fe87eea | Giuseppe Lavagetto | 11 February 2015, 16:51:15 UTC | memcached: add mc1018 to the mediawiki pool as shard 18 Change-Id: I5eb0b50f0138b9580faee2cbf89e26e64e2eb37e Signed-off-by: Giuseppe Lavagetto <glavagetto@wikimedia.org> | 11 February 2015, 16:51:15 UTC |
1ad2c51 | Giuseppe Lavagetto | 11 February 2015, 16:48:45 UTC | memcached: add mc1017 to the mediawiki pool as shard 17 Change-Id: Ic960c20fbd51c16f5afaad4065c5ee424595bbda Signed-off-by: Giuseppe Lavagetto <glavagetto@wikimedia.org> | 11 February 2015, 16:48:45 UTC |
b5c3b7a | Giuseppe Lavagetto | 11 February 2015, 16:26:23 UTC | dhcpd: mc1018 and mc1017 are ubuntu precise Change-Id: I6d53a49a3a53beb2a927a6add9170b98363a2304 Signed-off-by: Giuseppe Lavagetto <glavagetto@wikimedia.org> | 11 February 2015, 16:30:03 UTC |
294d323 | QChris | 11 February 2015, 09:23:43 UTC | Revert "Temporarily keep 40 instead of 31 days of webrequest data" The jobs that needed to run on the older data finished. This reverts commit 2e4bb3eb014da4475f89525b6a9773dee7e3ac32. Change-Id: I3241766a4f0350ae9825faa8dd0b25a6dde3e019 | 11 February 2015, 15:53:08 UTC |
330f6c1 | cpettet | 11 February 2015, 14:45:26 UTC | phab fail back to mysql search for now refs T75854 requires bin/search index --all --background Change-Id: I54d3da767cf9b2d428f1c27522914af3c994bd30 | 11 February 2015, 14:45:26 UTC |
f56208a | cpettet | 11 February 2015, 14:26:08 UTC | phab local.json should trigger a phd restart * convert phd to a init style service (hopefully should manage fine with /etc/init.d link) * tell template to refresh phd on change Change-Id: Ie393e690c644467d66163e652257f19d170ac84f | 11 February 2015, 14:27:16 UTC |
62624a0 | cpettet | 11 February 2015, 14:14:29 UTC | phab link phd to /etc/init.d for service management Change-Id: I0cb9b07d99d6a2ed88dfc7e578bab7bb9dc64149 | 11 February 2015, 14:19:13 UTC |
877f619 | Sean Pringle | 11 February 2015, 06:08:27 UTC | clean up basic monitoring checks for all mariadb roles Change-Id: Ic166a6eba156da8ea3960b2172ca46f00a208a2e | 11 February 2015, 06:08:27 UTC |
db69bd6 | Sean Pringle | 11 February 2015, 05:23:43 UTC | incorrect private variable reference, mysql_repl_pass Change-Id: Ib4520de2af42c2d293f9e8e8cffb9cc5e77cae60 | 11 February 2015, 05:23:43 UTC |
099fdce | Ori Livneh | 11 February 2015, 03:11:39 UTC | Merge "vbench: on chrome crash, try to continue rather than bailing" into production | 11 February 2015, 03:11:39 UTC |
3c35bc0 | Ori Livneh | 11 February 2015, 03:10:01 UTC | vbench: on chrome crash, try to continue rather than bailing Change-Id: I5474d4be4cac95359ab6ddfa418d1e10ae05685b | 11 February 2015, 03:10:01 UTC |
1eb8152 | Daniel Zahn | 11 February 2015, 02:47:05 UTC | Merge "randomize times the planet cron jobs run" into production | 11 February 2015, 02:47:05 UTC |
0c38025 | Daniel Zahn | 11 February 2015, 02:34:31 UTC | randomize times the planet cron jobs run Ensure the cronjobs that update the planet language versions don't all run at the same minute to even out load. Use fqdn_rand() but use $title as seed to get different numbers for each cron even when they are on the same node. I don't have evidence that this really affected Etherpad on the same node in any way, but the linked bug made me do this anyways, just in case. Bug:T89174 Change-Id: I8bc7b5c2231faa6567e2721d3baa0ed82ff80692 | 11 February 2015, 02:42:18 UTC |
