Otto 06 July 2016, 19:16:11 UTC Frack uses analytics-eqiad kafka cluster! Switching back to ALL_NETWORKS Change-Id: Ic81f7b730407cfc50bc37797f406b59a299c26de 06 July 2016, 19:16:11 UTC
f86af83 Nathaniel Schaaf 06 July 2016, 14:39:05 UTC Change path to proxy node requests Bug: T134782 Change-Id: I5c972cc1ebd8445f7d89db1acbc710b75526429a 06 July 2016, 18:41:04 UTC
2d6134b Eric Evans 06 July 2016, 17:17:02 UTC Upgrade rack 'b' Cassandra nodes to 2.2.6 Bug: T126629 Change-Id: I3aa8c7557f7a306d3059c6b99e3ef1bd64e37265 06 July 2016, 18:33:57 UTC
09c7dcc Daniel Zahn 05 July 2016, 22:36:10 UTC DHCP: fix promethium.eqad entry This was broken by accident in I470aa3600e3317d84, just restoring to how it was before. Bug:T120262 Change-Id: I895180a64efa4bd32533d8e1133e3e4dc7e3f248 06 July 2016, 18:28:04 UTC
13ad760 Elukey 06 July 2016, 17:35:54 UTC Revert "Raise the Hadoop HDFS datanode heapsize to 2GB." Unexpected behavior of the two Namenodes, it seems that the old setting is not correctly taking precedence as expected and tested. Rolling back a precautionary step. This reverts commit 01128856710f3618d0780cdc8cb40f81270460ed. Change-Id: I3a18e43383846ba37f55ad7a7dff0d11c5110e44 06 July 2016, 17:36:24 UTC
1b66a27 YuviPanda 06 July 2016, 16:54:14 UTC tools: Don't include k8s::ssl when not necessary It's now only neded when you want the private key as well, which is right now just the worker node Change-Id: Ia2e6c3a02a52aff027cde84245048a99fb357446 06 July 2016, 17:17:57 UTC
afd3fa8 YuviPanda 06 July 2016, 16:25:23 UTC tools: Don't set CA explicitly for kube2proxy Bug: T139461 Change-Id: Ie0e48627318fcf45bfe147e47bff1bbd385e48a0 06 July 2016, 17:17:57 UTC
24aefaa YuviPanda 06 July 2016, 14:04:55 UTC tools: Use provisioned cert instead of puppet cert Bug: T139461 Change-Id: I52f9b35dbf6b5a6d503c619ada3d1157ed13ca61 06 July 2016, 17:17:57 UTC
5b6c1f1 YuviPanda 06 July 2016, 14:33:23 UTC tools: Don't specify CA explicitly for client config Since we've merged the puppet CA to be part of the default CA bundle anyway Bug: T139461 Change-Id: I6f213b6806b5ea9327487167de299473f4f0e684 06 July 2016, 17:17:57 UTC
40b2106 YuviPanda 06 July 2016, 13:32:31 UTC tools: Provision star.tools.wmflabs.org cert for k8s master Bug: T139461 Change-Id: I4527f18eaf01fdbaed0d8e172e2fd43299d60f56 06 July 2016, 17:17:57 UTC
dbf1605 YuviPanda 30 June 2016, 14:13:58 UTC tools: Provision accounts for all tools maintain-kubeusers does the following: 1. Generate accounts for all tools 2. Generate abac file granting appropriate access to all tools 3. Read up infrastructure user config from a different file and make sure they are present in final token auth file list 4. Restart kube-apiserver if needed This needs to run as root since we're writing to many differnt users' homedirs, so we attempt to limit the damage by whitelisting capabilities and making most filesystem paths be ReadOnly Bug: T133999 Change-Id: I7a7a3dd951db2209e820752c4d14d77eb836b929 06 July 2016, 17:17:57 UTC
c23304b Antoine Musso 05 July 2016, 10:29:43 UTC spec fix for aptrepo and installserver Align installserver to use puppetlabs_spec_helper. Install server had the apt_repository class moved to the module 'aptrepo' but the spec hasn't been migrated. Do so and elevate aptrepo with the puppetlabs_spec_helper bits (.fixtures.yml, Rakefile, .rspec). Adjust installserver/install_server_apt_repository_spec and split it in two bits: - aptrepo_spec - distributio_spec (for apt::distribution) Inject some facts to please the wmflib os_version() function. For installserver, remove our custom fixtures setup in the Rakefile and just replace it with the puppetlabs_spec_helper + .fixtures.yml Add some "it { should compile }" except for the installserver::web_server that has a non trivial compile issue. Bug: T78342 Change-Id: Ica92f23e8cb6421c76bd4e12d5702f23da674dd0 06 July 2016, 15:52:10 UTC
