9cdc56a | Guillaume Lederrey | 15 February 2016, 21:13:47 UTC | Ship Elasticsearch logs to logstash Introduce gelf4j as a log4j appender to send logs to logstash. Sending log to logstash is disabled by default and is activated if $elasticsearch::graylog_hosts is not `undef`. * 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 * hiera config to activate this in labs (deployment-prep) * hiera config to activate in prod/codfw Bug: T109101 Change-Id: I683bcb75c91e6c2f4dd648517435a477ab23c2a7 | 29 February 2016, 12:12:10 UTC |
90213af | Jaime Crespo | 29 February 2016, 11:41:54 UTC | Reimage dbproxy1005-11 with jessie Bug: T126251 Change-Id: I4ed6b4ad4efbee79d67817c7b6065e047f92723b | 29 February 2016, 11:45:05 UTC |
8a2c1c1 | Jaime Crespo | 29 February 2016, 11:36:09 UTC | Set up dbproxy1005 for m5 load balancing/failover It will default to db1009 (current m5-master), failovering to db2030. Bug: T126251 Change-Id: I5194de4a346fa347d4c594e3452dfe2a7331097e | 29 February 2016, 11:43:55 UTC |
bfc8ea3 | Ariel T. Glenn | 03 October 2015, 16:38:35 UTC | git deploy: update for salt bug fix, pylint salt returners work with batches so update service-restart for that, also pylint changes (var names, unused var/import, logging fix) Change-Id: I07c045fa174d4a18a34d1475289d628b0f4b10a0 | 29 February 2016, 10:42:51 UTC |
b41e9e5 | Filippo Giunchedi | 29 February 2016, 09:53:50 UTC | cassandra: add restbase1009-b instance Change-Id: I0afeada2db425af88a4c8fa9118f634f928c0a23 | 29 February 2016, 09:53:50 UTC |
564ba4d | andrewbogott | 22 February 2016, 03:13:27 UTC | Designate.conf: Specify some more pool_target settings Change-Id: I060ca94d196b21df9a6aebbbd7210c5d2cb33897 | 22 February 2016, 03:16:34 UTC |
11ca8da | cmjohnson | 26 February 2016, 18:04:31 UTC | Adding dhcpd entries for new restbase1010-1015 Change-Id: I4e95d3196092ef5842336081870a6e5600d27774 | 26 February 2016, 18:08:13 UTC |
62a9aca | dzahn | 26 February 2016, 17:56:05 UTC | ipmi: remove from iron, add on neodymium remove ipmi tool from iron since iron is becoming the 2fa test server, requested by Moritz instead add it on new salt master, neodymium it can already be used, and is used, on the puppetmaster separate from this also fix wrong path in file comments Change-Id: I460e9114ca171c7a5ba38ab0710d4b9395681011 | 26 February 2016, 17:59:59 UTC |
a0222ed | Alex Monk | 23 February 2016, 15:27:31 UTC | admin: Clarify researchers vs. statistics-users rights Change-Id: I91c2873c7a0ab24d4e93dda93f626f5601bf1514 | 26 February 2016, 16:49:49 UTC |
6bc80ad | andrewbogott | 15 February 2016, 11:04:39 UTC | Added config files for Openstack Liberty Keystone: - Changes to driver path specifications - Added a bunch of new hashed-out settings from the sample config. Nova: - Updated libvirt driver Glance: - Purged obselete config setting scrubber_datadir Designate: - No config changes but there's probably something to do here re: designate-zone-manager/tooz Change-Id: Ic61f50d57d4c319472724375434692f057565918 | 21 February 2016, 22:26:15 UTC |
72186e3 | elukey | 26 February 2016, 12:27:16 UTC | Add mc1017/mc1018 back to the memcached/redis pools after maintenance. Bug: T123711 Change-Id: I609ca121172ca0201cb2f1ea19b97a8528350a2c | 26 February 2016, 12:27:16 UTC |
23b6290 | Brandon Black | 26 February 2016, 11:05:03 UTC | vcl_config: pass_random setting, disable for upload-fe Change-Id: I83613c8bc9ebabea30f8d1ffb745119b43794455 | 26 February 2016, 11:06:54 UTC |
