[00:03:42] (CR) AndyRussG: Pare down email pref ctr config (2 comments) [wikimedia/fundraising/dev] - https://gerrit.wikimedia.org/r/683725 (owner: Ejegg) [00:03:47] (PS2) AndyRussG: Pare down email pref ctr config [wikimedia/fundraising/dev] - https://gerrit.wikimedia.org/r/683725 (owner: Ejegg) [00:13:07] (PS3) AndyRussG: Pare down email pref ctr config [wikimedia/fundraising/dev] - https://gerrit.wikimedia.org/r/683725 (https://phabricator.wikimedia.org/T268510) (owner: Ejegg) [00:13:48] ejegg: you shouldn't be working today :) but if you happen to see this and would say it's OK for me to copy that civiproxi api command from fundraising-dev to wmff/install.sh in buildkit, and then +2 that patch, and abandon the other, I could definitely do that..... ;) [00:23:05] oh sure, go ahead AndyRussG ! [00:23:15] I was about to do that myself and then the baby woke up [00:24:10] ejegg: okok thanks, yeah I got it, baby is important :) [00:28:47] (CR) AndyRussG: [C: +2] Pare down email pref ctr config [wikimedia/fundraising/dev] - https://gerrit.wikimedia.org/r/683725 (https://phabricator.wikimedia.org/T268510) (owner: Ejegg) [00:28:53] (CR) AndyRussG: [V: +2 C: +2] Pare down email pref ctr config [wikimedia/fundraising/dev] - https://gerrit.wikimedia.org/r/683725 (https://phabricator.wikimedia.org/T268510) (owner: Ejegg) [00:34:22] (PS4) AndyRussG: Delete drupal settings override [wikimedia/fundraising/dev] - https://gerrit.wikimedia.org/r/683995 (https://phabricator.wikimedia.org/T281609) (owner: Ejegg) [00:54:04] (CR) AndyRussG: [C: +2] "Works great, thanks!" [wikimedia/fundraising/crm/civicrm-buildkit] - https://gerrit.wikimedia.org/r/683994 (https://phabricator.wikimedia.org/T281609) (owner: Ejegg) [00:54:35] (CR) AndyRussG: [C: +2] "Wooohooo thanks for this!!" [wikimedia/fundraising/dev] - https://gerrit.wikimedia.org/r/683995 (https://phabricator.wikimedia.org/T281609) (owner: Ejegg) [00:54:37] (CR) AndyRussG: [V: +2 C: +2] Delete drupal settings override [wikimedia/fundraising/dev] - https://gerrit.wikimedia.org/r/683995 (https://phabricator.wikimedia.org/T281609) (owner: Ejegg) [01:14:20] (CR) AndyRussG: [V: +2 C: +2] Restore fredge db, move drupal extra settings [wikimedia/fundraising/crm/civicrm-buildkit] - https://gerrit.wikimedia.org/r/683994 (https://phabricator.wikimedia.org/T281609) (owner: Ejegg) [02:43:34] (PS2) Ejegg: Set wmff CIVI_SITE_KEY for Docker dev setup [wikimedia/fundraising/crm/civicrm-buildkit] - https://gerrit.wikimedia.org/r/684121 (https://phabricator.wikimedia.org/T281651) (owner: AndyRussG) [02:44:15] ejegg: ah btw getting an error on the fredge schema thing (already merged) [02:44:36] only happens when I run buildkit without having reset the persistent storage [02:44:39] oh hmm, let me see if I have any more changes locally that were making it work [02:44:42] ohhh [02:44:46] DatabaseSchemaObjectExistsException: Table payments_initial already exists. in /srv/civi-sites/wmff/drupal/includes/database/schema.inc:731 [02:44:59] i was always resetting [02:45:09] yeah me too [02:45:20] I think it just needs a "if exists" somewhere [02:45:21] ah, probably buildkit drops all the tables in civicrm.* and drupal.* ? [02:45:36] ah also yes true I think [02:46:12] hmm, that 'if exists' would be in drupal 7 core code [02:46:31] ejegg: here's the full error: https://paste.toolforge.org/view/e091f048 [02:48:14] if buildkit really is blowing away all those tables, I guess the solution is to have the wmff config also blow away the fredge.