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
1bb5d19 Yuri Astrakhan 28 August 2015, 19:09:12 UTC Added maps-cluster referer rules (e.g. Phab) * Allow phabricator, mediawiki, wikitech, and direct root linking to the maps.wikimedia.org - in order to better document the system and engage the development community. Change-Id: I2f7246b246be91d0b4b4489666afe508d6ffdbfd 31 August 2015, 19:22:34 UTC
cfff82c dzahn 31 August 2015, 15:19:49 UTC icinga: libssl0.9.8 for NRPE checks to run T110893 says that one of the manual steps was to install libssl0.9.8 (in addition to libssl1.0.0) for some NRPE checks to run and that we should puppetize that. Indeed both packages are installed on neon. Bug:T110893 Change-Id: I7c255b300ae106c6cf752c0325ddd97fff264ffa 31 August 2015, 19:05:15 UTC
8ce37ae dzahn 31 August 2015, 15:23:26 UTC icinga: no suprise upgrades with 'latest' Let's not have suprise version upgrades of Icinga because monitoring itself is pretty critical. Let's do manual scheduled upgrades instead in this case. Change-Id: Ia3b2762bc53a11d183a07cfc7bc69b53e3ec04be 31 August 2015, 19:04:58 UTC
7a756dd Faidon Liambotis 30 August 2015, 18:30:47 UTC openstack: fix non-capitalized resource references Fixes deprecation notices. Change-Id: Ia7e7bd6a267ec8fbe923d8d660db883950a3c006 31 August 2015, 17:42:52 UTC
9e62d04 Brandon Black 31 August 2015, 16:17:35 UTC conftool: disable etcd1002 Seems etcd100x are dead (no ping), assuming this should be removed from the list that otherwise contains just conf100x Bug: T110899 Change-Id: Ic8efc279cc6ed1fa60f496eba5639a7807db6ee5 31 August 2015, 16:18:36 UTC
e9c0ba9 Federico Leva 30 August 2015, 21:41:56 UTC [Italian Planet] Update Wikimedia Italia feeds Change-Id: I059c0cbdef82774513eb14b4b60004abc4a4a869 31 August 2015, 16:17:16 UTC
cfa675e andrewbogott 31 August 2015, 15:43:46 UTC Revert "Add more nslcd notifying" Didn't help and introduced broken deps in a few places. This reverts commit 00d0344fc2ecfac04c8d0b294bc7b6b85612f1f8. Change-Id: Ibb9a4e60074d883914e824ace41d909b51aea851 31 August 2015, 15:47:56 UTC
9df1191 Brandon Black 27 August 2015, 14:32:12 UTC add various text backend defs to mobile This adds the backend defs for: backend_random in tier-two and the rendering, sec audit, and restbase backends in tier-one. It doesn't actually make use of them yet, which will be in a followup patch. Bug: T109286 Change-Id: I0304dc83296e0789f78b894b45d948afe7fd39c2 31 August 2015, 15:46:25 UTC
2ffdf7f cpettet 28 August 2015, 20:40:22 UTC elasticsearch: ferm for 4-7 Change-Id: I60014738a9c37cfcded99988df5825b7700cdb07 31 August 2015, 15:38:50 UTC
b6e36a6 Alex Monk 29 August 2015, 14:47:37 UTC Fix /static 404s in beta mobile Since there is no www.wikimedia.org host in beta cluster, when the mobile-frontend cache sends a request to that host using the mobile-backend ip and port the result is a 404; however if you use a valid beta host the static directory works as expected. Follow-up I7ad956fb Bug: T105541 Change-Id: Ief3a78d538d016be5eb25b9f58ba5d1c56988088 31 August 2015, 15:31:00 UTC
a1fe872 Alex Monk 29 August 2015, 00:09:57 UTC Fix wikitech beacon 204 This was applied to the port 80 VirtualHost, which is not particularly useful. Follows-up: I7bfdac7b Bug: T104359 Change-Id: I9b1ec64162f9916bcbe1458cd320ea90e1e89deb 31 August 2015, 15:29:39 UTC
