ad2194f | Guillaume Lederrey | 15 February 2017, 10:12:13 UTC | elasticsearch: adding new servers elastic1048-1052 Bug: T155790 Change-Id: I79d994e7f6ad2460bb33c082f80538eb33312add | 15 February 2017, 10:12:13 UTC |
0d293fa | elukey | 15 February 2017, 09:46:20 UTC | Update the zookeeper module Change-Id: I08d78e5f7d852922daec36b2c7808b9f1aa19897 Ref: https://gerrit.wikimedia.org/r/#/c/337413/ Bug: T157968 | 15 February 2017, 09:46:20 UTC |
b946b29 | Guillaume Lederrey | 13 February 2017, 10:19:51 UTC | elasticsearch - reimage to jessie and move data to /srv - preliminary work This move data directory configuration to host specific hiera files (those will be deleted priori to reimaging the hosts). elastic1* nodes are configured to install on Jessie. elastic1017-1031 have smaller disks, and are configure for RAID0 partitions, the previous elasticseach-raid0.cfg partman config has been changed to mount a data partition on /srv elastic1032-1052 are now using the raid1-lvm-ext4-srv.cfg parman config. The elasticsearch-raid1.cfg partman config isn't used anymore and has been removed. Bug: T151326 Bug: T151328 Change-Id: I76d0d1aeba02b1a3a8466ec7c912abb686041d95 | 15 February 2017, 09:05:34 UTC |
d78b68c | YuviPanda | 08 February 2017, 05:23:45 UTC | tools: Upgrade docker on tools k8s workers Bug: T157180 Change-Id: I00dc34bcc3c4d54d89d7fbe2c618fce91ade183e | 15 February 2017, 04:47:56 UTC |
acf1ec6 | Daniel Zahn | 11 February 2017, 01:09:37 UTC | lint: 'include standard' -> 'include ::standard' Change-Id: If0a848c683b7b149d86a2afbc67694d7d989b376 | 14 February 2017, 23:35:09 UTC |
8357b7b | Daniel Zahn | 13 February 2017, 22:13:22 UTC | aptrepo: rsync the entire /srv/ automatically, not just /srv/wikimedia/ So far we have setup 2 seperate rsync modules, one for /srv/wikimedia/ and one for the entire /srv/, so manually you can sync either of them but we have only automated the /srv/wikimedia/ part. The latter is just the data for apt.wikimedia.org, the extras that are in /srv/ are things like "firmware", "junos" and "megacli". And "tftpdata" which gets provided by puppet though. Let's just sync the whole /srv, it will make sure that install1002 and install2002 are the same also for the extra data. Change-Id: I4e4d2ccfb494046bc61a7c5661c49596812a4595 | 14 February 2017, 23:06:45 UTC |
49a7ad2 | juniorsys | 26 January 2017, 10:40:28 UTC | quarry: Linting changes Use full names for class names, as relative names are not allowed in future Puppet versions Add trailing commas to abide by the Coding Style guidelines Bug: T93645 Change-Id: Ie0cb6a9680c05c76ade5691d3e91e2d24de6126f | 14 February 2017, 23:02:08 UTC |
3754de0 | Daniel Zahn | 11 February 2017, 01:06:04 UTC | lint: 'include base::firewall' -> 'include ::base::firewall' Change-Id: I38d8459aa27450dc7934de7a9b1823e5328293be | 14 February 2017, 22:22:01 UTC |
81ac2da | Madhumitha Viswanathan | 14 February 2017, 22:11:21 UTC | labstore: Remove misplaced init in DirectorySizeCollector Change-Id: I8903148effe637436268703deadf41167a668e51 | 14 February 2017, 22:11:21 UTC |
06cfc91 | Madhumitha Viswanathan | 14 February 2017, 21:41:56 UTC | labstore: Fix sudo priveleges for user diamond Change-Id: Ide2f16d6a5274bb085d8c93810c51e61bad3b97e | 14 February 2017, 21:41:56 UTC |
08066cd | Madhumitha Viswanathan | 03 February 2017, 19:39:02 UTC | labstore: Diamond collector to track directory sizes Bug: T126623 Change-Id: I41c68ae468f0de3d5c368344f189071226339a24 | 14 February 2017, 21:10:43 UTC |
b82846b | Andrew Otto | 14 February 2017, 20:38:03 UTC | Update published-datasets-readme.txt Change-Id: Iee8963a71428c231797df31856672d2594ca62ef | 14 February 2017, 20:40:43 UTC |