* tables when it starts [02:48:53] oh, and we should make a PR upstream for anything we locally merge to buildkit [02:49:10] and then pester eileen to merge them :) [02:49:24] so stuff doesn't break when we switch to upstream [02:50:20] ejegg: oh oooops [02:50:35] oops? [02:50:51] yes ooops [02:51:03] * AndyRussG merged [02:51:35] that's fine for now, just so long as we get those things upstream before we switch repos [02:52:41] ah hmmmm okok [02:56:23] K I can add that in the same place in install.sh [02:56:41] cool cool! [02:56:46] k, see you later [03:13:48] (PS1) AndyRussG: WMFF: Drop fredge db before creating it [wikimedia/fundraising/crm/civicrm-buildkit] - https://gerrit.wikimedia.org/r/684138 (https://phabricator.wikimedia.org/T281609) [03:30:31] (PS3) AndyRussG: WMFF: sync drupal users and set API key for Docker dev setup [wikimedia/fundraising/crm/civicrm-buildkit] - https://gerrit.wikimedia.org/r/684003 (https://phabricator.wikimedia.org/T281651) (owner: Ejegg) [03:49:32] (PS4) AndyRussG: WMFF: sync drupal users and set API key for Docker dev setup [wikimedia/fundraising/crm/civicrm-buildkit] - https://gerrit.wikimedia.org/r/684003 (https://phabricator.wikimedia.org/T281651) (owner: Ejegg) [03:52:31] (PS3) AndyRussG: Set Civicrm API key for Civiproxy [wikimedia/fundraising/dev] - https://gerrit.wikimedia.org/r/684006 (https://phabricator.wikimedia.org/T281651) (owner: Ejegg) [05:09:12] (PS3) AndyRussG: WMFF: Static CIVI_SITE_KEY for Docker dev setup [wikimedia/fundraising/crm/civicrm-buildkit] - https://gerrit.wikimedia.org/r/684121 (https://phabricator.wikimedia.org/T281651) [05:34:04] (PS4) AndyRussG: Set static CIVI_SITE_KEY for Civiproxy [wikimedia/fundraising/dev] - https://gerrit.wikimedia.org/r/684122 (https://phabricator.wikimedia.org/T281651) [07:20:21] Fundraising Sprint File Systems Stage Show, Fundraising Sprint Git Rebase Jump, Fundraising Sprint Humongous bacteria petting zoo, Fundraising Sprint Interstitial ads halfway down the coaster hill, and 2 others: Production of New Annual Fund Thank You Email... - https://phabricator.wikimedia.org/T278363 [08:12:49] PROBLEM - check_log_messages on frav1002 is CRITICAL: CRITICAL: Paypal_endpoint_critical 1 [=1] [08:17:49] RECOVERY - check_log_messages on frav1002 is OK: OK [16:03:25] Fundraising-Backlog: Spanish Latam donors seeing English survey - https://phabricator.wikimedia.org/T281731 (krobinson) [16:08:32] Fundraising-Backlog: Spanish Latam donors seeing English survey - https://phabricator.wikimedia.org/T281731 (DStrine) @krobinson would this be the survey on the TY page or something else? @KHaggard @MNoorWMF @TSkaff @Pcoombe FYI. Have there been any surveys emailed by the email team? Has the TY page survey... [16:14:59] Fundraising-Backlog: Spanish Latam donors seeing English survey - https://phabricator.wikimedia.org/T281731 (KHaggard) @DStrine TY page survey only - no surveys sent through email recently and no TY page survey updates recently. I've checked the es-419 and es TY pages and the survey links are correct and in... [16:22:32] Fundraising-Backlog: Spanish Latam donors seeing English survey - https://phabricator.wikimedia.org/T281731 (DStrine) Per slack this is the TY page survey. The donors have not mentioned that the Ty page is in english so it might just be the survey. Fr-tech can't do anything with this. It's possible survey mo... [16:25:43] Fundraising-Backlog, AbuseFilter, Analytics-Radar, BetaFeatures, and 46 others: Prepare User group methods for hard deprecation - https://phabricator.wikimedia.org/T275148 (AndyRussG) [17:53:14] howdy fr-tech: as of yet, do we have an idea of what the http requests to civiproxy will look like? i'm looking at possibilities for request regexes to make sure only the enpoints/queries