| 8abc8ba | Faidon Liambotis | 11 February 2014, 16:29:01 UTC | rsyslog: use provider => 'upstart' Similarly to Ifb657c7, switch rsyslog to upstart as well, as it fails on trusty otherwise. Both this and the SSH one can be removed once we upgrade to a newer version of puppet, as puppet gained in a 2.7 point-release the line: defaultfor :operatingsystem => :ubuntu which we don't seem to have in precise's 2.7, nor do we have support for newer style upstart overrides. Change-Id: I688954780d0c642d9033a940d45584203c9c8369 | 11 February 2014, 16:31:53 UTC |
| 59d733c | Matanya Moses | 11 February 2014, 13:12:27 UTC | remove streber from puppet,dsh,dhcpd RT #2186 can be decom'ed per mark Change-Id: I546db3c0345a1ca503f550e08bac569c7d865163 | 11 February 2014, 16:28:40 UTC |
| 47239a7 | Faidon Liambotis | 11 February 2014, 16:08:25 UTC | base: remove initcwnd if-up.d hook for >= oneiric We currently have an if-up.d hook to set the TCP initial congestion window to 10*MSS on all boxes, but this is currently unneeded on all of our recent boxes. Linux upstream made 10 the default with commit 442b9635c, which is included since 2.6.39, way before any of our precise systems. As checking for a running kernel is kind of evil, make the check conditional to lsbdistcodename, and remove the if-up hook in everything post-oneiric (natty = 2.6.38, oneiric = 3.0). Change-Id: Idfe83887b5601fe1c02766ea6d7121a2bf516ef4 | 11 February 2014, 16:15:37 UTC |
| 533e8f7 | dzahn | 10 February 2014, 21:48:38 UTC | remove locke from puppet,dsh,dhcpd,ud2log filter RT #6168 can be decom'ed anytime per ottomata delete from dsh groups, ganglia comment, logging role, site.pp, dhcpd and ... Change-Id: I3fe32f4830289ad297a4335472db8137d09cd21a | 11 February 2014, 15:56:36 UTC |
| a6b65b2 | Faidon Liambotis | 11 February 2014, 15:27:17 UTC | autoinstall: set up apt pinning from early in d-i We currently have the need that we need to install a lower version of puppet (2.7 instead of 3.4) from our own apt repository in trusty, but we have no apt pinning set up while in the installer, and hence when both are set at 500, the highest version wins. Unfortunately, apt-setup does not support setting up apt pinning and both early_command & late_command are unsuitable for doing so (the former runs before we have /target, the latter runs after apt-get update has been run). Work around this in a very hacky way, by having the early_command shell script install a shell script hook under base-installer's prehooks runpartsdir, which runs before the apt_update step. It also runs before the install_base_system step and there's no hook step between those two, but this should be okay as long as we mkdir /etc/apt/preferences.d ourselves. Change-Id: I6dcb9d340aa54f52cdb96ef3cbcc51628e4001d9 | 11 February 2014, 15:34:06 UTC |
| f1b93ba | Faidon Liambotis | 11 February 2014, 14:32:27 UTC | autoinstall: disable multiarch by default multiarch is set to i386 on amd64 systems by default since apt-setup 1:0.79, i.e. Ubuntu saucy. We don't need it, so disable it. Change-Id: Id2c4d52617a05776c40dcc900dda4125e06e6b6d | 11 February 2014, 15:34:06 UTC |
| 3d395ab | Faidon Liambotis | 11 February 2014, 14:24:41 UTC | autoinstall: move disable_dhcp to early_command There is no reason for the disable_dhcp/disable_autoconfig boilerplate to be present in every single subnet definition where it can be moved in the place where we actually set the static IP address and call netcfg itself. This has the downside that it to a shell script, but it has the benefit of DRY, plus that we can document all netcfg exceptions. While at it, remove the netcfg/choose_interface setting from a couple of subnets, as it's defined in common.cfg anyway. Change-Id: I72681530ab76211e96a45ee46acb6e2b177e36e4 | 11 February 2014, 15:34:06 UTC |