75ccadf Andrew Otto 28 August 2015, 15:01:43 UTC Set up varnishkafka instance on cache servers to log raw client side events to kafka This will soon fully deprecate the UDP based varnishncsa instance, allow us to turn off the eventlogging client side UDP -> ZMQ forwarder. Bug: T106255 Change-Id: I470e37575fae87cd85665af88c7212229e68b16b 31 August 2015, 15:23:43 UTC
6ebf588 John F. Lewis 31 August 2015, 14:49:25 UTC icinga: puppetise apache mods Bug: T110893 Change-Id: I0f9de9df2d82e81a56e7c65765df5d593841535b 31 August 2015, 15:13:15 UTC
0f62ccf Moritz Muehlenhoff 28 August 2015, 11:12:35 UTC Enable debdeploy for an initial set of servers Will be moved into the base class at a later stage. Change-Id: I07b775ede04c141885ef239e290507c1680c52c4 31 August 2015, 15:01:15 UTC
00d0344 andrewbogott 31 August 2015, 14:54:07 UTC Add more nslcd notifying Because it still isn't getting restarted when it needs it. Bug: T110891 Change-Id: I5c89284a7856cc47ecbd9e37b54467a199dda1a5 31 August 2015, 14:58:20 UTC
7509808 Faidon Liambotis 31 August 2015, 14:51:55 UTC network: add bast2001 to bastions Change-Id: Ie8f26ac3c3808a5dfaadbf68a3384860e40aa00f 31 August 2015, 14:52:16 UTC
0ec05cd Faidon Liambotis 31 August 2015, 14:49:27 UTC network: indent/reflow $special_hosts (also kill a couple of long-forgotten pmtpa Labs IPs) Change-Id: I796bf1c0a2ce594241bae8c5d575c86283d14b62 31 August 2015, 14:52:16 UTC
342dc06 andrewbogott 31 August 2015, 14:40:46 UTC Explicitly include nslcd package. Bug: T110891 Change-Id: I2528330140d97080af3123584dc2f624b9771470 31 August 2015, 14:41:58 UTC
5d28d43 andrewbogott 31 August 2015, 14:33:37 UTC Make ldap/nscd/nslcd install order more explicit. This seemed pretty haphazard, and right now I'm seeing an issue where I have to restart nslcd after an instance is up to get it working. Bug: T110891 Change-Id: Id673119b6f971c00d2bb0c6d6a1fa8e038640f8f 31 August 2015, 14:35:42 UTC
6047bb1 Andrew Otto 31 August 2015, 14:02:08 UTC require_package openjdk-7 in zookeeper role Change-Id: Ic550b1c5478712f49218aa15041040965b86a7b4 31 August 2015, 14:10:19 UTC
ac09a37 Moritz Muehlenhoff 31 August 2015, 10:52:08 UTC Move base::firewall include into the role definitions Now that all existing swift systems have moved to ferm, move the include::firewall from site.pp to the respective roles Change-Id: I6359df383e63397d176c88014b0f31190e1b1dd5 31 August 2015, 14:09:55 UTC
ac2a9fd Moritz Mühlenhoff 06 August 2015, 13:10:41 UTC Add ferm rules for Hadoop worker nodes Bug: T83597 Change-Id: If36667152b3bcc679922f47cc0197faa4ecf2d93 31 August 2015, 13:57:22 UTC
7987bf6 Brandon Black 31 August 2015, 13:44:35 UTC varnishncsa: fix process monitoring We've had these in WARNING state and ACK'd for a long time now, because the old check was for 2x processes, and at this time some nodes have 2 defined and others have 3. This refactors as a separate 1-process check on each specific varnishncsa instance. Change-Id: Ia33609bef1e137f9086c5129f1fe895cc09c5deb 31 August 2015, 13:45:52 UTC
2f7e6ce Ariel T. Glenn 31 August 2015, 12:28:55 UTC dumps rsync to labs: generate and use list of lastr 3 good dumps updating from last 2, because we may include partial or in progress dumps Change-Id: Ib5358de8559a1c21bbebb4cb3f8f5d62c616c0ff 31 August 2015, 12:28:55 UTC
8b50899 Ariel T. Glenn 31 August 2015, 10:54:55 UTC dump lists for rsync: copy partial or in progress dump if most recent this means labs will get copies of dumps in progress, which will mean files show up as soon as they are available instead of waiting for all stages of all dumps to complete Change-Id: I154e1da7b4f3052260687464662eead3070b7107 31 August 2015, 12:05:57 UTC