we want hit the proxy layer. [17:54:38] dwisehaupt: we have that all set up in dev so it shouldnt be too hard to capture a couple examples [17:55:22] cool. not a rush on this but it would be good to get in from the start if possible. [17:56:07] Fundraising-Backlog, FR-Email: Civi: Contact records can be opted in even when on the Master Suppression List - https://phabricator.wikimedia.org/T274146 (KHaggard) @DStrine and @Eileenmcnaughton - Brian found a lot of things in the MSL purge, so I've summarized it a Google Doc for now. Please let me kno... [18:00:17] dwisehaupt ejegg https://gerrit.wikimedia.org/r/plugins/gitiles/mediawiki/extensions/DonationInterface/+/refs/heads/master/extras/civiproxy/CiviproxyConnect.php#18 [18:00:36] contact_id is numerical and hash is [0-9a-f] [18:03:38] Here is the civi API that eventually is hit: https://gerrit.wikimedia.org/r/plugins/gitiles/wikimedia/fundraising/crm/+/e08f34fba4db3c2e066833b538f62dd698c48445/drupal/sites/default/civicrm/extensions/wmf-civicrm/api/v3/Civiproxy/Getpreferences.php [18:50:42] thanks. i'll have a look. [20:22:53] fr-tech cstone here are the last patches in the chain for fundraising-dev, which should get you the civipreference setup automagically [20:23:46] for buildkit: https://gerrit.wikimedia.org/r/c/wikimedia/fundraising/crm/civicrm-buildkit/+/684121 [20:24:01] for fundraising-dev: https://gerrit.wikimedia.org/r/c/wikimedia/fundraising/dev/+/684122 [20:24:08] thanks AndyRussG [20:24:14] cstone: thank u :) [20:24:34] cstone fr-tech don't forget to do rm -rf src/civ-sites/wmff/vendor first [20:24:45] and also docker-compose down before updating fundraising-dev [20:25:07] after those steps, if you run it through setup.sh it should just work [20:25:37] or if you want to run the civi build step directly on the command line in the container, first do rm /srv/civi-sites/wmff/wmff.sh [20:25:46] and then you can do civibuild create wmff [20:27:12] you'll get the proper fredge database created with all that, too, btw [20:27:17] ahhhhh hopefully anyway ;p [20:40:11] woo AndyRussG got further than friday [20:40:42] ahhh hmmmm and.....? [20:41:42] still installing [20:41:48] ahhh okok good news! [20:42:01] dwisehaupt Jeff_Green I just confirmed locally that I can set Redis port on a per-queue basis [20:42:02] works fine [20:42:10] great! [20:42:14] yeah! [20:42:19] so I think no need to even make a task for that! [20:42:45] Jeff_Green dwisehaupt what I'm not clear on is who does what to make the email-pref-ctr-specific SmashPig config? [20:43:00] wellllllllll [20:43:26] Jeff_Green is it puppetized or in localsettings? hmmm I guess I can check heheh [20:43:28] ideally we'll be able to have that fully served by puppet [20:43:31] it is right now [20:43:38] Jeff_Green yeah that sounds reaosnable [20:44:10] it's in puppet, deployed to /etc/fundraising and f_c_u deploys a symlink to it [20:44:33] hmmm [20:44:52] Jeff_Green if you point me to the current puppet template I can propose a one for email-pref-ctr and put it in /tmp like e jegg did for localsettings [20:44:55] if it turns out there's something that has to be outside of puppet, I think there are other contexts where we source the /etc/fundraising file from one in localsettings [20:45:15] I don't think there is anything [20:45:22] ok on sec [20:45:57] also I don't know how the symlink works, but basically the requirement is just that it be different for the boxes that run e-p-c [20:46:23] there's a template that is specific to donorwiki in the puppet tree [20:46:41] puppet/modules/fundraising/templates/donorwiki/LocalSettings.php.erb [20:47:10] Jeff_Green: but