184112d | elukey | 26 February 2016, 10:46:26 UTC | Remove mc1017/mc1018 from the redis/memcached pools for maintenance. Bug: T123711 Change-Id: I4bea3808ef4ea44c786d884187ee5436fbab270d | 26 February 2016, 10:47:44 UTC |
63b4b39 | elukey | 26 February 2016, 09:41:22 UTC | Add mc1016 back to the redis/memcached pools after maintenance. Bug: T123711 Change-Id: I4ad9696657101231c8b3e462de846740043d0f37 | 26 February 2016, 09:41:22 UTC |
09b36ce | dzahn | 24 February 2016, 19:44:37 UTC | admin: add nikerabbit to researchers User nikerabbit is an existing shell user who has access to statistics servers. This additional group will let him read a mysql config file including a password, so he can then access the research database which has private data. Currently, the whole purpose of the researchers group is to handle that permission to get the database credentials because it was done manually before. Already has approval on ticket and does not involve sudo so it does not have to be in the ops meeting. Technically this is an access requests for database admins. Bug:T127808 Change-Id: I774d8d1673ff0a969f59937d7b9c0101ec80f6a4 | 26 February 2016, 09:27:51 UTC |
8faf0fd | elukey | 26 February 2016, 08:02:58 UTC | Remove mc1016 from the redis/memcached pools for maintenance. Bug: T123711 Change-Id: Ibd2caf2b4986733f9c3faeefa9c1a2a728f4f1db | 26 February 2016, 08:05:25 UTC |
546571a | Moritz Muehlenhoff | 15 February 2016, 07:47:55 UTC | Blacklist aufs kernel module It's similar to the already blacklisted overlayfs and carries similar risk. I made a salt run across the cluster and no system currently has it loaded. Change-Id: Ia070e746d41f9173f3e7aa28ee33d17bcb370d40 | 26 February 2016, 08:01:17 UTC |
66bb6ee | Brandon Black | 25 February 2016, 23:03:17 UTC | text VCL: misspass_mangle: fix bitscompat retval... Change-Id: I4535dfe34d790b401531980aefd3bf5e59cde57c | 25 February 2016, 23:03:17 UTC |
e2235bc | andrewbogott | 21 February 2016, 14:23:26 UTC | Stupid typo Change-Id: Ic0830e80e9c35d954ce6093817732d716c48accd | 25 February 2016, 22:53:49 UTC |
0461378 | Brandon Black | 25 February 2016, 13:24:55 UTC | text VCL: clean up 1be misspass mangling Change-Id: I5634d86aa320213417966d6f312c28ec25a31813 | 25 February 2016, 22:51:20 UTC |
7e542fa | Brandon Black | 25 February 2016, 13:16:58 UTC | VCL: rename cluster-common funcs for clarity Change-Id: I9243cb969f8df1b3c77a65d0cdcf443a2bdaa14a | 25 February 2016, 22:49:49 UTC |
4dac237 | andrewbogott | 21 February 2016, 12:20:37 UTC | Lower max-mthreads for pdns recursor. This is in response to a bug documented here: http://blog.powerdns.com/2014/02/06/related-to-recent-dos-attacks-recursor-configuration-file-guidance/ Previous setting (2048) allowed consumption of 4096 file descriptors, well exceeding the ulimit of 1024. Bug: T124680 Change-Id: I34d4152d329d895d33a9396a60889938a1f3288b | 25 February 2016, 22:45:35 UTC |
43ce719 | Brandon Black | 25 February 2016, 22:41:08 UTC | VCL: move role include up, attempt #2 Change-Id: Icdbd0886eaceb496341e81ada3fd11b663adff05 | 25 February 2016, 22:42:42 UTC |
80f8d6a | Brandon Black | 25 February 2016, 22:39:51 UTC | Revert "VCL: move role include up" Of course, the cluster files need common's code for directors and backends defined first, so it can't quite be done this way... This reverts commit 4c139575baa4b2e7f8d7fd60f2cdae100a49a1d0. Change-Id: I097e2aba162a6776c547ef232e4beff9552ba4f7 | 25 February 2016, 22:41:49 UTC |