7c50996 Filippo Giunchedi 31 August 2015, 09:40:51 UTC swift: enable firewall for ms-fe in eqiad Change-Id: I6bf0fd575ff7362cfd11cd20e5771964d8d0f116 31 August 2015, 09:43:37 UTC
4105374 Moritz Muehlenhoff 28 August 2015, 11:06:04 UTC Add debdeploy master to palladium Change-Id: Iaa731e538e15a1d85388912e636409d16298a124 31 August 2015, 09:08:29 UTC
8c100a8 Moritz Muehlenhoff 31 August 2015, 08:34:14 UTC Remove stray lines (nop commit to test puppet-merge after enabling ferm on puppetmaster backends) Change-Id: I940105e25c7d58f87fb047d33395e4b11790c995 31 August 2015, 08:35:43 UTC
b972586 Moritz Muehlenhoff 31 August 2015, 08:24:24 UTC Enable ferm on strontium (and merge it to the role) Change-Id: I3e7b5088ac3ab99eed0cd5802c2c624c08cf63b7 31 August 2015, 08:28:30 UTC
5ed040e Ori Livneh 31 August 2015, 08:16:07 UTC ssh-agent-proxy: break out of select loop once client is done By moving the `if code is None: return` code into a subroutine that is called from the select loop, we no longer break out of the loop when the client disconnects, leading to file descriptor exhaustion. Bug: T110794 Change-Id: I499c82490f77887d33c038e49c9b174bfcd0e57c 31 August 2015, 08:18:42 UTC
5ea62a1 Moritz Muehlenhoff 31 August 2015, 08:10:01 UTC Enable ferm on rhodium Change-Id: I7c43b07ff9ed82619877376c6bae6f7ed7d8b0d5 31 August 2015, 08:10:01 UTC
1ab63ed Moritz Muehlenhoff 28 August 2015, 17:16:35 UTC Query the puppetmaster from hiera instead of $::serverip On strontium the $::serverip resolved to strontium's IP address instead of the IP address of palladium, which made the ferm rule rather ineffective. Change-Id: I0fe4a0bb283c0d139b23f59ae88bbe4e726253ea 31 August 2015, 07:39:54 UTC
f8d55c8 Ori Livneh 30 August 2015, 18:22:55 UTC dotfiles: add script for guesstimating gzipped log file's line count * Add `lcz`, a small script I wrote to guesstimate the number of lines in a log file by decompressing a small chunk. * Use that to speed up log-trends. Change-Id: Ia98f4533c2f0d6a6e22e1fcc9d140f7fb55c6598 30 August 2015, 18:28:18 UTC
4f2f2a0 Ori Livneh 28 August 2015, 23:37:35 UTC harden ssh-agent-proxy Prior to this patch, ssh-agent-proxy only read the parts of each incoming message that it thought it cared about: the request code, and (in the case of signing requests) the public key blob. This is potentially insecure because a malicious client could potentially trick the proxy into forwarding a request that would otherwise get filtered by appending it to a legitimate request and deliberately under-reporting the message length. (Similar to the Heartbleed bug.) To protect against this, we make ssh-agent-proxy read and validate the entire message and its length against the agent protocol spec before forwarding it to the proxy. Change-Id: Ia00ccafcd0ab8a6856be56e276616bceb0ae9d74 30 August 2015, 18:27:44 UTC
310d7d0 Alex Monk 28 August 2015, 22:46:05 UTC Add link to developer app guidelines from dumps pages footer Bug: T110742 Change-Id: I76090696063b837aeff66b5124289336c5859641 30 August 2015, 13:29:33 UTC
c464ee9 dzahn 29 August 2015, 01:08:31 UTC mailman: fix Apache 2.4 require syntax pt.2 one more follow-up to I2c70d87922bcc65 Syntax error on line 27 .. negative Require directive has no effect in <RequireAny> directive <snip> When multiple Require directives are used in a single configuration section and are not contained in another authorization directive like <RequireAll>, they are implicitly contained within a <RequireAny> directive. Thus the first one to authorize a user authorizes the entire request, and subsequent Require directives are ignored. </snip> http://httpd.apache.org/docs/current/mod/mod_authz_core.html Change-Id: I25ff59808fcb077b3611637bf19cb7699163d755 29 August 2015, 01:21:14 UTC