for SmashPig, it's like a main.yaml file that ends up in /etc/smashpig [20:47:25] looking [20:48:58] the regular smashpig config is puppetized, but not (yet) deployed to frdata servers [20:49:52] we can configure it differently as needed, without any template modification we can tweak db/queue settings for it [20:50:31] that template is puppet/modules/fundraising/templates/smashpig/main.yaml.erb [20:51:31] if we need to vary the structure of the template otherwise that's pretty straightforward puppetwise [20:56:36] Jeff_Green: ok so we need to add another key after unsubscribe: [20:56:55] woohoo AndyRussG !!! fredge db and api key works [20:57:09] cstone ahhhhh fantasmic yaaaay [20:57:51] running to post office now ill be back [20:58:07] cstone though I'm not sure we should merge the fundraising-dev patches until the buildkit patches get through, which would be I guess more in eileen's corner? [20:58:21] Jeff_Green the new key will be 'email-preferences' [20:58:40] and instead of saying <<: *REDIS right below it, like the others [20:59:22] it'll be followed by entries that are similar to what's right under the top-level redis-servers key [20:59:25] ok [20:59:37] i.e. class: PHPQueue.... etc [20:59:53] and then with the e-p-c-specific port and password and such [21:00:23] does that make sense? [21:00:51] I think so? I'll try to mock it up [21:01:33] Jeff_Green: cool beans, thanks much! [21:02:17] Jeff_Green: I think potentially we can also leave off all those other queues, but I think also it won't make much difference, since they're there by default in SmashPig in any case, IIRC [21:03:05] what I may do is change the structure of the template to only include a stanza if there's config for it [21:04:15] i.e. if we don't specify fredge_db_reader we won't add the fredge-db data store block [21:04:44] Jeff_Green: right... ah also no provider-configuration-directory needed [21:04:50] or sequence-generator [21:27:20] AndyRussG: I'm a little unclear what the config would look like on the civicrm server [21:27:50] i.e. where we would want two queue server stanzas [21:28:50] Jeff_Green: so what worked locally is just adding email-preferences as a queue name, and putting the specific Redis config under that, instead of the <<: *REDIS backreference [21:29:32] can you put a copy of that config in tmp on frpm1001 or something? [21:29:58] Jeff_Green: yep one sec [21:31:11] Jeff_Green: oh I just noticed it already has an email-preferences key [21:31:18] just not where I was expecting it should go [21:31:56] oh, right, that must be for the unsubscribe stuff in paymentswiki [21:34:06] Jeff_Green: no I think someone just already added it? [21:34:16] there's the unsubscribe queue for that I think? [21:34:20] I might be wrong tho [21:34:33] Jeff_Green: so, see frpm1001:/tmp/new-draft-main.yaml.erb [21:37:03] AndyRussG: you're right, git says Dallas added it a few days ago [21:37:06] looking [21:37:19] ahh cool beans :) [21:37:58] ok I see [21:38:23] Jeff_Green: so also I added the new template parameters that I think would be needed? [21:38:26] like @queue_redis_epc_port etc [21:38:39] yup [21:38:50] :) [21:38:51] yeah. i added it in for T268512, but that was before we discussed the separate instance. [21:38:52] T268512: Add new queue settings for email-preferences - https://phabricator.wikimedia.org/T268512 [21:38:58] I'm on the fence about making two separate templates [21:39:28] Jeff_Green: welp we definitely don't want the provider directory on the new wiki [21:39:36] right [21:39:38] and it feels kinda odd to have that in the config pointing to nothing [21:39:50] though I guess technically