2a5c366 | Ori Livneh | 11 February 2015, 02:30:21 UTC | vbench: log console messages; abort when target crashes * On Inspector.targetCrashed, print summary statistics and bail. * Forward console.{log,debug,warn,...} messages from JS to Python. * Add --verbose arguments for debug logging. Change-Id: I27e0828ed3ffcd4895fe04720dba43f82d634740 | 11 February 2015, 02:32:26 UTC |
aaa8d7d | andrewbogott | 11 February 2015, 01:53:14 UTC | Merge "Switch to the 2.4-style 'Require host' syntax" into production | 11 February 2015, 01:53:14 UTC |
4e0cce0 | andrewbogott | 11 February 2015, 01:42:35 UTC | Switch to the 2.4-style 'Require host' syntax Change-Id: I02a255932541a8f479688d292ebbd18f4602662d | 11 February 2015, 01:49:00 UTC |
7876c3a | Marius Hoch | 11 February 2015, 01:17:22 UTC | role::mariadb: Remove references to undefined $shard Also use variable paths in labs.my.cnf.erb instead of hard coded ones. Change-Id: Ia5b49790d534452a49ed503104309922c068832b | 11 February 2015, 01:28:09 UTC |
e713939 | andrewbogott | 11 February 2015, 01:27:18 UTC | Merge "Allow silver to query its own server status." into production | 11 February 2015, 01:27:18 UTC |
383498d | andrewbogott | 11 February 2015, 01:24:30 UTC | Allow silver to query its own server status. Change-Id: I7c7bb18e2d70276500f45ec76b53b4a1e8ff67a3 | 11 February 2015, 01:24:30 UTC |
3a6af2d | Sean Pringle | 10 February 2015, 23:40:46 UTC | remove labsdb100[123] hacked role. use the mariadb config class. Change-Id: I478ad2a7cb8a4479207610ae95d4ac41fa674b78 | 10 February 2015, 23:46:37 UTC |
bfb29ce | Brandon Black | 10 February 2015, 22:04:51 UTC | repool cp1064 upload backend cache (testing experimental kernel) Change-Id: Id3f3e73849f0b52f72858a7900b1ad17952e824a | 10 February 2015, 22:05:25 UTC |
694ca92 | andrewbogott | 10 February 2015, 19:53:16 UTC | I guess we need these target dirs if we're going to backup to them. Change-Id: I336b26fb788efb2b23c55050b4a6053d23a0efb3 | 10 February 2015, 21:09:02 UTC |
48928e3 | Brandon Black | 10 February 2015, 20:40:51 UTC | downsize upload frontend mallocs to 1/12 This effectively undoes the effect of f7c47353 on the upload frontend caches; in this case backend disk cache is probably preferably to more frontend malloc. Change-Id: Ibe79365a02f9a3a0188efae19c4fbe837605d77e | 10 February 2015, 20:41:25 UTC |
92136d2 | Brandon Black | 10 February 2015, 20:33:47 UTC | jessie vm tuning: back to more-conservative-ish Change-Id: I6498186042b6249c2b45d6ab9754695f4f0e8673 | 10 February 2015, 20:33:47 UTC |
42e2f26 | cmjohnson | 10 February 2015, 20:09:03 UTC | Merge "Fixing dhcpd entry for mc1017/18" into production | 10 February 2015, 20:09:03 UTC |
caae251 | cmjohnson | 10 February 2015, 20:07:37 UTC | Fixing dhcpd entry for mc1017/18 Change-Id: I89559ea8b7af790d8e1372a6026cb18312df3c7f | 10 February 2015, 20:07:37 UTC |
819bf67 | andrewbogott | 10 February 2015, 19:46:08 UTC | Merge "Add a backup cron to the nova database class." into production | 10 February 2015, 19:46:08 UTC |
8835c2e | andrewbogott | 10 February 2015, 19:41:05 UTC | Add a backup cron to the nova database class. This happened as a side-effect of the wikitech backup when wikitech and the nova controller were on the same node. Now that they're separate, we still want to backup the nova dbs. Change-Id: If0de1102491c8af41728b3cedbbbc0075cde5020 | 10 February 2015, 19:42:22 UTC |
df3bff6 | Dzahn | 10 February 2015, 19:15:38 UTC | Merge "Change Blocking Tasks to 'Blocked By' Tasks" into production | 10 February 2015, 19:15:38 UTC |