30028ab John F. Lewis 29 August 2015, 00:08:53 UTC mailman: correct apache 2.4 usage of require not Apache 2.4 requires all granted followed by the require not statement. Fixes I58fe910e204d6d0d51b85ef1f589f0936a65b0ca Change-Id: I2c70d87922bcc65f3a7ca0aff53a15a09924eb8a 29 August 2015, 00:08:53 UTC
580d143 dzahn 28 August 2015, 21:28:36 UTC mailman: make Apache config 2.4 compatible Apache config needs to be adjusted for 2.4 here as well. still "client denied by server configuration" for monitoring and all Bug:T109925 Change-Id: I58fe910e204d6d0d51b85ef1f589f0936a65b0ca 28 August 2015, 23:46:58 UTC
f4e33f3 RobH 28 August 2015, 17:04:48 UTC ldap-eqiad.wikimedia.org cert renewal renewed cert, new cert content (no change in key) T106604 Change-Id: I3aa37df2157cf364d5cff38c6fbd51c97b9434bb 28 August 2015, 20:41:55 UTC
48dbf51 dzahn 28 August 2015, 00:51:33 UTC mailman: fix I/O monitoring on virtual machine On sodium we did not have to specify a device name, on fermium we have to, because it's /dev/vda, or: ERROR: Device incorrectly specified Let's try and use "is_virtual" here to know if we are on a VM or not and then set the device to be checked based on that. https://docs.puppetlabs.com/facter/1.6/core_facts.html#isvirtual --- notify{"am i virtual? $is_virtual - device: $iostat_device": } fermium:/Main/Notify[am i virtual? true - device: vda] sodium: /Stage[main]//Notify[am i virtual? false - device: sda] Bug:T109925 Change-Id: I6c453046b00ca1b12db354f1bcd938d01029948c 28 August 2015, 20:24:15 UTC
2d602fd cpettet 28 August 2015, 19:13:00 UTC icinga: fix host def for en.m.wp.o-legal-html check Change-Id: I4f7cfcbf746f844081d4e7d5cded4742d613419c 28 August 2015, 19:13:29 UTC
cb8c096 cmjohnson 28 August 2015, 18:36:58 UTC Adding the remaining Hadoop nodes an1053 and an1057 Change-Id: I67ea9dfd41c643664e952f032e8148ce0697368f 28 August 2015, 19:09:37 UTC
0c78c8a cpettet 28 August 2015, 18:58:59 UTC icinga: dupe resource name fix for legal Change-Id: I16775355cf09169917df95a734c2720aac1ac1ad 28 August 2015, 19:01:05 UTC
58dcfcd dzahn 26 August 2015, 17:49:18 UTC analytics: do not use node inheritance "Deprecation notice: Node inheritance is not supported in Puppet >= 4.0.0." So do not use node inheritance anymore. It did not save a lot of lines and node definitions in a role class are odd anyways. Change-Id: I840da1bf2a6ca54e1b3c06d92d7ac36ce61f6b91 28 August 2015, 18:57:32 UTC
6801c75 cpettet 28 August 2015, 18:34:34 UTC icinga: apply icinga::monitor::legal to role ref T108081 Change-Id: I4156e1595e3ff5305463c07d1bbb13796f681a77 28 August 2015, 18:36:15 UTC
8b5d4d2 cpettet 26 August 2015, 21:03:43 UTC icinga: watch for the existence of certain html This is going to be extremely iterative as there is little consistency across projects for these items: trademark, copyright, privacy policy, and terms of use. ref T108081 Change-Id: I7cdd33d6b097c0a34c97742110afaa6cf5a27cd5 28 August 2015, 18:17:21 UTC
ab6cca0 RobH 28 August 2015, 17:07:47 UTC ldap-codfw.wikimedia.org cert renewal renewing ldap codfw cert, no key change T106604 Change-Id: I7c0748338f664f522495df6d3369f6997aea9d4f 28 August 2015, 18:09:23 UTC