it might not break anything [21:40:11] it's easy to leave out stuff like that in puppet, once you can envision all the variants you need [21:40:42] ah right I see [21:40:44] I'm not entirely clear on how this config file works though [21:41:00] the top stanza breaks my brain [21:41:03] so still two different configs, just created from a single template with different parameters [21:41:07] right [21:41:21] three, actualy [21:41:28] hmmmm [21:41:32] okok [21:41:36] payments/payments-listeners vs civicrm vs frdata [21:41:40] right [21:41:58] Jeff_Green: wrt the stanza, I'm also fuzzy on the details of how yaml backreferences or whatever they're called work [21:42:01] but basically it works [21:42:14] and that's about as far as I get [21:42:17] right [21:43:32] I'll see if I can figure it out, if I can standardize on "all the queue-server stuff goes in redis-servers: and all the queues get their config from there by a backreference, I think it will be easier/more logical to include/exclude stuff per server role [21:44:06] I'll have to work on this tomorrow though, it's late here! [21:44:39] Jeff_Green: okok.... just one quick note, definitely for the civicrm we need two different versions of the queue config [21:44:45] or at least two different ports, in a single file [21:45:07] yep! [21:45:10] since that's the single box that needs to consume queue messages from both redis instances [21:45:13] :) cool :) [21:49:58] Jeff_Green: I think we need the provider-defaults.yaml file still on the email pref center wiki [21:50:41] oh wait, two different redis servers? [21:50:51] * ejegg needs to read more backscroll [21:51:21] ejegg: yeah the proposal is to have a different redis server on a different port for email-pref-ctr [21:52:12] oh ok [21:52:56] ejegg: btw I haven't tested any of the provider-defaults stuff with the current patches [21:53:03] that file is fully commented out on Docker [21:53:20] so yah, you can set the queue params on a per-queue basis in the main.yaml no problem [21:53:35] ejegg: yep! [21:53:42] AndyRussG: oh, if it's all commented out in docker, then we're just using the defaults from the SmashPig repo [21:53:48] okok [21:54:10] ejegg: I added some notes on the localsettings.php.erb in temp on the etherpad [21:54:12] https://phabricator.wikimedia.org/diffusion/WFSP/browse/master/config/provider-defaults.yaml [21:54:44] ejegg: hmmm [21:55:02] so just logging to syslog, no failmail [21:55:20] ejegg: so not sure if any of the logging calls for e-p-c go through SP? [21:55:34] ejegg: I don't really understand how the backreferences work in the existing file, is that quick to explain? otherwise I can research it tomorrow [21:55:34] though I do see the email entry there [21:56:34] Jeff_Green: they're standard yaml backreferences [21:57:06] Jeff_Green: it's basically just that the part below &REDIS is considered to be copied and inserted in all the places that it says <<: *REDIS [21:57:06] the reference just populates the node it's under with the content of the node it's referring to [21:57:34] it's just interpreted as such by the yaml interpreter [21:58:04] so, completely outside the domain of erb/puppet stuff, if that's a question [21:58:11] ejegg: so without familiarity with yaml backreferences yet, what I'd like to do is separately name something to the effect of *DONATIONQUEUE vs *EMAILPREFSQUEUE [21:58:48] Jeff_Green: if you're going to just use the different settings once, then just add them directly under the email-preferences node [21:58:54] and don't use any backreference ther [21:58:56] *there [21:59:09] fr-tech just gonna accompany Doggo on his "errands", back