9d09b03 Eric Evans 06 July 2016, 15:06:18 UTC Upgrade remaining rack 'a' nodes to 2.2.6 Bug: T126629 Change-Id: Ia06e9a8dc0b296fa01b44bc57eb8498c38b149b6 06 July 2016, 15:06:18 UTC
0112885 elukey 01 July 2016, 12:11:08 UTC Raise the Hadoop HDFS datanode heapsize to 2GB. We have noticed Java OOM errors anticipated by long GC pauses on a lot of Analytics Hadoop worker nodes. There is plenty of space in these servers to accomodate more heap space for a single daemon; this should resolve temporarily spikes in allocations causing OOMs and possibly reducing garbage collection occurrences. More heap size means also more time spent on garbage collection, but since the Hadoop cluster is not a latency mission critical system (but more focused on throughput) it should not be a problem. Bug: T139071 Change-Id: I4791115317770b213a9e927458709d70e42bac74 06 July 2016, 14:46:51 UTC
e7a87d2 Eric Evans 06 July 2016, 14:34:11 UTC Revert "Bootstrap Cassandra instance restbase1009-c" This reverts commit 2348573f543f478a450e8a0c9be64b296cf216cb. Bug: T139362 Change-Id: I1a9dfd3a645da9aa1cbfcf887daa5d9ee80d38d3 06 July 2016, 14:34:44 UTC
cf2cbd8 Andrew Otto 08 June 2016, 09:48:09 UTC Include librdkafka-dev in contint::packages::python This will allow testing using confluent-kafka, python bindings for librdkafka Bug: T133779 Change-Id: I65f691c563ee6da2c074ba3defb82df20ad71b1d 06 July 2016, 14:22:48 UTC
2bb434d Filippo Giunchedi 04 July 2016, 09:18:54 UTC install_server: pre-provision swift uid/gid Bug: T123918 Change-Id: Ic2e6ec6acc68b4046a202d1da76fffdef0c5e0ca 06 July 2016, 14:09:18 UTC
2c7fc94 Jcrespo 06 July 2016, 14:01:31 UTC Remove /a directory from db1048 Bug: T138460 Change-Id: Id0627ccd47de743c9b5e7520f15f363ffb309f03 06 July 2016, 14:01:31 UTC
21b215d Guillaume Lederrey 05 July 2016, 16:02:08 UTC Ensure base cassandra directory is created. This introduces a new cassandra::default_data_directory_base parameter, used only for default instances (not multi-instances), which is used to ensure the base directory is created. This is fairly ugly as it introduces one more difference between single and multi instance (note that we already have quite a few parameters following this pattern). A probably better alternative would be to remove everything related to instances from ::cassandra and let the caller manage either single or multi instance. But that's a bit more than I am willing to risk with my current understanding of the usage of this module. Better proposal welcomed! Bug: T138092 Change-Id: If69ef8dfff534d178472527c8c4fc47cffa7df10 06 July 2016, 13:40:25 UTC
0e74dbc Jcrespo 06 July 2016, 13:18:16 UTC Preparing db1048 for jessie install Set db1043 and db1048 as installing jessie by default. Changing puppet to use MariaDB 10 on db1048. Bug: T138460 Change-Id: I260e9c83cb00625805d44473d9f23d4dbd9274c3 06 July 2016, 13:27:17 UTC
65f735a Jcrespo 06 July 2016, 12:59:57 UTC Change otrs dumps to use the fqdn, as it fails from codfw Change-Id: Id84a35e8031798eeea8c4f187c9c7a0cdd5f84d6 06 July 2016, 13:01:35 UTC
b420cc9 Jcrespo 06 July 2016, 10:56:45 UTC Change dump-otrs.sh script permissions to 755 Backup failed due to lack of execution permissions. Also, the script itself does not contain any private information. Change-Id: I202ee8e076454dc230f83617ba0795f32cf7fcda 06 July 2016, 12:51:30 UTC
586c057 Alexandros Kosiaris 05 July 2016, 19:49:24 UTC Provide labtest realm with it's own copy of network::subnets The labtest realm will hopefully stop existing at some point really soon, making all of this redundant, but until then provide network::subnets as well in the labtest hiera data in order to allow slice_network_constants functionality Change-Id: Iffcdd8c605bbbb611bb110d0264f5e6363f9d999 06 July 2016, 12:29:39 UTC