| f2f864e | Faidon Liambotis | 11 February 2014, 14:05:36 UTC | autoinstall: another trusty early_command fix killall.sh for killing DHCP was (rightfully) renamed to kill-all-dhcp to indicate its purpose in netcfg 1.86. wheezy has 1.108 & trusty has 1.114 and are the first versions to have this change. Add support for in a backwards-compatible way and while at it, remove a useless "true" invocation. Change-Id: I282f8e2d7187e98b8e544335baba9501463b7570 | 11 February 2014, 15:34:06 UTC |
| 4f90841 | Andrew Otto | 05 February 2014, 20:54:40 UTC | [DO NOT MERGE] Setting batch_num_messages to 6000 Don't merge this yet. Magnus and are are trying to catch cp3019 misbehaving again, and we think that this setting might fix it. Change-Id: I2243fc06f410a5bdf8a8ce489e3a3b1fe16ecb2c | 11 February 2014, 15:23:23 UTC |
| 2a08452 | Ryan Lane | 11 February 2014, 09:39:00 UTC | Add Sphinx function documentation for deploy module This change adds basic Sphinx documentation to the deploy module. Change-Id: Ia8f04c8bc95ef521192e1ea9fa1a176b4bd58b8a | 11 February 2014, 15:18:10 UTC |
| 0d6d368 | Ryan Lane | 09 February 2014, 03:29:44 UTC | Ensure submodules are checked out on the deployment server If checkout_submodules is enabled on minions at any point and the deployment server doesn't have the submodules checked out then the minions will enter an unrecoverable state after the initial deploy. This change always does a submodule recursive init for repos to ensure this condition can't happen. Change-Id: I125e9275d813dc34baed10878c19be94e4a0251e | 11 February 2014, 14:56:07 UTC |
| 3266033 | Ryan Lane | 09 February 2014, 03:00:20 UTC | Continue on single repo failures for deployment server init Rather than failing immediately if a single repo is broken (for instance if the upstream is listed incorrectly), continue on to other non-broken repos. Change-Id: I27add880f7f4b864f7357a298d42cd5c8bd7e8cd | 11 February 2014, 14:55:23 UTC |
| fac3753 | dzahn | 07 February 2014, 23:56:44 UTC | add shell account for phuedx and add to mortals RT #6785 username: phuedx real name: Sam Smith email: samsmith@wm labs user/IRC user: phuedx UID: 3926 matching LDAP team: Growth requested: tin, deploy mediawiki-config, extensions/GettingStarted extensions/GuidedTours deployment works with: Matt Flaschen manager/approved: Terry Chay Change-Id: I42075d6828f287ec51a9b36588e2702a48f95ece key: https://office.wikimedia.org/w/index.php?title=User:SDSmith_%28WMF%29&oldid=106746 | 11 February 2014, 12:29:07 UTC |
| 80e0f2b | Ariel T. Glenn | 11 February 2014, 09:50:05 UTC | hourly logrot for parsoid for now Change-Id: I61fc3142ac7a553402f992295ccfaf880eb9159e | 11 February 2014, 09:53:58 UTC |
| 46c0af0 | Ori.livneh | 11 February 2014, 07:12:42 UTC | Merge "Fix submodule fetching in trebuchet" into production | 11 February 2014, 07:12:42 UTC |
| 2ce260c | Gabriel Wicke | 11 February 2014, 04:04:18 UTC | Fix parsoid log rotation The parsoid_log_file variable was not defined, so just use the static path. Can make this dynamic later if needed, but otoh we can also just switch to the Debian package. Also lower the rotation size as in the debian package. Change-Id: I0b85c64428c3801151c5facc283545317c1df1cb | 11 February 2014, 04:10:02 UTC |
| 6cd3828 | Ori Livneh | 11 February 2014, 01:47:36 UTC | Merge "scap: ensure local copy is up-to-date; chgrp wikidev" into production | 11 February 2014, 01:47:36 UTC |