4c13957 | Brandon Black | 25 February 2016, 13:06:28 UTC | VCL: move role include up Now that we have single-definition for the entrypoint routines, the order of the include for the cluster/layer -specific VCL file no longer changes the meaning of the VCL, so we can include it at the top of the common file. Change-Id: Id2746fa4c3902ab7b30058ee82240754decc4052 | 25 February 2016, 21:58:44 UTC |
000c6bf | Brandon Black | 25 February 2016, 12:53:59 UTC | vcl_error: single definition Functional no-op Change-Id: I997f9c63f5ead83a84a1356cd6c335a1841d2625 | 25 February 2016, 21:58:43 UTC |
afb807e | Brandon Black | 20 February 2016, 18:50:10 UTC | vcl_(hash|hit|miss|pass|fetch|deliver): single definition Functional no-op Change-Id: Ib46c90fd84d51ab9ea7cca19ed110a72a6c657d5 | 25 February 2016, 21:58:43 UTC |
9fe6064 | Brandon Black | 20 February 2016, 18:21:33 UTC | vcl_recv: single definition in wikimedia.vcl This is a functional no-op! This eliminates 'sub vcl_recv' from all places except the cluster-shared wikimedia.vcl, by creating hooks that call down the include chain logically as e.g. "sub cluster_be_recv" This simplifies the flow of control for humans to understand when looking from the top-down starting at mediawiki.vcl: it gets rid of the implicit concatenation of multiple vcl_recv, gets rid of the pattern of mw.vcl calling cluster.vcl which calls back into mw.vcl again, and gets rid of evaluation-order dependencies on where "include" statements exist at. The intent is to follow this up with similar work for other entrypoint subroutes, but vcl_recv was the most-difficult case and deserved its own commit and explanation. Change-Id: I610242c6e1440a8885407ac03e655cd250600f2b | 25 February 2016, 21:58:36 UTC |
3997891 | Brandon Black | 24 February 2016, 22:38:11 UTC | VCL: re-arrange subs into standard ordering Select one standard order for the entrypoint functions, so that things are less confusing: init fini recv hash hit miss pass fetch deliver error Change-Id: I38fbc050caba735ee0c82f5c76a5e7ab85ee6486 | 25 February 2016, 21:43:46 UTC |
fcf3ffa | Brandon Black | 24 February 2016, 20:30:24 UTC | upload VCL: refactor a bit Change-Id: I43fc1050cba93b91be910cb88970d1a26093a173 | 25 February 2016, 21:40:47 UTC |
e4c2420 | Brandon Black | 25 February 2016, 21:37:25 UTC | wikimedia-common VCL: small whitespace output fixup Change-Id: I9618581ec3adc6fe4e82387bfb176133ca7772e5 | 25 February 2016, 21:37:25 UTC |
b7f5a84 | Brandon Black | 25 February 2016, 21:16:51 UTC | wikimedia-common VCL: split output files Of course, various fe/be -specific parameters don't template correctly into a single on-disk copy... Change-Id: I8f971c946b2d13836cad5f49c40dac418fbee93e | 25 February 2016, 21:16:53 UTC |
47a4afd | Brandon Black | 25 February 2016, 00:48:06 UTC | vcl_layers: split deliver|error Change-Id: I84aee62e230233aa13d8e035011c6b16139da43f | 25 February 2016, 21:08:19 UTC |
5b52ba2 | Brandon Black | 25 February 2016, 00:42:56 UTC | vcl_layers: split hit|miss|pass|fetch Change-Id: I2c55d5e2ed7867db0dc81b3d57b784abbe337bbe | 25 February 2016, 21:07:51 UTC |
7c89a9c | Brandon Black | 25 February 2016, 00:31:13 UTC | vcl_layers: split up vcl_recv itself Change-Id: Iea402d409d53af030b09b9bab1e40149b2834476 | 25 February 2016, 21:06:26 UTC |
