[09:35:56] 06Traffic, 06MediaWiki-Platform-Team (Radar), 13Patch-For-Review, 07User-notice: [Main Rollout] Enable unified mobile routing on remaining wikis - https://phabricator.wikimedia.org/T403510#11245411 (10Krinkle) >>! https://gerrit.wikimedia.org/r/1190759 **merged**: > %%%varnish: Enable unified mobile routin... [09:36:13] 06Traffic, 06MediaWiki-Platform-Team (Radar), 13Patch-For-Review, 07User-notice: [Main Rollout] Enable unified mobile routing on remaining wikis - https://phabricator.wikimedia.org/T403510#11245412 (10Krinkle) [11:52:57] ATS lua question, how bad is it to have a big lua.conf file ? [12:32:44] claime: #define big [12:32:58] how big are we talking aobut? [12:33:00] *about [12:33:25] vgutierrez: I'm going for a different approach in the end, but what I was talking about was a a table with potentially every domain in group0/group1/group2 in it [12:34:09] vgutierrez: What I'll be doing instead is switching to a regex-based match instead of a litteral keymatch, and generating regexes on my side to add to the config file [12:34:12] .lua.conf files are effectively lua code that needs to be evaluated [12:34:27] Which would shorten the size significantly at the cost of doing a looped regex match [12:35:43] it would be nice to be able to run some benchmark [12:36:03] cause basically the question here is having a bigger table with literal matching or a shorter table with regex matching [12:36:05] so the host_litteral match would mean a 1k+ elements hash, each element being a two elements hash for now [12:36:50] The regex one would be significantly shorter but in file size, but means more computation [12:36:54] vgutierrez: exactly [12:38:25] maybe we can time the test suite to see what's better time-wise, but it won't tell the story on memory consumption [12:38:40] I don't know enough about lua profiling to know if there's a better way [12:41:44] claime: intuition says that regexes ordered by match probability should be enough [12:41:58] and it will be probably easier to maintain than a table of 1k elements [12:42:06] yes to both [12:42:11] even if regexes are tricky on their own :) [12:42:23] vgutierrez: they're going to be very naive regexes [12:43:25] basically taking the list of hosts in a group dblist, grouping by domain (wikibooks.org,wikiquote.org,etc.) and then creating an or-group of the subdomain [12:43:44] so something like (aa|ak|ang|ast|as|ay|bi|bo|ch|co|ga|gn|got|gu|ie|kn|ks|lb|ln|lv|mi|mn|my|nah|na|nds|ps|qu|se|su|sw|tk|ug|uz|vo|wa|xh|yo|za|zu)\.wikibooks\.org [12:43:55] for every domain [12:44:34] and or-grouping that for one rule-match per group [12:45:31] I'm not seeing a non-capture group pattern item for lua regex, am I right about that? [12:49:25] yeah... lua pattern matching isn't as powerful as true regex support [13:41:44] hello traffic! I was going to merge https://gerrit.wikimedia.org/r/c/operations/puppet/+/1191708 and child shortly [13:42:13] cdanis: thanks, no deployments from us AFAIK [13:42:40] vgutierrez: lol lua patterns don't have logical "or"... back to the drawing board [13:43:03] claime: is this lua within haproxy or within ats? [13:43:12] ATS [13:43:13] cdanis: this is ats lua [13:43:20] ah okay should have known from dblist [13:43:21] but i don't think that matters? [13:43:28] (in this specific case0 [13:43:37] claime: well in haproxy you get access to pcre :V [13:43:46] cdanis: hah [13:44:12] so yeah, without a logical "or" I don't see a way to implement that actually [13:44:32] At least not without exploding the complexity [14:02:28] 06Traffic, 06Infrastructure-Foundations, 06SRE, 10vm-requests: eqiad: 2 VM request for hCaptcha - https://phabricator.wikimedia.org/T406166#11246214 (10MoritzMuehlenhoff) p:05Triage→03Medium [14:09:00] 06Traffic, 06Infrastructure-Foundations, 06SRE, 10vm-requests: codfw: 2 VM request for hCaptcha - https://phabricator.wikimedia.org/T406167#11246249 (10MoritzMuehlenhoff) p:05Triage→03Medium [14:37:36] 10netops, 06DC-Ops, 06Infrastructure-Foundations, 10ops-eqiad, 06SRE: Eqiad C/D refresh: 2 x test hosts for config validation - https://phabricator.wikimedia.org/T405560#11246363 (10cmooney) p:05Triage→03Low [14:48:22] 06Traffic, 06Infrastructure-Foundations, 06SRE: Updating Netbox for LVS hosts in eqiad lvs10(1[789]|20) - https://phabricator.wikimedia.org/T334884#11246433 (10cmooney) @ssingh I think we can probbaly close this one? Currently the hosts are representing reality (even if the process could be improved), and g... [14:51:00] 06Traffic, 06Infrastructure-Foundations, 06SRE: Updating Netbox for LVS hosts in eqiad lvs10(1[789]|20) - https://phabricator.wikimedia.org/T334884#11246439 (10ssingh) 05Open→03Resolved a:03ssingh Thanks for following up @cmooney. Other than the fact that we owed you a response on your last comment... [15:21:46] 06Traffic, 06Data-Engineering, 06Data-Engineering-Radar, 13Patch-For-Review: improved x-analytics data on Edge Uniques status - https://phabricator.wikimedia.org/T405783#11246578 (10CDanis) a:05CDanis→03JAllemandou joal please deploy [[ https://gerrit.wikimedia.org/r/c/analytics/refinery/+/1192577 | re... [15:25:36] 06Traffic, 06DC-Ops, 10ops-codfw, 06SRE, 13Patch-For-Review: Q4:rack/setup/install cp20[43-58] codfw - https://phabricator.wikimedia.org/T392851#11246593 (10elukey) I was able to update the firmware cookbook for IDRAC 10, and now we can do idrac+bios (still working on some issue with ssd, should be solve... [16:15:38] 06Traffic, 06DC-Ops, 10ops-codfw, 06SRE, 13Patch-For-Review: Q4:rack/setup/install cp20[43-58] codfw - https://phabricator.wikimedia.org/T392851#11246838 (10Jhancock.wm) cp2050/52: try logging into root and using the racadm command. That's one of the symptoms when it needs to be used. cp2056: this is th... [16:47:54] 06Traffic, 06Data-Engineering, 06Data-Engineering-Radar, 13Patch-For-Review: improved x-analytics data on Edge Uniques status - https://phabricator.wikimedia.org/T405783#11246994 (10JAllemandou) Ingestion job restarted with the new sttings. The new field should start to populate soon. [16:55:50] 06Traffic, 10Prod-Kubernetes, 06serviceops, 07Kubernetes, 13Patch-For-Review: Handling inbound IPIP traffic on low traffic LVS k8s based realservers - https://phabricator.wikimedia.org/T352956#11247042 (10ssingh) Hi @akosiaris: Following up on this after a discussion during Traffic's planning with @Vguti... [17:36:13] 06Traffic, 06DC-Ops, 10ops-codfw, 06SRE, 13Patch-For-Review: Q4:rack/setup/install cp20[43-58] codfw - https://phabricator.wikimedia.org/T392851#11247151 (10Jhancock.wm) @elukey give 2056 a shot when you get on next. might have fixed it. if not i might have to get dell involved again. [19:55:09] 06Traffic, 06Data-Engineering, 06Data-Engineering-Radar, 13Patch-For-Review: improved x-analytics data on Edge Uniques status - https://phabricator.wikimedia.org/T405783#11247628 (10CDanis) 05Open→03Resolved