8d6979c dzahn 28 August 2015, 18:02:03 UTC mailman: convert French listinfo template File[/etc/mailman/fr/listinfo.html]/content.. ..failed: invalid byte sequence in UTF-8 listinfo.html: HTML document, ISO-8859 text -- iconv -f ISO-8859-1 -t UTF8 -> listinfo.html: HTML document, UTF-8 Unicode text Bug:110695 Change-Id: I09b0dfff2b814715b047044c669d28b1b703e407 28 August 2015, 18:08:11 UTC
405a54e cpettet 28 August 2015, 16:26:04 UTC phab: disable tools crons This is temporary as we are seeing large mysql connection spikes. I am suspcious of these jobs and I want to go a few days with them disabled to guage a baseline. There is no long term negative outcome here. Change-Id: I23cf4f30769e1bf03314ae88f8cbb5743725ba5d 28 August 2015, 17:59:04 UTC
247a062 dzahn 28 August 2015, 17:02:04 UTC mailman: fix "no gate_news on jessie" follow-up to Ib01baf00aca8c9 This was supposed to be "< jessie" not "<= jessie". Change-Id: I82040855cabca01da36ce83ede0f21b1116fc66c 28 August 2015, 17:50:32 UTC
9d3f43e Moritz Muehlenhoff 28 August 2015, 17:43:03 UTC remove stray line NOP commit to re-trigger puppet-sync from palladium to strontium/rhodium Change-Id: Ifce9955869afa0abc06a10edb96f384be12ce031 28 August 2015, 17:43:03 UTC
519f87e Moritz Muehlenhoff 28 August 2015, 16:27:45 UTC Disable ferm on the puppetmaster backends Still problems with the puppet merges, let's sort this out properly after the weekend Change-Id: Idef515cd2ad4a3bb19cf1e7922041f5411dfac0e 28 August 2015, 16:27:45 UTC
6cfb6a7 cpettet 28 August 2015, 15:59:08 UTC elasticsearch: ferm for [8-9][10-13] Change-Id: Ieea81bcf3b623b41774da6076b6122e8985fa880 28 August 2015, 16:00:03 UTC
782eb5c dzahn 28 August 2015, 02:30:50 UTC mailman: no gate_news debconf option on >= jessie "gate_news" is mailman's newsgroup to mail gateway. This option does not exist anymore in the jessie version of the package. @fermium:~# echo get mailman/gate_news | debconf-communicate 10 mailman/gate_news doesn't exist @sodium:~# echo get mailman/gate_news | debconf-communicate 0 false It causes puppet errors if we still try to set it on jessie: Mailman::Listserve/Exec[dpkg-reconfigure mailman]: Dependency Exec[debconf-communicate set mailman/gate_news] has failures: true Bug:T109925 Change-Id: Ib01baf00aca8c978e34b9553e29d1ef1bffee702 28 August 2015, 15:53:23 UTC
a3c073b Brandon Black 28 August 2015, 12:59:28 UTC refactor wikidata CA cookies workarounds a bit Bug: T109038 Change-Id: I243998e2c03d5e5e98f56293e89753ffab33921d 28 August 2015, 15:36:34 UTC
93529f3 Andrew Otto 28 August 2015, 15:10:01 UTC Don't alert on stopped eventlogging jobs on analytics1010 Change-Id: I0d325800554eb23bcc18d1d4e6300eff46200f4d 28 August 2015, 15:19:38 UTC
962cef3 cpettet 28 August 2015, 15:12:58 UTC elasticsearch: ferm for 14-17 Change-Id: I3c9614d5082ce0df240c17e902e707006698cc83 28 August 2015, 15:17:37 UTC
109a75f Brandon Black 28 August 2015, 15:14:07 UTC ssl_ciphersuite: disable DHE for apache for now Change-Id: I80ada9fb7c4b2f55731bd1de2bddd14f6f38b877 28 August 2015, 15:14:41 UTC
befeb77 Brandon Black 28 August 2015, 11:50:45 UTC ssl_ciphersite: bugfix for apache-2.4.8+ DHE selection Change-Id: I4101b7a1b962854e5decfc4003e547f2d9c46e85 28 August 2015, 15:00:35 UTC
1b1e580 Antoine Musso 28 August 2015, 14:22:39 UTC nodepool: update apt pinning Our Jessie servers no more included upstream jessie-backports. Update the pinning to jessie-wikimedia/backports. Bug: T110656 Change-Id: I61d99be3a735abe7f84dd5f4a9c63e3882839e25 28 August 2015, 14:22:39 UTC