2a3f98b | Brandon Black | 25 February 2016, 00:23:59 UTC | vcl_layers: move basic fe-only things to fe Change-Id: Ief2e2966c91e01b9a4c187e205ef011eebfb6cc9 | 25 February 2016, 21:06:09 UTC |
6828aca | Brandon Black | 25 February 2016, 12:55:58 UTC | vcl_layers: move cluster include to layer files Change-Id: I64248c9cdbbf58179105ab901e44cc2fafd53868 | 25 February 2016, 21:01:42 UTC |
bc7163a | Brandon Black | 25 February 2016, 00:10:48 UTC | vcl_layers: split wikimedia.vcl in puppet terms This step looks pointless on the surface, but sets up the 3x files the code will be split to, allowing the next commits to move code between them as appropriate... Change-Id: I00feca8ec41a4125e4f3d3a5f96ed40796765408 | 25 February 2016, 20:59:13 UTC |
cb2aa78 | Brandon Black | 24 February 2016, 23:57:47 UTC | vcl_layers: more layer-conditional redundancy Change-Id: I99052554fa38077edf4f5f1aa41bac25b8d2b09e | 25 February 2016, 20:30:45 UTC |
db94ff3 | Brandon Black | 24 February 2016, 23:56:39 UTC | vcl_layers: move xcache up a bit Change-Id: I9380bfae329ba28370455bd346280e7f48cb2fbb | 25 February 2016, 20:30:17 UTC |
5f28449 | Brandon Black | 24 February 2016, 23:54:38 UTC | vcl_layers: varnishcheck can happen earlier Change-Id: Ifb01b9ecd0312d64a17430760b8eef5f2e1a3a21 | 25 February 2016, 20:29:40 UTC |
8f25134 | Brandon Black | 24 February 2016, 23:53:35 UTC | vcl_layers: beacon should only be frontend Change-Id: Iafc5e401805735d9d35a20ca619da2ea4f533c16 | 25 February 2016, 20:29:16 UTC |
fae4faa | Brandon Black | 24 February 2016, 23:51:47 UTC | vcl_layers: move backend 403 check earlier Change-Id: I19a35d9c235200b5de856ea4b4b3d7a789a45f75 | 25 February 2016, 20:28:59 UTC |
994d3e2 | Brandon Black | 24 February 2016, 23:48:30 UTC | vcl_layers: https+fe layer-conditional fixups Change-Id: Ibd3c0714f90b6b6deebec224a1087eca5a3d6dad | 25 February 2016, 20:28:42 UTC |
d7968dd | Brandon Black | 24 February 2016, 23:47:27 UTC | vcl_layers: re-arrange fe_ip/xff/xcdis-clear Change-Id: Ie178f8cd74f0ad82049ee9f1d8129c4d9442f399 | 25 February 2016, 20:28:22 UTC |
5903f8f | Andrew Otto | 23 February 2016, 22:15:13 UTC | Use different my.cnf in labs for analytics-meta instance Change-Id: Ia1525020e068f95b4baf2d01924c0b4164429da3 | 25 February 2016, 19:27:12 UTC |
24fa296 | elukey | 25 February 2016, 19:10:08 UTC | Add mc1015.eqiad.wmnet back in the redis/memcached pools. Bug: T123711 Change-Id: I7144dd9e7da11a610d286524da7e0f6e0c5e3844 | 25 February 2016, 19:10:08 UTC |
ba9e78e | Brandon Black | 24 February 2016, 23:45:13 UTC | vcl_layers: consolidate 3x fe-only Change-Id: Ifb32a3c26512471d743b7e354f0e3ed77a619a10 | 25 February 2016, 18:42:57 UTC |
7339927 | Brandon Black | 24 February 2016, 23:43:07 UTC | vcl_layers: move recv_purge Change-Id: Ic3d4c67301f418d987531554bb9723c5172c05ce | 25 February 2016, 18:42:22 UTC |
b6f0233 | Brandon Black | 25 February 2016, 11:26:57 UTC | vcl_layers: consolidate fe-only sub blocks Change-Id: Ifce4314a70bad2cd616368dc06648052cf4f4405 | 25 February 2016, 18:41:44 UTC |
87d0b43 | Brandon Black | 25 February 2016, 11:25:20 UTC | vcl_layers: rewrite_proxy_urls fe-only Change-Id: I38358338f13a1edd8977dab5e6aa1949af79041f | 25 February 2016, 18:40:59 UTC |
283bc00 | Brandon Black | 25 February 2016, 11:39:07 UTC | wikimedia.vcl: trailing whitespace fixups Change-Id: Ib8c0a4e09f392943288033993a46f630ba90538f | 25 February 2016, 18:39:36 UTC |