2067fbd | Brandon Black | 09 February 2015, 23:17:41 UTC | fix PFS key rotation issues This reloads the ticket keys on the legacy precise stack with a daily cron, and eliminates them on the new jessie stack. Change-Id: I0fd68564a608d6cc5aafbd653a72350702abf31b | 10 February 2015, 18:21:05 UTC |
ef2a20d | cmjohnson | 10 February 2015, 18:10:35 UTC | Merge "Adding dhcpd entries for mc1018-8" into production | 10 February 2015, 18:10:35 UTC |
5a45ffa | cmjohnson | 10 February 2015, 18:09:17 UTC | Adding dhcpd entries for mc1018-8 Change-Id: I37502b446f54a51bffdc9a70065a5b721dd1a5b2 | 10 February 2015, 18:09:17 UTC |
dad6d71 | andrewbogott | 10 February 2015, 18:04:41 UTC | Merge "Change these mw config files to mwdeploy:mwdeploy to avoid scap confusion" into production | 10 February 2015, 18:04:41 UTC |
bca2457 | andrewbogott | 10 February 2015, 18:01:22 UTC | Merge "Revert "Temporary hack: Turn off wikitech-static dump crons."" into production | 10 February 2015, 18:01:22 UTC |
a5c79cc | andrewbogott | 10 February 2015, 17:59:15 UTC | Merge "Remove the nova::manager class from virt1000." into production | 10 February 2015, 17:59:15 UTC |
530468d | andrewbogott | 10 February 2015, 17:46:06 UTC | Change these mw config files to mwdeploy:mwdeploy to avoid scap confusion Change-Id: Ie57d4130731a7ba4fba3879262ab413a4f9034fe | 10 February 2015, 17:46:06 UTC |
1b08881 | Ariel T. Glenn | 10 February 2015, 17:33:23 UTC | snapshots: move the admin include into the role Change-Id: Icc9aba260bd8b10f82e483492953af38b30d31b6 | 10 February 2015, 17:33:23 UTC |
61bca3a | Marius Hoch | 07 August 2014, 14:39:09 UTC | Allow "hoo" to sudo into datasets Needed to keep and eye on and restart the Wikidata json dumps (and the future rdf and property dumps). This is done by introducing a new group snapshot-admin which allows me to sudo into the "datasets" user. RT: 8286 Bug: T86808 Change-Id: I1863679d72aeb4ab9aa527ec24bcbda290a8849f | 10 February 2015, 17:09:28 UTC |
c1921f5 | andrewbogott | 10 February 2015, 03:21:17 UTC | Revert "Temporary hack: Turn off wikitech-static dump crons." This reverts commit ae5643f84e86242739b30785cb1b31b450a21178. Now we have a dedicated wikitech host; maybe we can get wikitech-static up to date again. Change-Id: I27d09cdff30165be87ae0473bf0f3eff0ed1b720 | 10 February 2015, 13:42:17 UTC |
8b3b4f9 | andrewbogott | 10 February 2015, 03:25:10 UTC | Remove the nova::manager class from virt1000. This is now silver's job. Change-Id: I51ab695e74c6763aa6248882dcf4961d445c6a7a | 10 February 2015, 13:41:11 UTC |
1748e29 | Alexandros Kosiaris | 10 February 2015, 13:59:43 UTC | Merge "apparmor: minor lint fix" into production | 10 February 2015, 13:59:43 UTC |
fc271cd | Alexandros Kosiaris | 10 February 2015, 13:56:21 UTC | Merge "Update cassandra submodule" into production | 10 February 2015, 13:56:21 UTC |
02d63b0 | andrewbogott | 10 February 2015, 13:33:01 UTC | Add silver to the deployment train. Change-Id: Ib98fbd3ca15156aff25d624df5edfb45cd67d350 | 10 February 2015, 13:39:02 UTC |
b4e1ce8 | Matanya Moses | 10 February 2015, 12:40:24 UTC | apparmor: minor lint fix Change-Id: I7d544879d612b4ea4876df0139c24178e7165ea1 | 10 February 2015, 12:40:24 UTC |
e568e9a | Alexandros Kosiaris | 10 February 2015, 12:28:05 UTC | Merge "Bump cassandra memory on restbase test cluster" into production | 10 February 2015, 12:28:05 UTC |
4be85cf | Filippo Giunchedi | 10 February 2015, 11:45:33 UTC | introduce partman recipe for dm-cache also provision graphite1002 with it Bug: T88992 Change-Id: Icc8ebbc2c8d70507044c52a0b85f5ba3d80ce7e7 | 10 February 2015, 11:47:34 UTC |