379b167 Andrew Otto 28 August 2015, 14:09:25 UTC Rename server-side kafka processor to server-side-0 Change-Id: I1e09ea3f5b28f7fc05dceb86eaf8b2dda6e74fa6 28 August 2015, 14:09:25 UTC
cf20512 Andrew Otto 27 August 2015, 20:23:55 UTC Puppetize multiple kafka eventlogging processors on analytics1010 Bug: T104228 Change-Id: I9c9514e5461e3c34c608fb0543b1a3f7a21df836 28 August 2015, 14:04:59 UTC
13b642e Andrew Otto 26 August 2015, 15:08:29 UTC Improve description for statsistics-privatedata-users group in data.yaml Change-Id: I7f2f2f650cef8a8feb41c74452af40f9c360f302 28 August 2015, 13:53:07 UTC
a193363 Chad Horohoe 24 August 2015, 19:11:20 UTC Assign swift roles via ENC Bug: T91553 Change-Id: Ibb0139e8857d068f6a57e648824b9e2f12e69806 28 August 2015, 13:47:45 UTC
9841505 Moritz Muehlenhoff 28 August 2015, 12:40:54 UTC Allow SSH access from puppet master for puppet-merge Change-Id: I2da89c774718c0e9c6d01aaa1febe1426490b9ff 28 August 2015, 13:35:10 UTC
80a1f79 Jan Zerebecki 28 August 2015, 12:20:56 UTC Another CentralAuth double cookie workaround There was still another cookie that if duplicated prevents login. Bug: T109038 Change-Id: I0046aef6d8b294d9d079d778f72ee19e3085c751 28 August 2015, 12:36:44 UTC
60dd738 Alexandros Kosiaris 28 August 2015, 12:33:20 UTC maps: Fix typo in maps-admin name s/maps-admin/maps-admins/ Change-Id: I526f602fcbf7ce44460fbb9ddb7b17b06208d983 28 August 2015, 12:33:20 UTC
67b7817 Alexandros Kosiaris 27 August 2015, 14:12:27 UTC maps: ensure PostgreSQL's logs as maps-admin Ensure postgresql logs as maps-admin to allow maps-admin to read them Rely on logrotate's copytruncate policy for postgres for the rest of the log files in /var/log/postgresql We should find a better way of doing this. Abandoned efforts include: * sudo => wait to complex to right a good rule that does not make people's lives miserable but still works * adding adm group to maps-admins groups => not really possible with our current admin module or the puppet group resource. The puppet group resource provider on linux (groupadd) does not support the manages_members feature: https://docs.puppetlabs.com/references/latest/type.html#group-provider-features The admin module does not allow us to add people in groups on a per host basis. This means groups are global and that was a design goal back then. Bug: T106637 Change-Id: Ic27e3248c1357fe9797716a16301f3693c530e22 28 August 2015, 12:25:14 UTC
6173153 Moritz Muehlenhoff 28 August 2015, 09:35:33 UTC Enable ferm on swift/esams Change-Id: Ie86f4e7604ee55262ac073c668f87f2099f30314 28 August 2015, 09:35:33 UTC
082d3db Moritz Muehlenhoff 28 August 2015, 09:16:31 UTC also enable ferm on strontium (also move it to the role now) Change-Id: I1efc8f8cf2300442f58a5e539ef7cd26b41e1edc 28 August 2015, 09:16:31 UTC
270f3df Moritz Muehlenhoff 28 August 2015, 08:52:18 UTC Enable ferm on rhodium Change-Id: I26fbdd04959a203103862e1f87a72ad84433cee6 28 August 2015, 08:52:18 UTC
1a8a993 Tim Landscheidt 27 August 2015, 21:31:21 UTC gridengine: Fix status check for gridengine-exec Even though I am *very* certain that I tested both on toolsbeta-exec-101 and toolsbeta-exec-202 *exactly* that indeed the gridengine-exec service was identified as running when there was an sge_execd process, somewhere down the line reality shifted and now Puppet is behaving as documented at https://docs.puppetlabs.com/references/latest/type.html#service-attribute-hasstatus: | [...] | If a service's init script does not support any kind of status | command, you should set hasstatus to false and either provide a | specific command using the status attribute or expect that Puppet will | look for the service name in the process table. [...] So this change sets hasstatus to false to force Puppet to use the pattern attribute. Bug: T110532 Change-Id: I1244c55c77e605d3e98d0436ee7923a156223a93 28 August 2015, 07:54:22 UTC