| ba20f94 | Ori Livneh | 10 February 2014, 20:04:28 UTC | scap: ensure local copy is up-to-date; chgrp wikidev - set ensure => latest on the Git::Clone resource so that if an application server is down during a scap update, Puppet will eventually update it. - change group ownership to 'wikidev' so that scap is updateable by Bryan. Change-Id: I1de8ff88373d9640ebafde6c638fe89b54f2e14c | 11 February 2014, 00:33:59 UTC |
| 019b1ec | Ryan Lane | 10 February 2014, 23:51:45 UTC | Fix submodule fetching in trebuchet At some point fetch && fetch --tags was replaces with fetch --all, which doesn't do the same thing. This change reverts that code so that recursive submodule checkout will again work correctly. Change-Id: I8a878f4a79e0e6688bc7b98ab48e2284980a6517 | 10 February 2014, 23:51:45 UTC |
| 78ca8bf | Matanya Moses | 10 February 2014, 22:06:46 UTC | locke: decom, remove logging doc comments RT #6168 Change-Id: I5c9b75ecc14582bebb040a86087d7577e7b89801 | 10 February 2014, 22:18:50 UTC |
| 8d8cc30 | Chad Horohoe | 10 February 2014, 22:15:42 UTC | Revoke key per request Change-Id: Ib2429d9721e26ef569655a52a58a97aa0c925e41 | 10 February 2014, 22:16:05 UTC |
| d6597a1 | Matanya Moses | 10 February 2014, 22:00:16 UTC | locke: decom, remove udp2log filters RT #6168 Change-Id: I3b6e0dbe6950b43d8dff7c5fa97fa6124658fb31 | 10 February 2014, 22:00:16 UTC |
| c8ec830 | dzahn | 08 February 2014, 01:48:20 UTC | decom 'harmon' - rm from site/dsh/partman/dhcp RT #6267 Change-Id: Idc44f926f3bf1ccda9065e3fb685672fb051604a | 10 February 2014, 20:14:22 UTC |
| 5cec161 | Andrew Otto | 10 February 2014, 19:28:27 UTC | Adding new research leila, giving access to stat1 and stat1002 and bastions RT 6765 Change-Id: I60633e881d7b497a12ebaf95ffc348f27ce0ff87 | 10 February 2014, 19:28:27 UTC |
| 140db31 | Faidon Liambotis | 10 February 2014, 18:20:51 UTC | install-server: allow /sbin/ip for early_command ip is installed under /sbin in at least trusty installers (and precise & trusty systems). Leave /bin/ip as an alternative, just in case the older releases need it. This currently breaks trusty, as trusty installer has no ifconfig to fall back to at all. Change-Id: I0c4c9648802fe42cf44a4330cc9015b5d614b60c | 10 February 2014, 18:20:51 UTC |
| 781864d | Faidon Liambotis | 10 February 2014, 17:50:53 UTC | ssh: use provider => 'upstart' Apparently, on trusty, the behavior of the init.d SSH script has changed, and instead of shelling out to /lib/init/upstart-job when upstart is detected, it exits 1 instead on both "status" & "start" actions. Use provider => upstart instead, to make puppet use upstart directly. Change-Id: Ifb657c76c1b0d127acb262eee887d20d5919ba31 | 10 February 2014, 17:52:57 UTC |
| fd42e67 | Faidon Liambotis | 10 February 2014, 17:48:23 UTC | salt: remove pinning to specific package version The backport scheme fails when we actually use distribution packages (e.g. trusty has 0.17.5+ds-1). Pinning to a specific version from puppet is not great either, so remove and rely on a plain "installed" instead. Change-Id: Id21cde62ba620184d08ec8540118a43dd2e63db4 | 10 February 2014, 17:49:48 UTC |
| e7351ac | Faidon Liambotis | 10 February 2014, 17:37:14 UTC | reprepro: add i386 to trusty While we don't expect to ever install an i386 server anymore, trusty comes with multiarch enabled by default and it complains if it doesn't find it. Rather than use deb [arch=amd64] all over the place, create an empty i386 repository instead. Change-Id: I1ac4d4862af14d92555b24b0be61339462e0b6a4 | 10 February 2014, 17:37:14 UTC |