8e128bc | andrewbogott | 19 February 2016, 04:41:46 UTC | Updates to designate/mdns/pdns setup for Labs internal dns This change is in response to a long conversation with a designate developer, Kiall MacInnes. It adds a second pool_nameserver representing the secondary dns server (holmium) This /may/ resolve confusion in interaction between pdns and mdns. It also creates a clear path forward for splitting out the two pdns servers onto separate dbs, which I have more faith in as a proper fix for the associated bug. Bug: T124680 Change-Id: Ib12f6a9c6771d43e4eff479cc412ebb7db759023 | 21 February 2016, 09:37:40 UTC |
0ee3b68 | Filippo Giunchedi | 15 September 2015, 13:00:28 UTC | restbase: make statsd metric prefix configurable Bug: T112644 Change-Id: I1a4ace18d2b1a3f1227691ae054da1d588646302 | 25 February 2016, 17:46:06 UTC |
0c5b127 | elukey | 25 February 2016, 16:52:30 UTC | Remove mc1015.eqiad.wmnet from the redis-memcached pool for maintenance. Bug: T123711 Change-Id: Ib3bd9ca3a4c72d5b6c8f179d9a716ff1bbed3066 | 25 February 2016, 17:33:05 UTC |
5c86c90 | Tyler Cipriani | 23 February 2016, 16:39:17 UTC | Update scap to 3.0.2-1 Changes since 3.0.1-1: * loose mtime comparison for generated cdb files * only use 'sudo' in sudo_check_call if calling user is not euid * symlink add -T to treat LINK_NAME as normal file Change-Id: If731262a768f158358cd9097196912bca6adb35c | 25 February 2016, 17:23:05 UTC |
5809a4e | Tyler Cipriani | 09 February 2016, 20:34:59 UTC | Beta: Move deployment server This patch is everything that was cherry-picked on beta's puppetmaster in order to move the deployment server in beta from deployment-bastion to deployment-tin. * Retitle git::clones inside beta::autoupdater, conflict with git::clones in scap::l10nupdate. * Don't mount nfs to deployment-tin since we want to get rid of NFS in beta. * Change all instances of "deployment-bastion" to "deployment-tin" * Change all instances of the old ip (10.68.16.58) to the new ip (10.68.17.240) Bug: T126377 Bug: T126568 Bug: T126537 Change-Id: I5322b68dfe2e34d0a9170abd31a03f3afbda3907 | 25 February 2016, 17:15:02 UTC |
afd2ed1 | Stanislav Malyshev | 23 February 2016, 23:15:55 UTC | Always add Access-Control-Allow-Origin for WDQS backend response Bug: T115476 Change-Id: I028f0188c5c72593ee0432fd8c75ebea0753af25 | 25 February 2016, 17:10:54 UTC |
bcde806 | elukey | 25 February 2016, 16:10:58 UTC | Add mc1014.eqiad back to redis/memcached pool after maintenance. Bug: T123711 Change-Id: Ide3990e440ca523343a37c1183f97aaeb4d1cdb8 | 25 February 2016, 16:13:35 UTC |
c11c6bd | Andrew Otto | 25 February 2016, 16:08:33 UTC | Install spark on hadoop worker nodes This allows spark python oozie launchers to load their dependencies from the local file system. This should work without having spark installed locally on every node, but it seems to have broken in Spark 1.5.0. Change-Id: Ifc18f29aae14106563834a2844bf97a15bf490e9 | 25 February 2016, 16:09:43 UTC |