384247c Moritz Muehlenhoff 27 August 2015, 07:58:25 UTC Exempt mediawiki/http from connection tracking Change-Id: I727f7f9d9571ffc138c33a77e56c0a705268893c 28 August 2015, 05:31:03 UTC
7b25d22 Bryan Davis 25 August 2015, 22:03:09 UTC Logstash: make sure all input defines deal with ferm * Add ferm::service for logstash::input::tcp * Add note about ferm::service for logstash::input::irc None of these inputs are used in WMF production today, but they are used in beta cluster and/or labs projects. Having the ferm::service ready/noted may make debugging things later easier. Change-Id: I7f29e3b736af7833b77a3512a610f13de1dd01e0 28 August 2015, 05:23:39 UTC
936cd3d Ori Livneh 28 August 2015, 00:51:08 UTC UPING is not a word. Fixes a bad regex in beautify_message() that causes it to condense 'UP: PING OK' to 'UPING OK'. Change-Id: I5b0acd9a7717bc0386c2c4c1771e1d0d80f35774 28 August 2015, 00:51:48 UTC
b906a5b Ori Livneh 28 August 2015, 00:36:43 UTC Small improvements to ssh-agent-proxy Change-Id: I97dd5d803996e930bf8ef118b77ce8b4942f5b6f 28 August 2015, 00:37:38 UTC
b1b2ed5 dzahn 28 August 2015, 00:10:16 UTC mailman: use role-based hiera lookup on fermium We need to use the special "role" keyword here not a regular include or role-based hiera lookup won't work and we are getting "Could not find data item mailman::lists_servername". Also compare to sodium. Bug:T109925 Change-Id: Iad3d8901168100d8a0ac5e6892d79700c04c7a1e 28 August 2015, 00:12:17 UTC
ab7bdfe dzahn 25 August 2015, 23:03:37 UTC mailman: apply list role on fermium Bug:T109925 Change-Id: Icee8d71cc6bbc0cf70409ebec4db5d86bd364d65 28 August 2015, 00:03:29 UTC
6f65e22 YuviPanda 27 August 2015, 23:32:59 UTC k8s: Remove unnecessary etcd dependency Was making the unit fail Change-Id: Ia40295d7576d8e15157e95ff8d9ef41b92aae962 27 August 2015, 23:38:01 UTC
a125954 YuviPanda 27 August 2015, 23:12:14 UTC k8s: Increase verbosity of logging for kube-proxy Change-Id: I3a6a4e2576017a81214e04959e632a283907c9f4 27 August 2015, 23:37:31 UTC
aaf8c18 dzahn 27 August 2015, 16:57:10 UTC exim: temp hack to stop exim when on fermium Add an ugly but temp. hack to ensure exim is NOT running when on fermium but leave it unchanged on all other hosts (sodium). This is to be able to apply the puppet role and test it for the migration without having exim send out any emails. Bug:T109925 Change-Id: I1e336990d6e836c8b3a4c5da3530eabfebbf57f5 27 August 2015, 23:18:22 UTC
a49c92d YuviPanda 27 August 2015, 23:11:35 UTC k8s: Don't turn on all the admit controllers We're still playing around with some of those Change-Id: I7d3315a6baaaa55d4dd7da246a2e49a77b5c3539 27 August 2015, 23:12:46 UTC
b94a101 cpettet 27 August 2015, 19:50:58 UTC elasticsearch: ferm for 18-23 Change-Id: Ia99bd9f4e09e76ea5f2894730b9d0d53a062be53 27 August 2015, 20:07:42 UTC
0608f9e cpettet 27 August 2015, 17:57:20 UTC elasticseach: ferm for 27-31 Change-Id: Ia6f06903d048a4ffd0f56f7317433e217920176c 27 August 2015, 18:00:39 UTC
e6652e7 dzahn 27 August 2015, 17:10:21 UTC fermium: add mapped IPv6 address Give fermium a mapped IPv6 address, just like sodium has. Change-Id: Ia2c9c48dc0bf878fcf8b6a42e064805d1bee62c2 27 August 2015, 17:34:57 UTC