| 2d7f401 | Faidon Liambotis | 10 February 2014, 17:27:51 UTC | install-server: fix late_command for trusty Remove wikimedia-base from the packages that late_command (& standard-pacakges) installs. While it's easy to copy it from precise, it's also a useless package as the only thing that it actually does is to depend on openssh-server, which we install anyway. Make the ixgbe-dkms package installation conditional on the version of the module itself, as our ixgbe patch is not needed in recent kernels (incl. trusty). While this did no fail on the test system, it would fail on systems that actually have ixgbe, since we don't (nor plan to have) ixgbe-dkms on trusty. Change-Id: I47d40e31fb715972ce045382d290317a9c602361 | 10 February 2014, 17:33:05 UTC |
| a031bc3 | Faidon Liambotis | 10 February 2014, 16:31:06 UTC | Merge "install-server: disable biosdevname on trusty" into production | 10 February 2014, 16:31:06 UTC |
| fccad52 | Faidon Liambotis | 10 February 2014, 16:28:11 UTC | install-server: disable biosdevname on trusty trusty wants to name devices using Dell's biosdevname (they call it "consistent naming") in Dell servers and hence the primary embedded network card would get a name of "em1". This breaks our early_command, and while we could fix that, it's better to stay consistent with our old naming for now to help with server reinstalls and other changes across the board (interface:: definitions etc.). Set biosdevname=0 in the installer's kernel command-line to revert to the old behavior per https://lists.ubuntu.com/archives/ubuntu-devel/2012-August/035670.html Change-Id: I753123e54df70ce8777c7e08a79fcc4c79f490c9 | 10 February 2014, 16:28:11 UTC |
| 0552595 | Antoine Musso | 10 February 2014, 16:01:41 UTC | contint: file{} needs 'owner' not 'user' Change-Id: Ib631a4c38ef013b36b304615b95aa2880e5f9165 | 10 February 2014, 16:01:41 UTC |
| 66ce75b | Antoine Musso | 10 February 2014, 15:34:49 UTC | contint: .gitconfig for jenkins-deploy user The jenkins-deploy git configuration has been hacked up manually in its shared directory /home/jenkins-deploy. Generate it using git::userconfig, the email gives clue about the instance being used. Change-Id: Ic22d53bfb20b8c90987a749b40c3ce6cb2a07148 | 10 February 2014, 15:58:22 UTC |
| 2180748 | ArielGlenn | 10 February 2014, 15:57:58 UTC | Merge "contint: move jenkins-deploy user home to /dev/vdb" into production | 10 February 2014, 15:57:58 UTC |
| 54c771d | Antoine Musso | 10 February 2014, 15:25:04 UTC | contint: move jenkins-deploy user home to /dev/vdb `jenkins-deploy` user home dir ends up being in /home/ which is using GlusterFS on the integration project. Additionally, we need the home directory to be local to the instance since it is used by gems/npm as a local cache. To avoid using GlusterFS and race conditions, migrate it to /mnt/home/jenkins-deploy which is a /dev/vdb mount. WARNING: need to update Jenkins slaves configurations bug: 61144 Change-Id: Ie90874f23804e273bdfb4c698b19495f3aff56f8 | 10 February 2014, 15:56:27 UTC |
| 5de5a41 | Faidon Liambotis | 10 February 2014, 15:51:50 UTC | Switch copper to Ubuntu trusty copper is a Swift test/replication host; we can safely use it for such experiments. Change-Id: I6ee51feb48187629aeae2ac904245b65e97fc13b | 10 February 2014, 15:51:50 UTC |
| 8770c1a | Faidon Liambotis | 10 February 2014, 15:49:11 UTC | install-server: add Ubuntu 14.04/Trusty Tahr Add the debian-installer PXE configs for Trusty, as well as the reprepro bits to setup the new apt repository for the new distribution. Change-Id: I0ce1c067db7fe382ebe7282492c695dadc6edae9 | 10 February 2014, 15:51:35 UTC |
| 26abefd | dzahn | 08 February 2014, 01:00:44 UTC | Add shell account for hoo, admins restricted RT #6731 labs user: hoo UID: 2133 real name: Marius Hoch requested/approved: csteipp, Eloquence requested: accessing the DB (investigating bugs), possibly: general Wikidata deploys, debugging and config Giving admins restricted first. After getting used to the cluster this will be changed to mortals. Change-Id: I225f6d54d9bbf948639a301456b76a682fa8bb4c | 10 February 2014, 15:28:10 UTC |