441f499 | Giuseppe Lavagetto | 05 February 2015, 08:38:10 UTC | mediawiki: do not escape urls in the catchall redirect to https As specified in T88359, our APIs use url-encoded paths, so we need to use the NE flag in order not to break the redirect to HTTPS in such cases, whenever the redirect is a catchall. Change-Id: I31061fa2b1cf18bdea149d66d07e3e905fb31b9a Signed-off-by: Giuseppe Lavagetto <glavagetto@wikimedia.org> | 10 February 2015, 10:50:46 UTC |
f4c5a2e | Filippo Giunchedi | 10 February 2015, 10:46:59 UTC | install graphite1002 with jessie Bug: T88994 Change-Id: I3ba45be2f4f5e97d1b9f5ce91af4c7ec34402fa1 | 10 February 2015, 10:47:56 UTC |
96966bd | Giuseppe Lavagetto | 10 February 2015, 08:45:32 UTC | Revert "mediawiki: send .phtml files to HHVM as well" Alas, for some (obscure) reason, HHVM is extremely slow in processing /w/wiki.phtml requests, and it also gives wrong responses, so I am reverting this change and implement it as a RewriteRule instead. CFR https://phabricator.wikimedia.org/P276 This reverts commit ca057afc663a5dc2ceed32c1a89467cd006472c8. Change-Id: Ia0e101574441a8826ad04f4b867e97cb9c3055f3 | 10 February 2015, 08:45:32 UTC |
ca057af | Giuseppe Lavagetto | 09 February 2015, 08:38:41 UTC | mediawiki: send .phtml files to HHVM as well As noted in T88942 we still have a PHP file that's called "wiki.phtml" in the mediawiki repository, thus we have to catch cases of people trying to access that directly and send them to HHVM. Change-Id: Ia418bc129ad6e5748623128a637ef70435e76c43 Signed-off-by: Giuseppe Lavagetto <glavagetto@wikimedia.org> | 10 February 2015, 08:23:15 UTC |
1212b10 | andrewbogott | 10 February 2015, 02:44:20 UTC | Merge "Give wikitech (wherever it is) access to keystone and nova services." into production | 10 February 2015, 02:44:20 UTC |
0ba4d62 | andrewbogott | 10 February 2015, 02:38:46 UTC | Give wikitech (wherever it is) access to keystone and nova services. Change-Id: Ib6a0baae93e3a455ac6fc0288f6d81e039fc0179 | 10 February 2015, 02:38:46 UTC |
1a31ade | Ori Livneh | 10 February 2015, 00:54:58 UTC | vbench: don't trip click handler Change-Id: If230a96729bc36fb163297b0a0d7012fdaa07a2b | 10 February 2015, 00:54:58 UTC |
7962c87 | Ori Livneh | 10 February 2015, 00:47:35 UTC | vbench: pass all args from vb to vbench Change-Id: I17b9d0597f88045f8f476d21e7ce82db0d73cd5d | 10 February 2015, 00:47:35 UTC |
560ae29 | Ori Livneh | 10 February 2015, 00:34:06 UTC | Add missing file to I28ae265729 Change-Id: I2cf554d1b9e0c08e54eb1f321a1db2ab34637ab9 | 10 February 2015, 00:34:06 UTC |
46d7dc9 | Ori Livneh | 10 February 2015, 00:31:18 UTC | vbench: add 'vb' launcher Runs Chromium as part of the test suite, rather than using a global Chromium instance. Change-Id: I28ae265729171734b0efbc4eacaec6c8775a2551 | 10 February 2015, 00:31:18 UTC |
77c6203 | Ori Livneh | 10 February 2015, 00:01:18 UTC | vbench: tidy up string formatting code Change-Id: Iff118b42bdff1a0be941b45038c3a93842cf3ff7 | 10 February 2015, 00:01:18 UTC |
09a60df | Ori Livneh | 09 February 2015, 23:41:27 UTC | vbench: use 24-hr clock in timestamps Change-Id: I4ab6c83834111d49a5100881938ec709aaad833c | 09 February 2015, 23:41:46 UTC |
0efda1d | Daniel Zahn | 09 February 2015, 23:00:11 UTC | Merge "admin: add group for benchmarking with chromium" into production | 09 February 2015, 23:00:11 UTC |