bb9d993 milimetric 27 August 2015, 16:10:25 UTC Make statsd host configurable via hiera Bug: T110462 Change-Id: I68c7546e8298646abad844a26c341176f75eb8d6 27 August 2015, 16:48:30 UTC
b9f6673 Moritz Muehlenhoff 27 August 2015, 16:39:23 UTC Add ferm rule for grafana traffic Change-Id: I113d91d74960725d153fbbc27bc56a83e957dbb1 27 August 2015, 16:47:53 UTC
aa3c803 Alex Monk 26 August 2015, 13:55:35 UTC Add affcom wiki domain to apache config For renaming from chapcom Bug: T41482 Change-Id: I4112272181ea39c89820e88e1e13bc69b0c9f04d 27 August 2015, 16:35:05 UTC
75ff4f6 Alex Monk 26 August 2015, 22:39:43 UTC Remove other *.wikimedia.org stuff from beta apache config Change-Id: I2cc1f2596b9c0bde90ebe8919ca7b82a5ae8dc2c 27 August 2015, 16:24:11 UTC
57860c5 Alex Monk 26 August 2015, 22:36:08 UTC Remove apache vhost for *.labs.wikimedia.org from beta Beta cannot actually run *.wikimedia.org things... Change-Id: I0ef9eb2eb1a0df16986489328b827e0bf44f5695 27 August 2015, 16:23:36 UTC
9ce8caf Alex Monk 26 August 2015, 22:31:44 UTC Remove apache vhost for non-existent beta incubator wiki Change-Id: I56cedcec7ba5bea14949b192d8536e1bdd21ad0e 27 August 2015, 16:23:17 UTC
a6c95bb Yuri Astrakhan 27 August 2015, 01:45:03 UTC Relax REFERER restrictions for Maps cluster Per meeting between bblack, akosiaris, maxsem and yurik, relaxing referrer restrictions to include wikivoyage and all of labs. Bug: T109159 Change-Id: Iaaf2701a7d48bc236c32af083869c9fd98349721 27 August 2015, 16:13:10 UTC
f2aea62 RobH 27 August 2015, 16:05:43 UTC policy.wikimdia.org cert even though this is used by wordpress for the policy site, we need to keep a copy of it in our repo T110197 Change-Id: I307b2e258b75db058abb4e022ca04bdfcef4d458 27 August 2015, 16:08:31 UTC
289a1cf Moritz Muehlenhoff 27 August 2015, 15:32:35 UTC Enable ferm on logstash100[1-3] Change-Id: I150432c5c7efa8b95d350f371f530f2165bda899 27 August 2015, 15:32:35 UTC
c1d01be Andrew Otto 27 August 2015, 15:09:31 UTC Increase num.replica.fetchers to 12 across the Kafka cluster We were using 12 during the recent partition move, and things were working fine. With a lower number, some brokers start lagging during Camus runs. Change-Id: Ic1ca42f57c6700f1542f5cb8589832c16e8ec7a6 27 August 2015, 15:10:51 UTC
a74d480 Andrew Otto 27 August 2015, 14:23:20 UTC Increasing num.replica.fetchers all around the cluster Change-Id: I01b607d45383aa649764533630c30df2ff713a9c 27 August 2015, 14:32:58 UTC
10110d5 Moritz Muehlenhoff 27 August 2015, 14:18:00 UTC Enable base::firewall on eqiad storage servers Change-Id: I6c2b0b43cf2aafbf38b4c903e847bbface87e96e 27 August 2015, 14:19:21 UTC
5a55023 Filippo Giunchedi 27 August 2015, 13:27:49 UTC swift: enable base::firewall on ms-fe2* Change-Id: I30c38c2e50bcea53f05ed4c9f7ba1404adafe578 27 August 2015, 13:33:52 UTC
06af077 Brandon Black 27 August 2015, 13:27:28 UTC Require ethtool package for ethtool execs This caused some exec failures during initial puppetization of lvs2006 as jessie. Change-Id: Ic04e42c20fc44ddd5322c0d7ba8c77b71dc1e69e 27 August 2015, 13:28:10 UTC
f3c1c23 Antoine Musso 27 August 2015, 09:47:47 UTC gdash: fix parser cache metrics The MediaWiki parser cache related metrics have been moved up: - MediaWiki.stats.pcache_* + MediaWiki.pcache_* Fix https://gdash.wikimedia.org/dashboards/pcache/ Change-Id: Ieecc644e0d27b8ff762343131452db232ad1ff45 27 August 2015, 13:23:07 UTC
back to top