in about 20 min :) [21:59:23] right, I'm just thinking in terms of repeatability in the puppet template [22:00:00] and thinking ahead to when we decide we need more than one queue in the EMAILPREFSQUEUE redis instance [22:00:06] repeatability? so this second redis instance will have more than just the email-preferences queue? [22:00:49] not today, but if I can design the template well now maybe I can save pain in the future [22:00:59] hmm [22:01:13] in redis-servers: [22:01:53] I'd say go for simplicity for now and just do it without a backreference till we come up with that second queue [22:01:54] what is &REDIS and everything under it? is it all one value? [22:02:16] &REDIS 'captures' all of the stuff under it [22:02:29] and stores it as the REDIS set of nodes [22:02:40] to be pasted in anywhere you see <<< REDIS [22:02:47] ok [22:05:29] so is value "redis-servers" arbitrary, just there so that the &REDIS part will be interpreted? or is redis-servers called anywhere else? [22:05:59] yep, exactly, that node is just a placeholder [22:06:07] oop, gotta head out [22:06:12] ok thanks! [22:14:22] cstone: you are reviewing that last patch? do you need any help? [22:14:30] I'm gonna check about monolog deployment [22:15:08] eileen: im getting some food right now but im going to look at it after that, i got my local setup behaving which was the biggest issue on friday [22:15:17] cool [22:20:11] (PS1) Eileen: Vendor update, phpmailer, monolog handler [wikimedia/fundraising/crm/vendor] - https://gerrit.wikimedia.org/r/684510 [22:20:25] (CR) Eileen: [C: +2] Vendor update, phpmailer, monolog handler [wikimedia/fundraising/crm/vendor] - https://gerrit.wikimedia.org/r/684510 (owner: Eileen) [22:22:34] PROBLEM - check_log_messages on frav1002 is CRITICAL: CRITICAL: minFraud_endpoint_critical 2 [=1] [22:23:36] (CR) jerkins-bot: [V: -1] Vendor update, phpmailer, monolog handler [wikimedia/fundraising/crm/vendor] - https://gerrit.wikimedia.org/r/684510 (owner: Eileen) [22:25:33] (CR) Eileen: [C: +2] "recheck" [wikimedia/fundraising/crm/vendor] - https://gerrit.wikimedia.org/r/684510 (owner: Eileen) [22:25:46] woo consuming and updating !! :) [22:27:34] PROBLEM - check_log_messages on frav1002 is CRITICAL: CRITICAL: GlobalCollect_endpoint_critical 1 [=1],minFraud_endpoint_critical 3 [=1] [22:28:11] (CR) jerkins-bot: [V: -1] Vendor update, phpmailer, monolog handler [wikimedia/fundraising/crm/vendor] - https://gerrit.wikimedia.org/r/684510 (owner: Eileen) [22:28:11] hmmm... looking at those check_log_messages errors. [22:31:13] looks like it may be transient. i'll keep an eye on it. [22:32:30] PROBLEM - check_log_messages on frav1002 is CRITICAL: CRITICAL: Paypal_endpoint_critical 1 [=1],cURL_error 5 [=5],minFraud_endpoint_critical 4 [=1] [22:34:03] cstone yaaaay :) :) [22:35:02] question on the part when it logs when none are processed, is there a reason were logging that? [22:37:15] cstone: hmmm not sure that might have come form eileen's initial version of the patch, but I don't see why it has to be a conditional like that [22:37:32] Just seems it would cause a lot of spam [22:37:34] PROBLEM - check_log_messages on frav1002 is CRITICAL: CRITICAL: GlobalCollect_endpoint_critical 1 [=1],Paypal_endpoint_critical 3 [=1],cURL_error 5 [=5],minFraud_endpoint_critical 3 [=1] [22:37:37] as in, it seems fine if it just says "Processed 0 messages" [22:37:44] cstone: ah I see right [22:37:49] hmmm [22:37:59] Cause if it's running at whatever frequency and no one is unsubing at most times [22:38:01] we can also get rid of it [22:38:03] yeah [22:38:18] so we wouldn't need something like that for