| decb528 | Ariel T. Glenn | 10 February 2014, 12:24:28 UTC | caesium added as backend for varnish misc web cluster Change-Id: I1a5fda48b4a94711a108a5d562d57b20677e4241 | 10 February 2014, 12:24:28 UTC |
| 77ed10d | andrewbogott | 09 February 2014, 02:26:03 UTC | Switch nova-network vlan from 1118 to 1102. Stab in the dark! Change-Id: Iae87ebe9f7aed3e21221506fb051b1406bcdfc78 | 09 February 2014, 02:38:27 UTC |
| 7fece06 | andrewbogott | 08 February 2014, 22:48:50 UTC | Move ports-wikitech.conf out of templates and into files. Change-Id: I2134cb935aae7792a3631ffb0ed1593ee78cad5f | 09 February 2014, 02:33:19 UTC |
| 0e0af62 | Ariel T. Glenn | 10 February 2014, 11:34:40 UTC | caesium as releases webserver (mw tarballs etc) Change-Id: I9ea8e0397ba3cb6c44390b4aefe70a4003840608 | 10 February 2014, 11:34:40 UTC |
| 317a7fd | Ariel T. Glenn | 10 February 2014, 08:50:20 UTC | module for releases webserver (mobile and mw tarballs) Change-Id: Ife41e6f5c937e5af6b922cf118b7a7042dd794a9 | 10 February 2014, 11:13:24 UTC |
| 90acc91 | Antoine Musso | 05 February 2014, 13:35:36 UTC | contint: fix slave-scripts deployment on labs We can not use git-deploy from tin.eqiad.wmnet to update instances in labs. CI has a bunch of slaves that needs the slave-scripts repository to be kept up to date somehow, hence the use of git::clone in contint::slave-scripts. This patch: * prevents contint::slave-scripts on production since it will conflict with git-deploy * updates contint::slave-scripts clone path to match production ie: /srv/deployment/integration/slave-scripts * apply contint::slave-scripts to all labs slave by including it in role::ci::slave::labs::common Change-Id: I1795f325ed8afb8beb3d894ff5b0e8b89e7b085d | 10 February 2014, 11:09:27 UTC |
| 13823c3 | Antoine Musso | 05 February 2014, 13:31:20 UTC | contint: slave-scripts are deployed via git-deploy The slave-scripts on CI production slaves (role::ci::slave) are deployed via git-deploy (they have a deployment-target). There is no more need to use ontint::slave-scripts which was basically using git::clone to fetch the scripts in /srv/slave-scripts. Change-Id: Ida49e6ec11046e53d7a0fe088c555614d21b93d0 | 10 February 2014, 11:06:57 UTC |
| 9f24da0 | Antoine Musso | 04 February 2014, 16:31:21 UTC | contint: browsers for testing + xvfb for headless We want to run browser tests using Chromium and Firefox in headless mode. This is made possibly by installing the browsers and adding in xvfb, a fake X server. Change-Id: I58ccedfe6bf949e0bc8d07b211d7eaf26e27c8e0 | 10 February 2014, 11:03:38 UTC |
| 39120be | Alexandros Kosiaris | 10 February 2014, 07:55:56 UTC | Merge "puppetize ports-wikitech.conf on wikitech hosts." into production | 10 February 2014, 07:55:56 UTC |
| 49c612d | Ori.livneh | 10 February 2014, 07:52:15 UTC | Merge "logstash: Add normalized_message field to all events" into production | 10 February 2014, 07:52:15 UTC |
| 8cf6cd3 | andrewbogott | 10 February 2014, 06:55:51 UTC | Merge "Added role::nova::manager to virt1000, so I can debug OSM." into production | 10 February 2014, 06:55:51 UTC |
| 4c859f4 | Sean Pringle | 10 February 2014, 03:34:12 UTC | update coredb topology for s2 master switch Change-Id: Ie4209d3894b02a5db55b9d32d50837134caddbf7 | 10 February 2014, 04:08:44 UTC |
| 0013c08 | Ryan Lane | 09 February 2014, 02:27:23 UTC | Remove old parsoid deploy repos and remove parsoid salt module It seems at some point the parsoid module was removed from the repo, but wasn't removed from the salt config. The missing module breaks the module/grain/pillar refresh across the cluster. This change cleans that up and removes any reference to it (which removes the old parsoid deploy repos). Change-Id: I74d63e9dfd516f7afc27f111d5919f6f81db813a | 09 February 2014, 02:38:40 UTC |