8a386e4 | Andrew Otto | 14 February 2017, 20:35:47 UTC | Use --partition-type hive for refinery-drop-wdqs-extract-partitions job Bug: T146915 Change-Id: Iff8fe78bf438c5e8e4dad752948073a53d75c9b6 | 14 February 2017, 20:35:47 UTC |
ba84760 | Nathaniel Schaaf | 01 February 2017, 12:31:57 UTC | Drop wdqs_extract partitions older than 90 days Amended to run once per day Bug: T146915 Change-Id: I5d29490c0e8e0314a131d12a758bdfb3d2d8735f | 14 February 2017, 19:52:23 UTC |
bbdc556 | Paladox | 14 February 2017, 17:45:04 UTC | Gerrit: Converts ChangeSubject Velocity template into soy template In gerrit 2.14 velocity templates will be deprecated and replaced with soy templates. Bug: T158008 Change-Id: Ia58898c7f18b7cdb516898d6a1849f3b256bb192 | 14 February 2017, 19:42:42 UTC |
aaca189 | andrewbogott | 14 February 2017, 18:37:29 UTC | Horizon: add explicit "!" policies for unsupport services. Among other things, this should prevent the 'Admin' panel from appearing for non-admins. Bug: T158099 Change-Id: Ie56c578b66c1cdb7269aecc3b347b41e22ac6c68 | 14 February 2017, 19:01:37 UTC |
fabe075 | andrewbogott | 14 February 2017, 18:42:46 UTC | Horizon: Backport a newton fix to Mitaka This should fix https://bugs.launchpad.net/horizon/+bug/1653792 Bug: T158099 Change-Id: I5f467194a9305bed673cecc0c740d7d6f3b42760 | 14 February 2017, 19:01:22 UTC |
21a159c | Eric Evans | 13 February 2017, 22:03:40 UTC | Enable Prometheus exporter on restbase1007 (canary) Bug: T155120 Change-Id: I3479ca331f3f97bcab7e43496d89c1730df235a0 | 14 February 2017, 18:51:02 UTC |
2dc6120 | RobH | 14 February 2017, 18:01:00 UTC | adding info to nithum's shell account I neglected to include new expiry parameters in the shell access, so I'm adding it in after the fact. Bug:T157724 Change-Id: I1dd7d33bcb9202889e9b108de8170954c9d4dba3 | 14 February 2017, 18:01:00 UTC |
9305868 | RobH | 13 February 2017, 19:33:38 UTC | new shell user Nithum Thain Adding new user nithum to shell access and analytics-privatedata-users This change should only be merged by an opsen after reviewing the task and ensuring no objections after the three day wait has ended. This wait ends on 2017-02-14. Bug:T157724 Change-Id: Ie52aad113523fbb20170f66db2689f637b07950c | 14 February 2017, 16:56:58 UTC |
7c5e987 | andrewbogott | 14 February 2017, 14:47:01 UTC | Horizon: Upgrade to mitaka Change-Id: Iedf3497fa8b0d0b59a8b26d604c0c00b9ed19916 | 14 February 2017, 15:59:45 UTC |
e389e04 | Filippo Giunchedi | 02 February 2017, 10:38:25 UTC | scap: move udp2log from fluorine to mwlog1001 Bug: T123728 Change-Id: I67701eba2686b8836b3c43a56bdc4d440a6a8d69 | 14 February 2017, 15:26:27 UTC |
0655216 | Filippo Giunchedi | 02 February 2017, 10:49:58 UTC | udp2log: mirror traffic from mwlog1001 to fluorine Introduce $mirror_destinations to instruct an udp2log host to mirror traffic to other hosts. Bug: T123728 Change-Id: I08da3d975aa901a750a96c34a3814cf02b1c12d2 | 14 February 2017, 15:17:52 UTC |
3ce61fd | Manuel Arostegui | 20 December 2016, 12:02:25 UTC | Reporting tests with the private data script * For now just run the private data script and email me once something is found so it can be polished. * Scheduled to run once per week now: every Monday. Ideally it should be an icinga check eventually. Bug: T153680 Change-Id: I7796d6860f70c34b1758655f18a4ed8196724e97 | 14 February 2017, 14:56:21 UTC |
dc32ba5 | elukey | 14 February 2017, 14:26:15 UTC | Move mw224[45] from appservers to imagescalers Bug: T156023 Change-Id: Id8da627d8762a28581b5d45fefc8f65389c39916 | 14 February 2017, 14:27:27 UTC |
e816281 | Moritz Muehlenhoff | 14 February 2017, 13:59:50 UTC | Remove access credentials for bcohn Bug: T158051 Change-Id: I7e6fdc6b62996cda9ea3c5e5c5b25d0bfe2d6c0a | 14 February 2017, 13:59:50 UTC |