c37c69f | Daniel Zahn | 09 February 2015, 22:46:41 UTC | admin: add group for benchmarking with chromium Adds an admin group for people who run benchmarking tests with chromium (currently on node osmium). This is for the request Bug: T89038 so Roan is the only member for now. Control of the services chromium and xvfb (display server) is needed. Change-Id: I96d07556ecaafdefdf45676efc1bcac79ef3b314 | 09 February 2015, 22:57:38 UTC |
97dcdc8 | Ori Livneh | 09 February 2015, 22:36:18 UTC | vbench: make it easier to log to a file When stdout is not a tty, strip color codes, and output logs to stdout and stderr. Change-Id: I04a296486f08ff4c9d4d68b4cc625bddc1a0aa4e | 09 February 2015, 22:41:16 UTC |
9b4b63b | Marius Hoch | 09 February 2015, 21:16:00 UTC | Add hoo to the "dba" and "wikidata" contactgroups So that I get notified when eg. Wikidata goes crazy... Change-Id: I6455b17a629ac571a6c02a7a2f5646abc50317c2 | 09 February 2015, 21:55:22 UTC |
6b91501 | Marius Hoch | 09 February 2015, 19:17:13 UTC | Fix template name typo "paresercache" -> "parsercache" Change-Id: I68053741f89701a1195c0e2700ca44b2d6a8e3ae | 09 February 2015, 21:25:38 UTC |
4987f33 | Merlijn van Deen | 08 February 2015, 14:52:14 UTC | Change Blocking Tasks to 'Blocked By' Tasks Bug: T33 Change-Id: I89f3c6b9f2e09b3b405fd088f78c18461f206ea8 | 09 February 2015, 21:06:30 UTC |
494b04d | Dzahn | 09 February 2015, 20:37:53 UTC | Merge "Change 'Export to Excel' to 'Export (disabled)'" into production | 09 February 2015, 20:37:53 UTC |
d6f2565 | Dzahn | 09 February 2015, 20:35:41 UTC | Merge "Icinga: Drop qchris from analytics contactgroup" into production | 09 February 2015, 20:35:41 UTC |
b319e5c | Antoine Musso | 05 February 2015, 13:24:51 UTC | puppet-lint: ignore some var in single quoted strings We have some use of puppet style variables in single quote string. They are not meant to be interpreted by puppet itself but still cause puppet-lint to complains about it. Use puppet-lint control comments [1] to ignore the few occurences we have. That fix the last puppet-lint error we had in the repository: $ puppet-lint --error-level=error .` $ [1] http://puppet-lint.com/controlcomments/ Bug: T87132 Change-Id: Id00e0a7844a539b7bf3aeeeb22b0887c34326693 | 09 February 2015, 20:07:40 UTC |
9cca621 | Andrew Bogott | 09 February 2015, 20:07:05 UTC | Merge "Move puppet-lint options to .puppet-lint.rc" into production | 09 February 2015, 20:07:05 UTC |
457ee2f | Ori Livneh | 09 February 2015, 19:30:49 UTC | vbench: remove debugging code Change-Id: I211fefd4374b34f70c07abf5e491de15143e2336 | 09 February 2015, 19:30:49 UTC |
0240240 | Ori Livneh | 09 February 2015, 19:22:49 UTC | vbench: use defer.inlineCallbacks to chain commands Fixes twisted.internet.defer.AlreadyCalledError exceptions. Change-Id: I5912cd03e8f3a5b179a9fe95e4e7e242febed2c7 | 09 February 2015, 19:22:49 UTC |
62dd63b | Gabriel Wicke | 09 February 2015, 19:01:27 UTC | Bump cassandra memory on restbase test cluster 8g performs better than 5g. It is also the typical config for large-memory cassandra boxes, as larger JVM heaps result in unacceptably long GC pauses. Change-Id: I19d84072cce947cff8df2c1045133a0492700cb0 | 09 February 2015, 19:01:27 UTC |
5476718 | Gabriel Wicke | 09 February 2015, 18:55:19 UTC | Update cassandra submodule Bug: T88956 Change-Id: I405386417248c057976472712513fd7fd587000a | 09 February 2015, 18:55:19 UTC |
63e4919 | Ori Livneh | 09 February 2015, 18:45:22 UTC | vbench: create domain proxy objects for more python calling conventions Before: self.sendCommand('Page.navigate', url=url) After: self.page.navigate(url=url) Change-Id: I77acc49a15d09a304887838c864e1eb99030fb10 | 09 February 2015, 18:45:22 UTC |