5a8e60d Moritz Muehlenhoff 05 July 2016, 08:12:13 UTC role::labs::graphite: Further use of LABS_NETWORKS These are only accessed by labs instances, production hosts use a different graphite host. Change-Id: Idc5d0686b846b2a74b88bbc3d8a47422e3d96b7c 06 July 2016, 11:49:50 UTC
b7570b2 Moritz Muehlenhoff 05 July 2016, 14:55:40 UTC role::kafka::main::broker: Use DOMAIN_NETWORKS This is only accessed from production systems. Use DOMAIN_METWORKS to still allow setting up a test instance in labs. Change-Id: I15adb3c85bf85565b2374ce50569458dc9f16cab 06 July 2016, 11:19:48 UTC
144dfd9 YuviPanda 06 July 2016, 11:16:04 UTC tools: Add a check for k8s backed webservices Bug: T131929 Change-Id: I2101975c2c000b63ae880d5bf3ba607c7aa05e2d 06 July 2016, 11:16:35 UTC
bb9d3b2 Filippo Giunchedi 06 July 2016, 11:11:59 UTC swift: adjust group ownership for Ubuntu/Debian Special-case 'syslog' group on Ubuntu. Bug: T137397 Change-Id: If2a898f6c6b74f4d72892a8612bfb9021d65dc79 06 July 2016, 11:12:00 UTC
3734d85 Filippo Giunchedi 06 July 2016, 10:53:44 UTC swift: adjust group ownership/perms for /var/log/swift Make it work on ubuntu too, on ubuntu rsyslog runs as syslog:syslog, on debian runs as root. Bug: T137397 Change-Id: I4f6e6df334d75e2676d651a5e31140a3f3015441 06 July 2016, 10:57:19 UTC
c899510 Guillaume Lederrey 05 July 2016, 20:13:46 UTC Maps - notify tilerator of new expire files This corrects: * call should be POST, not GET * params should be passed as query string Change-Id: Ie5e45952140984acdcc9fe692fcf9b58db7c9d35 06 July 2016, 10:45:48 UTC
216bf79 Filippo Giunchedi 06 July 2016, 10:31:54 UTC swift: more inclusive rsyslog matching the 'swift-' prefix in syslog is used only as the daemon names (e.g. systemd) and not by swift itself. Bug: T137397 Change-Id: Ib5134b6943129f24897520c03efc34282a1020c5 06 July 2016, 10:31:54 UTC
86e42e6 Filippo Giunchedi 16 June 2016, 09:08:32 UTC swift: redirect syslog from all daemons to separate file This is to avoid spamming /var/log/syslog, also decrease the chance of logs filling up the disk with a more restrictive logrotate retention. Bug: T137397 Change-Id: Iec42489838e5e46c09764b15c96d3ff5e6ab68b6 06 July 2016, 10:23:19 UTC
d554cb1 Filippo Giunchedi 22 June 2016, 15:04:26 UTC install_server: smaller root for single-disk /srv rationale: we're standardizing on using /srv for service data, as a consequence the root filesystem doesn't need to be as big. It can also be resized later from space in the VG if needed. snapshot1006.eqiad.wmnet: Filesystem Size Used Avail Use% Mounted on /dev/dm-0 74G 6.0G 64G 9% / VG #PV #LV #SN Attr VSize VFree snapshot1006-vg 1 2 0 wz--n- 446.82g 89.37g snapshot1007.eqiad.wmnet: Filesystem Size Used Avail Use% Mounted on /dev/dm-0 74G 5.9G 64G 9% / VG #PV #LV #SN Attr VSize VFree snapshot1007-vg 1 2 0 wz--n- 446.82g 89.37g snapshot1005.eqiad.wmnet: Filesystem Size Used Avail Use% Mounted on /dev/dm-0 74G 5.9G 64G 9% / VG #PV #LV #SN Attr VSize VFree snapshot1005-vg 1 2 0 wz--n- 446.82g 89.37g labstore2003.codfw.wmnet: Filesystem Size Used Avail Use% Mounted on /dev/dm-0 74G 1.6G 68G 3% / VG #PV #LV #SN Attr VSize VFree labstore2003-vg 1 2 0 wz--n- 10.91t 0 labstore2004.codfw.wmnet: Filesystem Size Used Avail Use% Mounted on /dev/dm-0 74G 1.7G 68G 3% / VG #PV #LV #SN Attr VSize VFree labstore2004-vg 1 2 0 wz--n- 10.91t 0 Change-Id: Ia63d9af21c7a68d7236f540f611a8ffbe8a9c569 06 July 2016, 10:04:52 UTC
bd5386a Jaime Crespo 01 July 2016, 10:59:09 UTC Disable crons using the phabricator db slave due to maintenance Bug: T138460 Change-Id: Iab5e0fb5823532b158c649a4bfb78bb4755a61ac 06 July 2016, 09:26:45 UTC
4001f2d Guillaume Lederrey 04 July 2016, 15:41:37 UTC Correct hiera property name to create postgresql users Example of configuration for codfw cluster was wrong, which lead to configuration of labs instance to also be wrong. Bug: T138092 Change-Id: Iebe7ec41d8b98c78eacd359cf20c32aa09261a00 06 July 2016, 08:35:28 UTC