04af674 | Faidon Liambotis | 25 February 2016, 15:15:59 UTC | autoinstall: add ixgbe.allow_unsupported_sfp=1 to jessie (At least) our mc* servers have 10GbE ixgbe cards with non-Intel SFP+s (IIRC, we use DACs for these, so we're stuck between either breaking Intel or Juniper compatibility). Unfortunately and annoyingly, the ixgbe driver has been traditionally difficult with non-vendor SFPs. These systems have been running precise until now. precise kernel's ixgbe blocked non-Intel SFPs completely, so we had resorted into patching the driver ourselves, build the kernel module, embed it into the d-i initramfs and also provide it as a DKMS package for post-install kernel upgrades. Newer kernel versions do support it natively but they require a module parameter, allow_unsupported_sfp, to be set to 1 (defaults to 0). This passes this parameter to the kernel command-line from PXE so that those systems can boot and install. d-i will probably carry that option into the final system's grub config; we can decide to either keep that, or supply a modprobe.d config from puppet instead (and regenerate the initramfses). This will follow in a separate commit, if at all. Bug: T128068 Change-Id: Ic647884782861ad2d29b79a47dd663fe77be8fdc | 25 February 2016, 15:23:18 UTC |
58375b6 | Andrew Otto | 25 February 2016, 15:13:03 UTC | Remove symlinks made for $cdh::hadoop::gelf_logging_enabled when it is disabled Change-Id: I15763950afdfd60f1ad6be6d04b9c96a9430bbbd | 25 February 2016, 15:13:35 UTC |
f6f25ce | andrewbogott | 21 February 2016, 02:52:24 UTC | Consolidate labs pdns settings into a hiera dict. Among other things, this will allow us to specify realm- and host-based pnds passwords. Change-Id: I92c0d785ecb9d773a6ecab7a22b694015477e529 | 25 February 2016, 15:08:11 UTC |
47c7c2c | Giuseppe Lavagetto | 25 February 2016, 10:30:41 UTC | nutcracker: do not declare redis_codfw in labs Bug: T127845 Change-Id: I32030e3a5acf7434d2f9ff1a773865d7e6b51f2a | 25 February 2016, 10:30:41 UTC |
20f21a9 | elukey | 25 February 2016, 09:19:16 UTC | Remove mc1014 from memcached/redis pools for maintenance. Bug: T123711 Change-Id: I737d8ecd472eadbf0ab92a0d9b91a786c456c3d1 | 25 February 2016, 09:19:16 UTC |
81b92e5 | Giuseppe Lavagetto | 25 February 2016, 07:43:26 UTC | varnishkafka: update for logrotate Change-Id: I9084a83f0b92855fe1145746e624846f22fb19b4 | 25 February 2016, 07:43:26 UTC |
308ed80 | dzahn | 25 February 2016, 02:03:09 UTC | wikistats: no interactive password for mysqldump Change-Id: Ifae3e39b822fbb155065f7472e4b0ba46f77f94c | 25 February 2016, 02:03:09 UTC |
ce55ec4 | dzahn | 25 February 2016, 01:25:03 UTC | wikistats: fix date in mysqldump file name again Change-Id: I7bee1ea3a4d2f5a3d8fac3383af247aad12f4b2e | 25 February 2016, 01:25:03 UTC |
769d8de | dzahn | 25 February 2016, 01:00:38 UTC | wikistats: fix date in filename for db dumps Change-Id: I705c9c3953abd4b2a28cf6bb1eaa35ae5e057826 | 25 February 2016, 01:00:38 UTC |
374cb32 | dzahn | 05 September 2015, 01:41:01 UTC | wikistats: crons for db backup Create and keep some db backups, but also make sure older ones get deleted again. Change-Id: I0636faff52b778f62d5bd1ea80da6740fc9369a9 | 25 February 2016, 00:29:36 UTC |
7e7e8fb | dzahn | 24 February 2016, 20:28:57 UTC | wikistats: let puppet git clone vs. deb Letting puppet git clone this, then will need a small deploy script to copy things in the right places. In this case, building a deb for each update was just making it harder for volunteers to get stuff deployed. Later, operations/debs/wikistats can be removed. Change-Id: Iaecb2863e769686a7cea5bf0a33253b6f9e09ea1 | 24 February 2016, 21:24:06 UTC |
fcf0fa9 | Andrew Otto | 24 February 2016, 20:31:18 UTC | Update cdh module with spark fixes for CDH 5.5.2 Change-Id: I71deb7fbbe5ad4bd53271ad770fb26439d6b39c2 | 24 February 2016, 20:31:57 UTC |