2ebb3b1 | Marko Obrovac | 09 February 2015, 13:57:14 UTC | vbench: Fix minor bug in std() std(), introduced in https://gerrit.wikimedia.org/r/#/c/189305/ , calculates the standard deviation of a given data set. However, there was a small bug using an undefined variable - 'c' - rendering the calculation incorrect. This path fixes it by using the mean value computed in the previous step. Change-Id: I32ed2ab16da7b952e61f6d4059b46691b33509c8 | 09 February 2015, 18:06:04 UTC |
6e84030 | andrewbogott | 04 February 2015, 19:18:01 UTC | Rename cassandra-roots to cassandra-test-roots; add mobrovac and jdouglas. Bug T85492 Change-Id: Iae2aee82d742a3eee949aca12f11cbbb9ccb6a3c | 09 February 2015, 17:49:56 UTC |
696e54d | Filippo Giunchedi | 09 February 2015, 16:25:02 UTC | gdash: fix graphite disk dashboard sda->md1 Bug: T85909 Change-Id: I5b02c11a4da9e8225c1a9ac29d3782ca22515819 | 09 February 2015, 16:25:18 UTC |
adc9cc5 | Christian Aistleitner | 09 February 2015, 15:54:41 UTC | Icinga: Drop qchris from analytics contactgroup Change-Id: I9af6798e084b693bec67d674869986d99a176ab1 | 09 February 2015, 15:54:41 UTC |
2e0a4b9 | Ariel T. Glenn | 09 February 2015, 15:31:55 UTC | dumps nginx: fixup ca cert name Change-Id: I332d9593f443b550a6e5d46673e87badbcc908ad | 09 February 2015, 15:31:55 UTC |
fc88db8 | Alexandros Kosiaris | 09 February 2015, 13:11:12 UTC | Merge "Fix regexp for wikidata icinga check." into production | 09 February 2015, 13:11:12 UTC |
0eff584 | Jan Zerebecki | 09 February 2015, 13:00:06 UTC | Fix regexp for wikidata icinga check. The anchor at the the end was missing, so it accepted any lag. Now it alerts at over 59. Bug: T88980 Change-Id: I74f24afbfac35feb8de97f917204e48d28ea8a8f | 09 February 2015, 13:00:06 UTC |
44276c9 | dzahn | 05 February 2015, 01:03:28 UTC | redisdb: add ferm::service for redis-server Add firewall hole for redis-server on rbf hosts. On the new hosts in codfw we could start out with base::firewall right away (If12b08e406edf3d) On the old hosts we eventually can after checking what else we need to open. This is checked against rbf1001. Bug:T86898 Change-Id: I0957b96b99d525df690d406c3162929bcc376d5c | 09 February 2015, 11:40:35 UTC |
484e196 | Ariel T. Glenn | 09 February 2015, 11:19:58 UTC | dumps nginx: enable ipv6 Change-Id: I9e6dde249df2c893c186132f1f4b5e0272ca9912 | 09 February 2015, 11:19:58 UTC |
e35cf05 | Ariel T. Glenn | 09 February 2015, 10:36:09 UTC | dumps nginx, fix up limit_conn_zone directive Change-Id: Ie3d2e35830ccb509e57da58d5ea4cf6a38fd2587 | 09 February 2015, 10:36:42 UTC |
e357382 | Ariel T. Glenn | 09 February 2015, 10:27:39 UTC | connection limits in nginx need shared memory and key defn Change-Id: I89a2ff33899a403d8d36921b1e04d03f46322fec | 09 February 2015, 10:27:39 UTC |
e36ce54 | Ariel T. Glenn | 09 February 2015, 09:44:07 UTC | reinstate bandwidth and conn caps for dumps.wm.org turns out that the pipe can handle more than the disks. values may need to be tweaked. Change-Id: I7076b65d30b7abf24375816e3dd30875854485c3 | 09 February 2015, 10:14:09 UTC |
2a504e7 | Ori Livneh | 08 February 2015, 07:12:03 UTC | vbench: various improvements * Use deferreds to avoid races. * Re-write some argument names to be clearer. * Avoid depending on numpy by having local mean / median / std functions. * Get rid of warm-up loops. * Run 5 loops by default. Change-Id: I4249a4eb156a23890887c83735b6b15c660692ed | 09 February 2015, 04:50:41 UTC |