e4df1c6 Nicko 19 June 2016, 15:13:20 UTC Include a cassandra::instance::monitoring class In order to follow the single responsability principle, it could be great to externalize the monitoring setup from the general cassandra setup. This is the next step to include the monitoring setup in all the calls to ::cassandra Bug: T137422 Change-Id: I9177f9797bfb5a13696fa3ad2b9af5264def5513 06 July 2016, 07:59:52 UTC
9129b87 Daniel Zahn 05 July 2016, 17:18:21 UTC site.pp: split hydrogen/chromium, move restbase-test Instead of using a regex for hydrogen and chromium to then follow with an "if hostname" inside that, split them up, at least until they are actually the same (again?). Also move restbase-test nodes into alphabetical order. http://puppet-compiler.wmflabs.org/3272/ Change-Id: If38f67af0e52def7203ff17bd1734c50c3d46284 05 July 2016, 23:51:13 UTC
86ed0ce Chad Horohoe 05 July 2016, 20:23:41 UTC Gerrit: Don't install python-paramiko anymore Can't even remember why we had this. Old hooks I think? Whatever, don't need it now... Change-Id: I8446448d27a34af3a68f0e119b8226b43b415fa7 05 July 2016, 22:44:52 UTC
634f441 andrewbogott 05 July 2016, 22:28:22 UTC Depool labvirt1011, pool labvirt1001. In theory the issue with 1001 has been fixed, and 1011 is suddenly badly broken. Change-Id: I369012ac1794c826b7ecdd339e4831e58828a5d2 05 July 2016, 22:28:22 UTC
bf14a9d andrewbogott 05 July 2016, 22:19:10 UTC Fix typo in I14c786544a27d9a540da1928d54ae6b9839cc482 Change-Id: Ife80f58a07d31804aeb781b544e33e0da4785af4 05 July 2016, 22:19:10 UTC
83bda78 andrewbogott 05 July 2016, 21:56:08 UTC Disable user creation of new VMs until we increase capacity. Change-Id: I14c786544a27d9a540da1928d54ae6b9839cc482 05 July 2016, 21:56:35 UTC
869f5ee Andrew Bogott 05 July 2016, 20:58:50 UTC Revert "labs: Take out all hosts other than labvirt1011 out of pool" I don't think they were ever full. In any case, T137857 may be resolved. This reverts commit 005dba709296dbc9711aedd69803c3241ef458f3. Change-Id: I411b7bf5fbb3d70bc9bd466e869e56b7fa21cb93 05 July 2016, 21:02:15 UTC
2be2335 andrewbogott 05 July 2016, 12:01:46 UTC Partman recipes for labvirt1012-1014. Bug: T138509 Change-Id: I890b8b9290736336c7b09ec6e2d37c582b5acdc2 05 July 2016, 19:28:31 UTC
3d008f2 andrewbogott 02 June 2016, 06:51:01 UTC Ensure mariadb running on labservices hosts. Without it pdns will fail. Change-Id: Ic6e649995e18b9321934f2e72cebcd6142a451ae 05 July 2016, 19:23:02 UTC
ad40b0c Merlijn van Deen 05 July 2016, 17:55:11 UTC toollabs: set nano as default editor Bug: T100526 Change-Id: I637a83ed52fcfbe1ac0f203fdaff69d5e289df5f 05 July 2016, 17:58:32 UTC
3b16771 andrewbogott 02 June 2016, 06:49:06 UTC Ensure mariadb service is running on wikitech host. Bug: T125987 Change-Id: I7a5cbd70be824a4e29192caa0c7482a5bcd2869b 05 July 2016, 16:19:28 UTC
2348573 Eric Evans 05 July 2016, 15:26:28 UTC Bootstrap Cassandra instance restbase1009-c Bug: T139362 Change-Id: I6991673fba6bb0052968c196c449d62c8edcc1ff 05 July 2016, 16:07:27 UTC
937582d Daniel Zahn 29 June 2016, 21:17:09 UTC admin: add wdqs-admins to deploy-service group As requested in T138628, adding the current members of wdqs-admins to the deploy-service group. Reasoning has been "order to be able to deploy WDQS via scap3". Bug:T138628 Change-Id: Ib2557f150f385476a3b5073ed75711bdd3524b0e 05 July 2016, 16:02:41 UTC
c61d1fc Moritz Muehlenhoff 24 June 2016, 06:45:12 UTC url_downloader: Use DOMAIN_NETWORKS in ferm rules The url downloaders in production are only accessed by production servers, but we also have an url_downloader instance in deployment-prep. Change-Id: I4be5ac7cca504fa4cee6855522d130b9aa13bce7 05 July 2016, 15:46:45 UTC
back to top