| 6ca91d4 | andrewbogott | 08 February 2014, 22:32:12 UTC | puppetize ports-wikitech.conf on wikitech hosts. Change-Id: I49fa3244cd3a74f320b3b53fa3bf59057c4fe123 | 08 February 2014, 22:36:41 UTC |
| 721d80d | andrewbogott | 08 February 2014, 21:39:06 UTC | Added role::nova::manager to virt1000, so I can debug OSM. Change-Id: I2f0da424bd773144b160d9d004a51e73d9079351 | 08 February 2014, 21:39:06 UTC |
| b529b09 | andrewbogott | 08 February 2014, 15:35:16 UTC | Set multi_host to False. As best I can tell, multi_host means that nova-network is running on multiple hosts. I don't know why setting this to True in Folsom worked... Change-Id: I6b490f6c7885957e2963dda2dab23821b141ad7b | 08 February 2014, 15:35:16 UTC |
| a93dda0 | andrewbogott | 08 February 2014, 13:25:52 UTC | Revert "Include the nova-api-metadata service in havana." This is interfering with nova-api, somehow... This reverts commit 6213f688681a8fd0972fa762744d3431ae533604. Change-Id: Ide865b707c2b4807f3f73df414b78a34ccb3266d | 08 February 2014, 13:26:26 UTC |
| 6213f68 | andrewbogott | 08 February 2014, 11:03:40 UTC | Include the nova-api-metadata service in havana. Change-Id: I860b580328b93648ea9c54899acb16d86a035856 | 08 February 2014, 13:16:53 UTC |
| 9126397 | andrewbogott | 08 February 2014, 13:13:31 UTC | Add rabbit login/password in case that helps. Change-Id: Idc582880316877ecf0fe23fcb3b4102d0c94d83f | 08 February 2014, 13:15:58 UTC |
| db8ed98 | andrewbogott | 08 February 2014, 10:52:38 UTC | Merge "Set network_api_class for nova-network" into production | 08 February 2014, 10:52:38 UTC |
| 2c5b46f | andrewbogott | 08 February 2014, 10:49:34 UTC | Set network_api_class for nova-network Change-Id: Ie3e3d76b26b6a8d7c057441d4a4e457bda99eaad | 08 February 2014, 10:50:02 UTC |
| 82b00a5 | Sean Pringle | 08 February 2014, 09:26:36 UTC | Appears to need same fix as If0117511497ee1457e63d2710f1c3e29f4f44bc0 Change-Id: Ibdc0e3db56849b8bd2115a1880dd9d22f5a464db | 08 February 2014, 09:27:33 UTC |
| 5203f27 | andrewbogott | 08 February 2014, 06:02:37 UTC | Added auth_uri for nova. Change-Id: I03356ea3dff96850f2eec8f5c7021f8f013e9902 | 08 February 2014, 06:02:37 UTC |
| fb138a5 | andrewbogott | 08 February 2014, 05:29:53 UTC | Set up eth4.1118 on labnet1001 Change-Id: Ie141a20471189b68d83cdfe7e51e81932f5951f7 | 08 February 2014, 05:29:53 UTC |
| 1524852 | andrewbogott | 08 February 2014, 05:18:47 UTC | Use 10g interface for labnet1001 Change-Id: I4662d24173f687f6c891bd79dfcb4770643fb7d5 | 08 February 2014, 05:18:47 UTC |
| b21bda9 | andrewbogott | 08 February 2014, 04:37:36 UTC | Add role::nova::network to labnet1001 Change-Id: Iab39eb38402ca2c8d20e147501bc710866fb578e | 08 February 2014, 04:37:36 UTC |
| 40ec00e | andrewbogott | 07 February 2014, 05:22:13 UTC | Add the use_neutron switch. - Wrap neutron-specific code in nova roles and classes with use_neutron - Turn neutron OFF in eqiad. - Replace (optional) nova-network settings in havana nova.conf Change-Id: I5ca96c867b6cc04cf9a0c03192d5b414647eb427 | 08 February 2014, 04:11:17 UTC |
| cc5ef32 | andrewbogott | 07 February 2014, 05:25:40 UTC | Replace some nova-network configs for eqiad. There was no need to chop this in the first place, it's activated by switchable code elsewhere anyway. This is a partial revert of abd0ded4213e94a822d448c3c53f75e1ba5f02a2 Change-Id: If3b72498aa98a080feb9bd167234871a06ea6f62 | 08 February 2014, 03:46:14 UTC |