4255ef2 | Moritz Muehlenhoff | 14 February 2017, 13:43:17 UTC | Record extented account expiry date for nettrom Change-Id: I51cef14953894c7a6090e900495221aa2ed5a25a | 14 February 2017, 13:43:45 UTC |
6df5a59 | elukey | 14 February 2017, 12:55:17 UTC | Fix and tune the new Analytics Hadoop alarms Bug: T88640 Change-Id: I1e47c128ca04dc48690ecbd5d70fa7ee154b7423 | 14 February 2017, 13:01:39 UTC |
aaa05b1 | Filippo Giunchedi | 13 February 2017, 12:33:22 UTC | install_server: fix graphite partman recipe Switch to ext4 and fix root filesystem size to be 50GB (25GB per partition in raid10) Change-Id: Ica8b4404bd5cd1f67031b9b84be6b2b32f3e5ec2 | 14 February 2017, 11:57:34 UTC |
2dcac1a | elukey | 14 February 2017, 11:50:23 UTC | Move mw222[123] from appservers to api_appservers (conftool) Bug: T156023 Change-Id: I3304de1913aad0316a25501d49d7d8fdabb5a676 | 14 February 2017, 11:50:23 UTC |
5601b62 | elukey | 14 February 2017, 11:34:02 UTC | Change role to mw222[123] (appservers -> api_appservers) Bug: T156023 Change-Id: I0898d4173b351c353cab3b468a9802e2037cf1c8 | 14 February 2017, 11:34:02 UTC |
90986c5 | Moritz Muehlenhoff | 13 February 2017, 14:45:11 UTC | Only run the timesynd_ntp_status Icinga check every 30 minutes Only run the timesynd_ntp_status every 30 minutes (similar to the Icinga check for ntpd). Even if timesyncd would have crashed in the mean time, the clocks are not going out of syncronisation during the interval anyway. Also, timedated is a socket-activated daemon and any invocation of timedatectl (as done by the Icinga checks) logs "Starting Time & Date Service" and "Stopping Time & Date Service", so this reduces log spam as well. Bug: 157798 Change-Id: I47535c3d140521a0d33974cba6b6d7ffc6ae59aa | 14 February 2017, 10:57:13 UTC |
b6c18a1 | Volans | 14 February 2017, 10:32:34 UTC | Revert "Testreduce: allow to decide the state of the services" This reverts commit 34141b66e3fe70ad347dceb5732773a70ab02051. Bug: T156177 Change-Id: I9ba4310dd79dec5392f5f5fdbc3884b95378eb90 | 14 February 2017, 10:52:36 UTC |
7a1b26e | Elukey | 03 February 2017, 12:32:07 UTC | Revert "Revert "Add JVM Heap usage alarms for basic Hadoop daemons"" This change was rolled back to establish if it was causing an issue to graphite1001. This reverts commit 71607cfff7b770939adbdd7a7dbee4bbafe49c76. Change-Id: I9a44472d2a3a42fb769fe607334f4052a30b7112 | 14 February 2017, 10:48:06 UTC |
a60da1d | Riccardo Coccioli | 14 February 2017, 10:40:06 UTC | Testreduce: renamed environmental variable Bug: T156177 Change-Id: I973a3901b070fb09e03fa1e8cce0b7c9ea3b1820 | 14 February 2017, 10:40:06 UTC |
0cea11f | Riccardo Coccioli | 14 February 2017, 10:23:12 UTC | Testreduce: use address instead of IP for web proxy Bug: T156177 Change-Id: I64babd343efc761913582601423ce22046dca09d | 14 February 2017, 10:23:12 UTC |
6f9c214 | Filippo Giunchedi | 09 February 2017, 14:53:41 UTC | diamond: require $handler to be defined AFAICS the check was introduced in https://gerrit.wikimedia.org/r/#/c/144718 but nowadays diamond is enabled everywhere in labs too, therefore handler is always defined. Also reload diamond on handler config changes. Bug: T157022 Change-Id: I6a10871d304ba1fc942896d3d88aa714f27874cc | 14 February 2017, 10:15:14 UTC |
1acb63f | Emanuele Rocca | 14 February 2017, 09:20:50 UTC | Analytics VCL: default to 'org' if top_domain is not set Bug: T138027 Change-Id: Ia66270c54363634f43ac663789d0919e868bf5fe | 14 February 2017, 09:20:50 UTC |
d8f2b63 | Emanuele Rocca | 09 February 2017, 13:29:16 UTC | VCL: Add support for WMF-Last-Access-Global analytics cookie Bug: T138027 Change-Id: Ib2eacbd0479462c894c214d663b11143586edd50 | 14 February 2017, 09:13:20 UTC |