debugging? [22:38:33] Hmm like a general it did x thing ? [22:38:46] Is there a success one sorry I'm on phone now walking back [22:39:03] If it's just the results of the run then that makes sense [22:41:51] cstone: I think that's it? looking in other queue consumers, it's the same pattern, see queue2civicrm drupal module [22:42:34] PROBLEM - check_log_messages on frav1002 is CRITICAL: CRITICAL: GlobalCollect_endpoint_critical 2 [=1],Paypal_endpoint_critical 2 [=1],cURL_error 7 [=5],minFraud_endpoint_critical 2 [=1] [22:47:34] PROBLEM - check_log_messages on frav1002 is CRITICAL: CRITICAL: Paypal_endpoint_critical 2 [=1],minFraud_endpoint_critical 1 [=1] [22:50:34] ahh i see AndyRussG I was thinking it was doing it an extra time but nope [22:50:59] ok that makes sense [22:52:34] PROBLEM - check_log_messages on frav1002 is CRITICAL: CRITICAL: Paypal_endpoint_critical 2 [=1],cURL_error 6 [=5],minFraud_endpoint_critical 4 [=1] [22:54:11] there is a problem in vendor just updating the phpmailer code - which is odd [22:54:31] since the monolog extension is not in that & I can't think what else [22:55:28] eileen: did you see that when the vendor directory is there and you run civibuild create wmff, it doesn't create the civicrm_monolog table? [22:55:41] AndyRussG: sorry this is gerrit [22:55:49] & without the monolog patch [22:56:02] ahh okok [22:56:09] https://gerrit.wikimedia.org/r/c/wikimedia/fundraising/crm/vendor/+/684510 [22:56:27] (that should merge to vendor & not have much to do with monolog) [22:56:32] ah okok [22:56:37] eileen: I was looking at https://phabricator.wikimedia.org/T281647 [22:56:49] but maybe that's the fix' [22:56:51] ? [22:57:17] AndyRussG: yeah - I wasn't going to look at that until we had merged monolog to deployment - but this is pre-monolog in deployment [22:57:34] PROBLEM - check_log_messages on frav1002 is CRITICAL: CRITICAL: GlobalCollect_endpoint_critical 1 [=1],minFraud_endpoint_critical 2 [=1] [22:57:40] hmmm [22:57:46] that's a hefty patch [22:59:00] (CR) Eileen: "recheck" [wikimedia/fundraising/crm/vendor] - https://gerrit.wikimedia.org/r/683482 (owner: Eileen) [22:59:10] the composer update one? [22:59:38] I'm gonna retry merging just the phpmailer - my current theory is that maybe the other package update touches symfony [22:59:54] if this still passes https://gerrit.wikimedia.org/r/c/wikimedia/fundraising/crm/vendor/+/683482 - I'll merge it [23:00:05] & get the phpmailer update done first [23:01:48] eileen: K I'm not understanding much but pls lmk if I can help w stuff :) [23:01:57] (CR) jerkins-bot: [V: -1] phpmailer update to 6 [wikimedia/fundraising/crm/vendor] - https://gerrit.wikimedia.org/r/683482 (owner: Eileen) [23:02:07] lol ok - I'll see what happend [23:02:34] PROBLEM - check_log_messages on frav1002 is CRITICAL: CRITICAL: GlobalCollect_endpoint_critical 1 [=1],Paypal_endpoint_critical 2 [=1],minFraud_endpoint_critical 6 [=1] [23:02:38] nope phpmailer patch failing now too - even though it passed last week - same patch [23:07:34] PROBLEM - check_log_messages on frav1002 is CRITICAL: CRITICAL: cURL_error 6 [=5],minFraud_endpoint_critical 3 [=1] [23:12:34] PROBLEM - check_log_messages on frav1002 is CRITICAL: CRITICAL: GlobalCollect_endpoint_critical 2 [=1],Paypal_endpoint_critical 1 [=1],minFraud_endpoint_critical 3 [=1] [23:13:39] these alerts are annoying but not fatal. i'm going to ack them for a while as i have to kid taxi in a few. [23:14:54] ACKNOWLEDGEMENT - check_log_messages on frav1002 is CRITICAL: CRITICAL: GlobalCollect_endpoint_critical 2 [=1],Paypal_endpoint_critical 1 [=1],minFraud_endpoint_critical 3 [=1] Dwisehaupt