9904659 | cpettet | 24 February 2016, 20:03:03 UTC | labstore: persist cfq io scheduler Change-Id: I7ccfa5f165bb2c26677061f2fe6f7236b39570e4 | 24 February 2016, 20:03:03 UTC |
0634e6d | cpettet | 12 February 2016, 22:45:42 UTC | labstore1001: persist cfq ioscheduler Bug: T126090 Change-Id: I9ecf948b509bed7a36a6288ccbd6b7561328f4f3 | 24 February 2016, 19:56:52 UTC |
7222579 | cpettet | 24 February 2016, 19:38:47 UTC | nfs: deprecation warnings on /etc/modprobe.d All files here should end in .conf Change-Id: I8c4829db7ddba5e46291efc1339660bdfdd419cc | 24 February 2016, 19:39:56 UTC |
a45df75 | cpettet | 24 February 2016, 19:30:16 UTC | labstore: tcp-setup path fix Change-Id: I57d56e3360f5416b4dcd20ad18d5281d68873574 | 24 February 2016, 19:37:19 UTC |
ba5f0b7 | cmjohnson | 23 February 2016, 20:36:13 UTC | Fixing partman recipe that wmf4727-test uses. Needed gpt Change-Id: Ib34f764bc661ddf15c30c4a2433a7f0458c096bd | 24 February 2016, 19:31:57 UTC |
d60539d | cpettet | 23 February 2016, 22:40:46 UTC | labstore: real time application of tc setup This is intended to be merged post careful manual post staggered rollout of 272891 Change-Id: Id3892fc74cb42cbb92622eebf783979ce119d70d | 24 February 2016, 19:04:04 UTC |
8e83330 | Giuseppe Lavagetto | 24 February 2016, 18:41:15 UTC | nutcracker: fix beta config Bug: T127845 Change-Id: I7e4179f6922dbe4c6008c6451b01d9532fdc190f | 24 February 2016, 18:43:15 UTC |
bfbc4cd | Brandon Black | 24 February 2016, 18:11:13 UTC | Revert "cache_misc: disable do_gzip" This reverts commit 3a24cabfd9dcef496cd280c73c3c5a489dc7c6a8. Bug: T127294 Bug: T127931 Bug: T96846 Bug: T107749 Change-Id: I3c18ca48518ccf084c10c950d2d50f9a29ca1053 | 24 February 2016, 18:11:42 UTC |
f1fe37b | Brandon Black | 24 February 2016, 18:03:50 UTC | disable keepalives everywhere Bug: T127294 Bug: T107749 Change-Id: Id6821f3e5f6b0c42a5031c8e6f5036d8b3c243c1 | 24 February 2016, 18:03:52 UTC |
78babbf | dzahn | 24 February 2016, 17:41:56 UTC | lint: fix some of the few remaining warnings There is not that much left in this area, a small follow-up to Ib062d7142a79. Very few warnings left in general. Change-Id: I71dbdb10b6a480d120893262a2db05d23d311a05 | 24 February 2016, 17:56:26 UTC |
1279037 | andrewbogott | 20 February 2016, 21:37:37 UTC | Glance policy.json: Give glanceadmin a bunch more rights Bug: T127755 Change-Id: I795c329aab01ec2945d33f520a6ee7ef7e879fd6 | 24 February 2016, 16:56:42 UTC |
27d56ff | Brandon Black | 24 February 2016, 16:20:15 UTC | disable keepalives on cache_text T127931 Change-Id: If6ccbc351f9ba10a33793491a2b2f0325465e5f9 | 24 February 2016, 16:32:23 UTC |
df1afb3 | Brandon Black | 24 February 2016, 16:20:29 UTC | Revert "disable do_gzip on cache_text (experiment)" This reverts commit 6c320dfab680c64ab19a23b1bc9f828bcd7f7d34. Change-Id: I8f2f78aa7cfc6dd2b4a29909b2b839f81cf96477 | 24 February 2016, 16:32:18 UTC |
4c4007c | Dan Duvall | 23 February 2016, 22:59:42 UTC | Fix role::programdashboard include and hieradata Change-Id: I74d9c18284bc9827e4a2dce2f9a5c8e08a1a0336 | 24 February 2016, 15:45:29 UTC |
bff2767 | Andrew Otto | 24 February 2016, 15:23:10 UTC | Add bmansurov to analytics-privatedata-users Bug: T113069 Change-Id: Ifdda4ec9ad4e9534f669cccb6dd3bfaac31361c6 | 24 February 2016, 15:23:10 UTC |
a265c2a | Moritz Muehlenhoff | 24 February 2016, 14:27:25 UTC | Assign salt grains for yubiauth servers and use in debdeploy Change-Id: I139f60f560fcc6843fd4327dc14c011902c5469d | 24 February 2016, 14:56:59 UTC |