e209e5e | Merlijn van Deen | 08 February 2015, 14:43:41 UTC | Change 'Export to Excel' to 'Export (disabled)' Once PHPExcel is available, we can drop the '(disabled)'. Bug: T152 Bug: T257 Change-Id: I7a08bcca727d9732eaed353137ceb44064a23d70 | 08 February 2015, 14:48:43 UTC |
2e7d329 | Faidon Liambotis | 07 February 2015, 21:38:09 UTC | icinga: remove mark from SMS Change-Id: Ib302a9bb99747718f5b7720582964165fa77dae8 | 07 February 2015, 21:38:09 UTC |
a72490a | Brandon Black | 07 February 2015, 10:32:20 UTC | depool cp1064 upload backend Change-Id: I5101dc0745f6b856a2c7bff20106b35eaa2cd329 | 07 February 2015, 10:32:32 UTC |
f7af4cc | Faidon Liambotis | 07 February 2015, 09:53:34 UTC | reprepro: switch Cassandra to 2.1, add to jessie Switch our update configuration to fetch Cassandra 2.1. Add the cassandra source to jessie-wikimedia and remove from {precise,trusty}-wikimedia. Task: T88850 Change-Id: I09793b47a6758c8a8081c1c340122c7c470ff918 | 07 February 2015, 09:54:32 UTC |
e280b02 | Brandon Black | 07 February 2015, 09:44:24 UTC | depool cp1070 in cache.pp: T88889 Change-Id: I41e2a4d763ac463706764dce6286a3fba9b4cb68 | 07 February 2015, 09:44:24 UTC |
ae8101b | Ori Livneh | 07 February 2015, 02:41:30 UTC | vbench improvements * Encapsulate profile data processing in a new ChromeCPUProfiler class. * Use Twisted log observer to log debug data. * Add --latency and --hidpi options. Change-Id: I856c68256f08fd6ef944eabec6bad74e6d3c2ebd | 07 February 2015, 02:58:04 UTC |
c7c21a8 | Christian Aistleitner | 07 February 2015, 00:44:16 UTC | Bump cdh module to increase heap on name nodes Bug: T88871 Change-Id: Id1b599a599330ab946e79535c290bfdaedba8682 | 07 February 2015, 00:52:40 UTC |
211ea1f | RobH | 06 February 2015, 19:59:18 UTC | reclaiming ms1004 to spares ms1004 was misc ganglia aggregator, so changed it to bast1001. removed all ms1004 entries, as it will not use this hostname in the future. T86933 Change-Id: I5447f18f26f815007acb3962c6bbc76150f22bc6 | 06 February 2015, 20:03:33 UTC |
8642f89 | YuviPanda | 06 February 2015, 19:19:30 UTC | toollabs: Fix webservice2 restart when no webservice is running Change-Id: Ia3ba48aadd6ca779902cfd08fd319d6a6942ce00 | 06 February 2015, 19:24:24 UTC |
75d044a | Gabriel Wicke | 06 February 2015, 01:08:03 UTC | Update the restbase config for v0.4.0 RESTBase is now full spec-driven, so the config file has grown a bit. Change-Id: Ibff24a8a59e17515a31e942ed0a0c1af6d70d181 | 06 February 2015, 16:29:17 UTC |
9fed3ff | Brandon Black | 06 February 2015, 14:52:15 UTC | more-aggressive vm tuning for jessie-varnish Change-Id: I3a7a3e685bcd20009f6fb972aecb3033e85f2f41 | 06 February 2015, 14:54:43 UTC |
38c7ac4 | Faidon Liambotis | 06 February 2015, 14:11:12 UTC | sysctl: brown-paper bag fix for service's Exec Change-Id: I41ca9a3a8daacf213943105415af7e20ac355d27 | 06 February 2015, 14:11:12 UTC |
0d9bf6f | Faidon Liambotis | 29 January 2015, 02:00:07 UTC | sysctl: make service call init system-agnostic Remove an onlyif that checked for an upstart unit, that was a remnant from a workaround for Ubuntu hardy boxes. Switch "start" to a "service" call so that it works equally under all init systems. Change-Id: I0952cb041036ff2bce4e3007c272e009e7780c74 | 06 February 2015, 14:04:40 UTC |
97f9663 | Alexandros Kosiaris | 06 February 2015, 12:46:51 UTC | Merge "Disable LWP SSL hostname verification" into production | 06 February 2015, 12:46:51 UTC |