| 05199d3 | Bryan Davis | 07 February 2014, 23:48:57 UTC | logstash: Add normalized_message field to all events Copy the message of all events destined for storage in Elasticsearch into a "normalized_message" field that is truncated to 255 characters. This can be used in dashboards as a term search to correlate common messages. Change-Id: I01c50456cf0e334075acacbe2aebe5d8fc941d31 | 07 February 2014, 23:48:57 UTC |
| ff2666d | Dzahn | 07 February 2014, 22:39:01 UTC | Merge "sudoers: remove sudoers.nrpe_fr & sudoers.search " into production | 07 February 2014, 22:39:01 UTC |
| ac7ca35 | Matanya Moses | 05 February 2014, 13:02:33 UTC | sudoers: remove sudoers.nrpe_fr & sudoers.search Change-Id: Idf973cb9125a7c6a14638393d479dcb94ce33688 | 07 February 2014, 22:36:00 UTC |
| 643d6f1 | Andrew Otto | 07 February 2014, 21:04:09 UTC | Adding alerts for Kafka Broker replication metrics Change-Id: I3116f74b7fcf6a66f865f4c5acba98593a513ffe | 07 February 2014, 21:04:52 UTC |
| 1ba6258 | dzahn | 07 February 2014, 20:21:15 UTC | fix whitespace in mysql nrpe check command "nrpe::monitor_service { 'mysql disk space':" from I100a7adb4211b7fc6f0e7a37471010d4371a795b results in files with actual whitespace, like file "check_mysql disk space.cfg" check_mysql disk space.cfg: ASCII text on the monitored servers and errors like: <+icinga-wm> PROBLEM - MySQL disk space on db35 is CRITICAL: NRPE: Command check_mysql not defined etc on all mysql servers Change-Id: If0117511497ee1457e63d2710f1c3e29f4f44bc0 | 07 February 2014, 20:21:15 UTC |
| 8834591 | dzahn | 07 February 2014, 19:13:34 UTC | fix mysql process monitoring and puppet run on pc I100a7adb4211b7fc6f0e7a37471010d4371a795b introduced a bug that caused puppet runs to break on mysql servers, f.e. on pc2: err: Could not retrieve catalog from remote server: Error 400 on SERVER: Invalid parameter nrpe_command at /etc/puppet/modules/mysql_wmf/manifests/init.pp:19 this should be nrpe::monitor_service for both of them Change-Id: Ib7411cadba127ca018bc5406ada4bf1e4643dd50 | 07 February 2014, 19:13:34 UTC |
| 03ac8b7 | Federico Leva | 31 January 2014, 09:55:49 UTC | Split exim stats to own class and add it to mchenry Bug: 57890 Change-Id: Ia24fc244c7201e96ff045b88786a895be718dfea | 07 February 2014, 17:43:40 UTC |
| f549a19 | andrewbogott | 07 February 2014, 16:00:16 UTC | Merge "Don't configure a nova-network interface in havana." into production | 07 February 2014, 16:00:16 UTC |
| bc02d41 | andrewbogott | 07 February 2014, 15:52:23 UTC | Merge "Add eth1.1102 interface to compute nodes" into production | 07 February 2014, 15:52:23 UTC |
| 962f431 | Matanya Moses | 02 February 2014, 11:59:13 UTC | mysql: change nrpe monitoring to use nrpe::monitor Change-Id: I100a7adb4211b7fc6f0e7a37471010d4371a795b | 07 February 2014, 13:05:33 UTC |
| bc8baea | Alexandros Kosiaris | 07 February 2014, 12:29:27 UTC | Merge "ganglia: puppet 3 compatibility fix: fully qualify variable" into production | 07 February 2014, 12:29:27 UTC |
| 6dc48ac | Alexandros Kosiaris | 07 February 2014, 12:26:17 UTC | Merge "ldap: puppet 3 compatibility fix: fully qualify variable" into production | 07 February 2014, 12:26:17 UTC |
| 21ec744 | Alexandros Kosiaris | 07 February 2014, 12:03:21 UTC | Merge "puppetmaster: puppet 3 compatibility fix: fully qualify variable" into production | 07 February 2014, 12:03:21 UTC |
| 5fe66dd | Alexandros Kosiaris | 07 February 2014, 11:41:52 UTC | Merge "Tools: Fully qualify variables for Puppet 3 compatibility" into production | 07 February 2014, 11:41:52 UTC |
| 18f755c | Alexandros Kosiaris | 07 February 2014, 11:23:41 UTC | Merge "Fix erroneous warning on puppet-merge" into production | 07 February 2014, 11:23:41 UTC |
| 8853424 | Alexandros Kosiaris | 07 February 2014, 11:23:06 UTC | Merge "protoproxy: puppet 3 compatibility fix: fully qualify variable" into production | 07 February 2014, 11:23:06 UTC |