b0770d1 | Daniel Zahn | 14 February 2017, 00:24:49 UTC | joe: move hosts file for carbon to install1002 carbon has been replaced by install1002 Change-Id: I160a5ebcdc873ab81b5b1fb0019c22e43f2afc30 | 14 February 2017, 07:42:10 UTC |
5c04d25 | Giuseppe Lavagetto | 06 February 2017, 15:57:46 UTC | stdlib: upgrade to 4.15.0 Change-Id: I11b9042f323fe3f1c2c9950c808ab16551d9d46f | 14 February 2017, 07:28:29 UTC |
e38eb5a | Daniel Zahn | 14 February 2017, 00:45:39 UTC | install: correct spare role name for carbon It's role::spare::system, not just role::spare. Change-Id: I46416bab212db270a8dfdb078322a38861f7f51b | 14 February 2017, 00:46:55 UTC |
797dd0d | Daniel Zahn | 11 February 2017, 00:14:35 UTC | install: remove roles from carbon, demote to spare carbon will be decom'ed in about a week but until then all roles should be removed and it should be using role::spare per the decom steps from server lifecycle. Bug: T158020 Change-Id: Ib58bd8ad52a2047e488aca64672f445daf62dc5b | 14 February 2017, 00:36:06 UTC |
88e4813 | Daniel Zahn | 11 February 2017, 00:16:59 UTC | let install1002 be the new source for APT data rsync Start syncing /srv APT data from install1002 to install2002 instead of from carbon to install1002. Bug: T132757 Change-Id: I513e1cd1a2cd381013675304e25df733ae780829 | 13 February 2017, 20:47:08 UTC |
1321bdf | andrewbogott | 09 February 2017, 12:30:35 UTC | Remove openstack::clientlib from icinga hosts This was used for the currently-in-limbo Keystone role tests. If/when those tests are revived they need to run an labcontrol rather than on the icinga host. Bug: T157760 Change-Id: I487453dcbed9e3b651e0704dc6c5ed06370da746 | 09 February 2017, 12:33:32 UTC |
5ebb4cd | RobH | 13 February 2017, 20:14:14 UTC | correct samtar's stat1003 access this patch corrects some initial confusion in the access request, changing access from statistics-users to the researchers group on stat1003 Bug:T157483 Change-Id: I8e9ba9ece7842ba9448000e9b6c2c1ee6a5b5c31 | 13 February 2017, 20:35:18 UTC |
d861f29 | Daniel Zahn | 13 February 2017, 20:01:35 UTC | install: enable Letsencrypt on install1002 After the apt.wikimedia.org CNAME switched from carbon to install1002, we need to enable Letsencrypt cert creation on install1002 to make the cert for "apt" work here. Bug: T132757 Change-Id: I4831d4e9d243af3c83333afa2d2f6d35cbfd0c8d | 13 February 2017, 20:09:31 UTC |
96c8899 | Daniel Zahn | 10 February 2017, 16:21:47 UTC | CI: decom scandium remove scandium from puppet, install_server, Hiera Bug: T150936 Change-Id: I8c3e877c2296617019dc8fbacfb6a6135c43c4f3 | 13 February 2017, 16:39:29 UTC |
fccf34e | Guillaume Lederrey | 06 February 2017, 14:10:46 UTC | WDQS - move metric collection to diamond Moving metrics to diamond also changes the path of the "lag" metric. Graphite check has to be updated to reflect that change. Bug: T146468 Change-Id: Iaaf06519d25e7b7941e9fb3b5693178514bd9c51 | 13 February 2017, 16:31:53 UTC |
9eaa8de | Faidon Liambotis | 11 February 2017, 02:00:17 UTC | Remove jzerebecki from Icinga contact groups After this is merged, the contact should be removed from puppet-private too. Change-Id: I0591eb1d67f903245f3130a9a87f3a778ed7085d | 13 February 2017, 15:59:01 UTC |
f389a32 | Guillaume Lederrey | 02 February 2017, 14:02:41 UTC | WDQS - move metric collection to diamond Lag and number of tuples are at the moment published via a custom PHP script and not via diamond. Moving everything to diamond ensures coherence and automatic configuration of new nodes. This is a direct port of the PHP implementation available at https://phabricator.wikimedia.org/diffusion/ADES/browse/production/src/wikidata/sparql/minutely.php Bug: T146468 Change-Id: I802bc5bc5324c052137f99eb8fdd1fee1b57e3b2 | 13 February 2017, 14:54:19 UTC |