161947d | Moritz Muehlenhoff | 24 February 2016, 14:24:53 UTC | Assign salt grains for cache::misc in esams/codfw/ulsfo and use in debdeploy Also drop references to obsolete cache::mobile and cache::parsoid Change-Id: I5f71d18f0b4ab05334f59e306702d16847d4b49d | 24 February 2016, 14:51:36 UTC |
6c320df | Brandon Black | 24 February 2016, 14:30:24 UTC | disable do_gzip on cache_text (experiment) Bug: T127931 Change-Id: I95c97b839ba7ce867994333bf1f1a645c1790e84 | 24 February 2016, 14:31:09 UTC |
ff5ac07 | Jcrespo | 24 February 2016, 13:18:41 UTC | labsdb1003 is a bit overloaded right now, move commonswiki to 1 There has been several reports of slowness on c3 (labsdb1003), while some were created by unproperly cached requests, that has been already solved and now there is still several users hitting hard commonswiki. Move those to labsdb1001, which is slightly less loaded. Bug: T127940 T127266 Change-Id: If8aee7d04e3818140f07f493ab989d965a07b935 | 24 February 2016, 13:18:41 UTC |
cec6388 | Giuseppe Lavagetto | 24 February 2016, 11:18:32 UTC | deployment-prep: fixup for I65dc207e5 Bug: T127845 Change-Id: I539a26d678768f11e6286d6d90f2995b9410dbe6 | 24 February 2016, 11:20:47 UTC |
0759603 | Filippo Giunchedi | 03 February 2016, 10:18:39 UTC | grafana: add dashboard import tool the idea is to dump dashboards from grafana into git/puppet and thus 'freeze' them Bug: T125644 Change-Id: If30f9dd06d0c3b265ed59f0f8427872fd5d2b178 | 24 February 2016, 10:39:51 UTC |
f25b659 | Filippo Giunchedi | 17 December 2015, 11:18:08 UTC | swift: adjust mount options for debian and ubuntu Bug: T117972 Change-Id: I715937b4b07a2dbd4702ab4cf73e1005680e6198 | 24 February 2016, 09:53:01 UTC |
61f230e | Chad Horohoe | 24 February 2016, 05:52:11 UTC | Drop bogus apache2 error "AH01075: Error dispatching request to :" So noisy, you don't see anything else in the apache2 log :) Change-Id: I8ff1ad21c650624412dd20b66c386f2748b13009 | 24 February 2016, 05:52:11 UTC |
af6f168 | Andrew Otto | 23 February 2016, 23:36:34 UTC | Add ori to analytics-privatedata-users Ori is already root, and also officially in charge of Analytics team, so I didn't think this needed the 3 day wait time. Change-Id: Ia694855e891eb52e4707001832796b8dcc56f930 | 23 February 2016, 23:37:37 UTC |
6f87602 | andrewbogott | 20 February 2016, 17:00:42 UTC | labtest: Update designate domain ids Change-Id: I9c56301d8ae23b7dc5f9043e350fce3a4673fb1e | 20 February 2016, 17:00:42 UTC |
daf43ef | andrewbogott | 20 February 2016, 16:42:12 UTC | designate.conf: remove terminal comment that was wreaking havok Change-Id: I10023c9315f50ae031068f7a86e56700f9d61d74 | 20 February 2016, 16:42:12 UTC |
aa43436 | cpettet | 23 February 2016, 21:57:56 UTC | labstore: stage tc script and enable on boot We have been piloting this configuration for a few weeks to limit the variable load exposure of labstore10* in tools. Labstore1001 especially has difficulty sustaining the extreme bursty load that any operation can trigger. These shaping values are based on data collected from the nfsiostat diamond collector deployed a few weeks ago. This will all most likely require tweaking over time. Change-Id: I19814c3da53f78666a8877937353f3987a904501 | 23 February 2016, 22:50:16 UTC |
6818e46 | dzahn | 23 February 2016, 22:45:57 UTC | programdashboard: move role to modules/role Change-Id: I79fe8a55e94d5f8c365b380c430befcead92c33f | 23 February 2016, 22:45:57 UTC |