| cec4955 | Alexandros Kosiaris | 07 February 2014, 11:17:47 UTC | Fix erroneous warning on puppet-merge Gerrit is the commiter on merges and as such, the script issues a false warning when a merge happens Change-Id: Ie560fa48172c8d68b8f481a615c0f32032bf7e72 | 07 February 2014, 11:19:39 UTC |
| dc269b5 | Alexandros Kosiaris | 07 February 2014, 11:15:55 UTC | Merge "lucene: puppet 3 compatibility fix: fully qualify variable" into production | 07 February 2014, 11:15:55 UTC |
| 2539ee6 | andrewbogott | 07 February 2014, 01:40:18 UTC | Don't configure a nova-network interface in havana. Change-Id: Ibb170d2b930fa1cb7b1c876cb3bfac8694a4da45 | 07 February 2014, 01:40:18 UTC |
| abd0ded | andrewbogott | 07 February 2014, 01:26:10 UTC | Add eth1.1102 interface to compute nodes Change-Id: I6982daf64133edfe5009baed21e4503871809b88 | 07 February 2014, 01:31:10 UTC |
| dd06ef4 | andrewbogott | 06 February 2014, 23:50:41 UTC | Cut down to just one compute node for now. Debugging is too confusing with multiple compute nodes. Change-Id: I66130d60ed62806f05689ee0d608e661dfe52170 | 06 February 2014, 23:50:41 UTC |
| 8874078 | andrewbogott | 06 February 2014, 23:16:49 UTC | Specify sysctl priority for openstack settings. Change-Id: I5f137c186169888db729dfd3faff267e9d95196d | 06 February 2014, 23:16:49 UTC |
| 0c9cf2b | andrewbogott | 06 February 2014, 21:13:59 UTC | Set tunnel_type and tunnel_types. Trying to get dhcp to work... Change-Id: I28cc714c500a7df71f01058a7c1526056a968b1e | 06 February 2014, 21:13:59 UTC |
| b0dbeb2 | andrewbogott | 07 February 2014, 04:58:27 UTC | Merge "Configure novnc, take two" into production | 07 February 2014, 04:58:27 UTC |
| 7d2b63b | andrewbogott | 07 February 2014, 04:47:07 UTC | Merge "Configure novnc for instance debugging." into production | 07 February 2014, 04:47:07 UTC |
| f3ba9b6 | Matanya Moses | 05 February 2014, 20:42:58 UTC | lucene: puppet 3 compatibility fix: fully qualify variable Change-Id: I424e006cd37ece2a8ab3590e2d33581b61d8c9b0 | 06 February 2014, 20:47:09 UTC |
| c69c1b5 | andrewbogott | 06 February 2014, 20:36:22 UTC | Configure novnc, take two Change-Id: If0e3d1c316136143167a13e938778b71b8866ef3 | 06 February 2014, 20:36:22 UTC |
| df09f41 | Christian Aistleitner | 06 February 2014, 20:32:45 UTC | Reflect emery -> erbium move in documentation Change-Id: I6548ca51584f65b907aad4269e9c9c2e84350b66 | 06 February 2014, 20:34:28 UTC |
| fa2c855 | andrewbogott | 06 February 2014, 20:23:10 UTC | Configure novnc for instance debugging. Change-Id: I005e9c8ea6b45712102c046f1b9c4fe0520da937 | 06 February 2014, 20:25:19 UTC |
| 9157654 | Nuria Ruiz | 06 February 2014, 17:24:45 UTC | Fixing domain name for erbium host Change-Id: Ib322242615d010f9d9b98cf5c606c40fce425d6e | 06 February 2014, 18:30:43 UTC |
| 2657fec | Ariel T. Glenn | 06 February 2014, 18:06:30 UTC | remove tantalum, add caesium to dhcp and netboot for install Change-Id: Ie7483fb0855b7cdac79b7c98f0efaa66e50007c1 | 06 February 2014, 18:06:59 UTC |
| d103c57 | Andrew Otto | 06 February 2014, 16:29:03 UTC | Increasing num_replica_fetchers to 2 on kafka brokers Change-Id: I6f46006079fff26cfe071b61d351a9b0d0d77a57 | 06 February 2014, 16:29:28 UTC |
| e185127 | Matanya Moses | 06 February 2014, 12:17:19 UTC | protoproxy: puppet 3 compatibility fix: fully qualify variable Change-Id: I36e30549abd8324a274947f8f2a6b6a9229677f0 | 06 February 2014, 15:58:04 UTC |
| 73b671b | andrewbogott | 06 February 2014, 14:55:26 UTC | Further attempt to actually add neutron_ldap_user_pass to nova.conf Change-Id: If1263737cd712106251c563a0eee8978f814fd3e | 06 February 2014, 14:55:26 UTC |