b1bfa59 | Filippo Giunchedi | 13 February 2017, 12:20:42 UTC | coal: run on jessie Add systemd unit for coal to run on jessie. Also initialize /srv/org/wikimedia as expected by coal, thus fixing puppet runs. Bug: T157022 Change-Id: Iccbd7c23a7ffe5823806e256eca9cbca1bc1852f | 13 February 2017, 14:34:34 UTC |
e151cfc | Moritz Muehlenhoff | 13 February 2017, 12:38:04 UTC | Add debdeploy salt grains for new dbmonitor hosts Change-Id: I5d21a405ec969b5300c1bae4be3ae4fdd0ee9695 | 13 February 2017, 12:51:53 UTC |
799e54c | Filippo Giunchedi | 13 February 2017, 11:41:43 UTC | hieradata: temporarily remove prometheus100[34] from prometheus_hosts When Iab83f351fb was merged the AAAAs for prometheus100[34] were not in place yet. This meant that reloading ferm would fail due to @resolve failing and the whole puppet run fail as a result. Note though that the next puppet run would succeed because 'ferm reload' won't be issued. Thus force a successful 'ferm reload' with this change, to be rolled back once AAAAs are available. Bug: T152504 Change-Id: Icf131539af64947ddf98cf309a6150eb88e0820c | 13 February 2017, 11:41:45 UTC |
7ef4f0a | Guillaume Lederrey | 06 February 2017, 15:22:43 UTC | wdqs1002 - move data to /srv/wdqs to follow the usual partitioning scheme cleaning up the now unused lvm-wdqs.cfg partman recipe Bug: T144536 Change-Id: I0d6fead29549ee0f3098d2a30b40833f8ba816e9 | 13 February 2017, 11:11:43 UTC |
29d2cac | Daniel Zahn | 07 February 2017, 01:49:07 UTC | add prometheus1003/1004 to site.pp Bug: T152504 Change-Id: Iab83f351fbc90bca2904cfe8f66ac36cdad58c92 | 13 February 2017, 11:02:38 UTC |
50efe0e | Moritz Muehlenhoff | 13 February 2017, 10:41:37 UTC | Update SSH for Sam Tarling Previous commit had a linebreak error in the key. Bug: T157483 Change-Id: Ib97a18e7008d5df4124b2e4a7c8a057b7b5ff1d7 | 13 February 2017, 10:41:37 UTC |
e94b69c | Emanuele Rocca | 10 February 2017, 13:06:52 UTC | varnish: remove ganglia vhtcpd python module Change-Id: I0b67cef2f345ad10a5c8cc5774a32f5920dd9d3b | 13 February 2017, 10:14:11 UTC |
c0c6dc2 | Guillaume Lederrey | 11 February 2017, 09:12:34 UTC | elasticsearch - reimage elastic20(33|34|35|36) to jessie and move data to /srv Bug: T151326 Bug: T151328 Change-Id: I2b2d00f6a7cf4d7e01f4f57cd7e11f46552a8044 | 11 February 2017, 09:12:34 UTC |
cde1a15 | Faidon Liambotis | 10 February 2017, 22:49:12 UTC | salt: use SHA256 master key fingeprint on newer systems stretch's salt-minion expects master_finger to be a SHA256 fingerprint rather than an MD5 one. While it's possible to change that with the hash_type argument, MD5 is cryptographically obsolete and shouldn't be relied on, so start using a SHA256 fingerprint instead. Change-Id: Id6315e6ca37234e4bd3c4728b25d35830ef94193 | 11 February 2017, 01:13:22 UTC |
bd464a0 | Daniel Zahn | 11 February 2017, 00:24:24 UTC | delete install1001/2001 from Hiera data Bug: T157840 Change-Id: I842c68783bd7beb431dbc23e93776ef75150b2cb | 11 February 2017, 00:50:35 UTC |
ad932f1 | Chad Horohoe | 10 February 2017, 23:36:27 UTC | Gerrit: Stop stuffing so many cache things into memory The disk cache is fast enough, plus is survives restarts Change-Id: Iee02c1d2049421c7c9da108ec4c450d484425ed9 | 10 February 2017, 23:36:27 UTC |
2b58b1b | Faidon Liambotis | 10 February 2017, 22:06:46 UTC | salt: add missing import to grain-ensure.py grain-ensure uses salt.minion.SMinion but doesn't actually import salt.minion. This is currently broken with at least salt 2016.11.1, as found in stretch. Change-Id: I985645bf559f6d674298480529f14231123a68f7 | 10 February 2017, 22:11:30 UTC |
13272de | andrewbogott | 09 February 2017, 02:33:57 UTC | Toollabs: Remove zsh from package list zsh is now included in Standard, which is already present on all toollabs nodes. Change-Id: I9315cec33af0eca84fe1c336d97df62df167b90e | 09 February 2017, 02:36:39 UTC |
2913b7c | Faidon Liambotis | 10 February 2017, 21:45:45 UTC | autoinstall: also pass net.ifnames=0 to the end system Otherwise the newly-installed system gets brought up without working network configuration. Change-Id: I04cf599e32a5cfe89827c4840e4ea8e926ec8bdc | 10 February 2017, 21:46:37 UTC |
29dceee | Faidon Liambotis | 10 February 2017, 21:42:57 UTC | Replace 'zsh-beta' with 'zsh' zsh-beta was a transitional package (in trusty/jessie) depending on zsh that is non-existent in stretch onwards. Replace it with zsh but don't add a zsh-beta removal as to not affect existing precise installs. Change-Id: I4209290edb88be684cd6012590bcf9f426480a5d | 10 February 2017, 21:46:37 UTC |
d568472 | Daniel Zahn | 10 February 2017, 20:54:29 UTC | remove install1001/install2001 from site.pp Bug: T84380 Bug: T132757 Change-Id: I6d839a03f05ee83f2ab21579fac0790ad077a40b | 10 February 2017, 21:07:57 UTC |
e301f4a | Faidon Liambotis | 10 February 2017, 21:02:40 UTC | aptrepo: add new RSA 4096 apt key Add a new RSA 4096 key for use by apt with a fingerprint of: B8A2 DF05 748F 9D52 4A3A 2ADE 9D39 2D3F FADF Use it initially just for stretch-wikimedia, as a) the machinery to use it in already installed systems is not there yet and b) stretch actually requires a stronger key. Change-Id: I1dc483fdc6c7b3bc374b6005937128c52698dd41 | 10 February 2017, 21:05:00 UTC |
cc991e9 | Daniel Zahn | 10 February 2017, 04:18:22 UTC | install/DHCP/TFTP: use install1002 and install2002 as next-servers - replace install1001 with install1002 - replace install2001 with install2002 - remove carbon Bug: T84380 Bug: T132757 Change-Id: I9f66945f045fed3ee72adbb17f32f7044a6501df | 10 February 2017, 21:02:06 UTC |
85f7dd2 | Andrew Otto | 10 February 2017, 20:46:29 UTC | Include geoip on refinery hosts This will put geoip on stat1004 and analytics1027 Change-Id: I319e6d9c3580aef4934060aacd79f99b2557704f | 10 February 2017, 20:46:29 UTC |
ee759a8 | Filippo Giunchedi | 10 February 2017, 20:24:25 UTC | install_server: reinstall graphite1001 with jessie Bug: T157022 Change-Id: Id920560b339b8b7e109230632a14f4b08628c3ac | 10 February 2017, 20:25:59 UTC |
bb7ac54 | Faidon Liambotis | 10 February 2017, 20:04:52 UTC | autoinstall: pass net.ifnames=0 to stretch d-i Turn off predictable network interface names, that are the default for new installs as of Debian stretch. They are not as predictable as they claim to be (e.g. our d-i scripts check for the "eth0", while the equivalent new name would not be predictable). They are not necessarily a bad idea, but they diverge significantly from our current setup and assumptions all over our tree (d-i and puppet, not to mention muscle memory). We can think about enabling them at a another point in time, separate from the stretch upgrades. Change-Id: I973eb1cde95286595fe1fd1b515ec479ce2551d5 | 10 February 2017, 20:17:43 UTC |
bffddff | Faidon Liambotis | 10 February 2017, 20:01:41 UTC | autoinstall: add virtual.cfg to d-i-test Clearly a VM, with /dev/vda for its root device. Change-Id: Ib7829b2985a2de2e4d8b1c77921ef2d9afe3d127 | 10 February 2017, 20:17:43 UTC |
b0ad2f2 | cmjohnson | 10 February 2017, 19:43:11 UTC | Adding elastic1048-1052 to dhcpd Change-Id: I95c469af1887f77ce79121ab74ac8079e9e09099 | 10 February 2017, 19:51:24 UTC |
54c0519 | Faidon Liambotis | 10 February 2017, 19:27:46 UTC | autoinstall: switch d-i-test to stretch ...and to the flat.cfg partman recipe. Change-Id: I59d052d5e09d0073298bf19a77b585fc56e3633c | 10 February 2017, 19:28:06 UTC |
bea065f | Faidon Liambotis | 10 February 2017, 19:18:22 UTC | autoinstall: add stretch Change-Id: Ib8cce2cd3021fddeed2e6b9ade7102b4723dbaed | 10 February 2017, 19:19:48 UTC |
19e7611 | Faidon Liambotis | 06 February 2017, 17:34:46 UTC | aptrepo: add suite stretch-wikimedia Change-Id: Id354f780cdae9bfc53ee85d2ab2b2515d2b07c7b | 10 February 2017, 19:19:47 UTC |
3c4b69a | Guillaume Lederrey | 10 February 2017, 18:19:15 UTC | elasticsearch - reimage elastic20(29|30|31|32) to jessie and move data to /srv Bug: T151326 Bug: T151328 Change-Id: I6cb3a28fec14704d8981ddad9f63eed92806fa76 | 10 February 2017, 18:19:15 UTC |
c314407 | Amir Sarabadani | 02 February 2017, 18:42:40 UTC | dumps: More UI cleanup Bug: T155697 Change-Id: I91a324a623c7466236035104aa21ac34ceaaa58d | 10 February 2017, 17:47:32 UTC |
e53bad4 | RobH | 09 February 2017, 19:43:00 UTC | Sam Tarling shell access + statistics-users granting sam tarling shell access alogn with access to the user group statistics-users. Please note this should be merged by ops clinic duty, AFTER the 3 day wait has expired and there are no objections noted on the phabricator task. Bug:T157483 Change-Id: I579c8a9abe359f6d1e6f1999200eb7558af74165 | 10 February 2017, 17:38:12 UTC |
ba5b323 | Antoine Musso | 10 February 2017, 14:34:25 UTC | Remove zuul-merger from scandium.eqiad.wmnet We have working zuul-merger process on contint1001 and contint2001, hence there is no more any need for scandium.eqiad.wmnet. Remove role::zuul::merger Will have to stop the daemon on it and refresh Icinga. Bug: T150936 Change-Id: I2a472ea915f3c67d4a3adc44c3aacd8a47516baf | 10 February 2017, 16:11:25 UTC |
c0f7677 | Eric Evans | 10 February 2017, 15:21:42 UTC | Fix broken path to Prometheus exporter config Bug: T155120 Change-Id: I62f512ebe6de395705d74e7b53cbb21b8f4e38a2 | 10 February 2017, 15:35:36 UTC |
141cb40 | Giuseppe Lavagetto | 09 February 2017, 17:32:57 UTC | prometheus::class_config: allow new selections for prometheus In some cases, we want to create a targets list for all the servers that include a certain class. This define allows to do that, and even allows querying by specific parameters of the class. Change-Id: I765f2dd0a97f7826e679d3da621777b64e7eda03 | 10 February 2017, 15:24:18 UTC |
465b11e | Jaime Crespo | 10 February 2017, 14:43:44 UTC | Deploy 3a09aee8dd90d8f to production (Introduce linters using rake) It also includes a3ded1b40909f9351610, make sure you deploy gerrit:331329 first (or both at the same time). Change-Id: Icc1911e583b63d1280df73d3a11175f74c498fd6 | 10 February 2017, 15:20:45 UTC |
807cedd | Jaime Crespo | 10 February 2017, 14:30:25 UTC | Apply a3ded1b40909f9351 (mariadb client install) to production Bug: T157702 Change-Id: I71ef814d7dae43d953e8a1d4d32139593f173ca6 | 10 February 2017, 15:16:33 UTC |
55cc42a | Eric Evans | 03 February 2017, 16:42:58 UTC | Enable JMX exporter on RESTBase Staging nodes in eqiad Bug: T155120 Change-Id: Ib8eac73657a01f13478e74bb7e656852a7615e85 | 10 February 2017, 15:04:21 UTC |
5a3a7f7 | Filippo Giunchedi | 03 February 2017, 08:33:56 UTC | graphite: move alerts to graphite2001 Move the actual icinga checks over to graphite2001. This is required because from icinga's POV the checks belong to graphite1001, therefore if the machine is down the checks are also down. The checks' puppet exported resources will need to be cleaned up manually on the puppet master for icinga to pick up the changes. Bug: T157022 Change-Id: Id8025d28ca6727f6840ee1f2f88d1245c3ed0ba5 | 10 February 2017, 14:35:52 UTC |
a9bfafd | Giuseppe Lavagetto | 10 February 2017, 13:00:46 UTC | profile::etcd::replication: attempt not to confuse icinga Apparently a dollar sign in the arguments of the command is harmful. here it was not giving us any real advantage, so just remove it. Change-Id: I9dbeee9bcf383c3ed2114b6cba8cf7f02c5a3819 | 10 February 2017, 13:01:08 UTC |
80cb119 | Moritz Muehlenhoff | 10 February 2017, 12:14:47 UTC | Record extended NDA/contract dates for CPS frtech consultants Change-Id: I1c3885ac878cc46caee1121a71eb36b778088742 | 10 February 2017, 12:14:47 UTC |
9417787 | Moritz Muehlenhoff | 10 February 2017, 12:09:41 UTC | Extended MOUs for ISI researchers Change-Id: I966ad3604ec1db884d2eb7a52144c550e1251a5f | 10 February 2017, 12:09:41 UTC |
b9074d8 | Giuseppe Lavagetto | 10 February 2017, 11:46:47 UTC | profile::etcd::replication: lag can be negative, quotes Change-Id: Iabf6bdc5a40bb7588fca21c762b721584eb5510b | 10 February 2017, 11:55:28 UTC |
803dcfc | Jaime Crespo | 10 February 2017, 11:32:03 UTC | admin-jynus: Update my alias to show static hostaname Otherwise, this shows localhost for \h- use hostname to run it once show it shows the hostname. Change-Id: Ieb5b8c7be06869d02986916ecd657fe364ce3a07 | 10 February 2017, 11:42:28 UTC |
35098e1 | Manuel Arostegui | 10 February 2017, 06:58:41 UTC | beta.my.cnf: Make beta mysql prompt like prod This change will make beta to show: mysql:user@host [database]> Just like we do in production Bug: T157714 Change-Id: I5578fda2913a5708cb5465ae165759a81c5eca1d | 10 February 2017, 11:03:03 UTC |
15758f6 | Ariel T. Glenn | 10 February 2017, 09:51:04 UTC | temp removal of wansec.com mirror from list, dns issues Change-Id: I002efc6b3b3b220e1f712bd6f8f00e537fb391d1 | 10 February 2017, 09:54:01 UTC |
7e1eee4 | Giuseppe Lavagetto | 10 February 2017, 09:37:00 UTC | profile::etcd::replication: fix monitoring port for icinga Change-Id: I8d5690deab2a0aca97ad1b7cdb463c4322be8472 | 10 February 2017, 09:37:00 UTC |
abb53c3 | Emanuele Rocca | 09 February 2017, 10:13:38 UTC | varnish: remove ganglia python module We are now using prometheus+graphana to plot graphs based on varnishstat: https://grafana.wikimedia.org/dashboard/db/varnish-machine-stats Remove the ganglia python module. Change-Id: I6b48216745a7c1a6e478e562ab1230072cf210fb | 10 February 2017, 09:01:55 UTC |
709de06 | Guillaume Lederrey | 10 February 2017, 08:42:04 UTC | elasticsearch - reimage elastic20(25|26|27|28) to jessie and move data to /srv Bug: T151326 Bug: T151328 Change-Id: Ib89e396b2e0abe135040590b245d94dc7a94de87 | 10 February 2017, 08:45:34 UTC |
28acc93 | Guillaume Lederrey | 09 February 2017, 19:47:52 UTC | wdqs - icinga process check more relaxed on arguments The actuall WAR name can change a bit over time. The classifier should not be part of the check. Change-Id: I458965b3f21a355ec7a98667aee71533da413f3a | 10 February 2017, 08:35:45 UTC |
09dc659 | Giuseppe Lavagetto | 09 February 2017, 11:06:36 UTC | profile::etcd::replication: refactor to make failover easier - Revert the etcdmirror::instance logic that was thought out to allow multi-source replication. We're not going to do that anytime soon and it made it impossible to replicate back the data to the original cluster easily in case of failover - Add monitoring and ferm rules - Install on all machines, just have the service active on one of them. Bug: T156009 Change-Id: Ie2a64ce9dcba6a3bf7dd85084c87836a8803dff3 | 10 February 2017, 08:16:14 UTC |
ddf3531 | Daniel Zahn | 10 February 2017, 04:38:13 UTC | zuul: add contint1001/2001 to zuul merger hosts for ferm change I1a66af3435bc67013 turned contint1001 and 2001 into zuul merger hosts. But this was not adjusted so ferm rules were not added and things broke with contint1001 refusing connection. https://integration.wikimedia.org/ci/job/operations-puppet-tox-jessie/13878/console 04:23:44 contint1001.wikimedia.org[0: 208.80.154.17]: errno=Connection refused Bug: T150936 Bug: T140297 Change-Id: I5a8ece9152bd237b4ee958c5386dbeebfe688b31 | 10 February 2017, 04:40:26 UTC |