transient errors on a few providers from different hosts. acking for now to give it time to settle. [23:21:58] Hmm Ok I think it's the changes made in https://phabricator.wikimedia.org/T281609 [23:22:41] hmm gerritt isn't loading for me [23:24:07] eileen: oh hmmm which of those? [23:24:08] I'm finding it does after a bit [23:25:06] AndyRussG: I'm pretty sure the reason this line was commented out https://gerrit.wikimedia.org/r/c/wikimedia/fundraising/crm/civicrm-buildkit/+/683994/2/app/config/wmff/install.sh is because it caused the problem we are hitting [23:26:32] for history - that whole precreated.sql thing is because it used to be the case that we created our db on vagrant and had to let vagrant pre-create because of internal wmff setup [23:28:21] hmmm okok [23:28:32] eileen: I can test commenting that out again... [23:28:39] any suggestions then on another way to do that? [23:28:57] ideally unravel the whole CI pre-create dbs thing [23:28:59] (somehow) [23:29:09] under wmff there is a bin/ci-create-dbs and that sets up the databases when jenkins is running it [23:29:50] hmmm [23:29:56] also I don't see any precreated.sql anywhere [23:31:21] no I'm trying to recall where our script causes that to be used [23:31:42] in the past the fredge thing has never worked quite right because we've had to prioritise CI over local [23:32:14] but if we could get the magic out of ci it would help - I was hoping that once we had the repo phab done we might figure that out [23:34:19] (PS1) Eileen: Attempt to unravel pre-create [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/684537 [23:34:38] AndyRussG: I don't think ^^ is magically going to work but it's the right code areas [23:37:27] (PS2) Eileen: Attempt to unravel pre-create [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/684537 [23:38:01] AndyRussG: in the meantime I think we have to comment out the line again because we have to prioritise CI [23:41:22] (CR) jerkins-bot: [V: -1] Attempt to unravel pre-create [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/684537 (owner: Eileen) [23:44:40] (PS3) Eileen: Attempt to unravel pre-create [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/684537 [23:46:12] eileen: before you dive too deep, what about if we just add directly create the db in setup.sh? [23:46:23] and leave that stuff commented out then? [23:46:27] would that be easier for now? [23:47:37] well we Do need to get rid of the precreated stuff at some point because it is a pain - but yeah unless it turns out to be easier than I think we might need to do that. [23:47:38] also cstone eileen not meaning to distract or change topic here, just a thought, if we could get the queue consumer merged today that'd prolly be nice? :) [23:47:57] eileen okok yeah I think that's fine eh [23:48:00] (CR) jerkins-bot: [V: -1] Attempt to unravel pre-create [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/684537 (owner: Eileen) [23:48:13] AndyRussG: its looking good just trying to break it haha [23:48:27] eileen without having looked, it sounds like it's a pretty labyrinthine rabbit hole [23:48:31] cstone ah cool beans thx! [23:48:46] i guess what i was trying to do was get to these messages https://gerrit.wikimedia.org/r/c/wikimedia/fundraising/crm/+/682349/13/drupal/sites/default/civicrm/extensions/wmf-civicrm/CRM/Queue/PreferencesQueueConsumer.php [23:48:51] cause i dont see them in the logs [23:51:23] cstone: right... well the one on line 21 ("No records updated from e-mail preferences message") I think would only happen if somehow a record wasn't written, or multiple records were written, and an exception wasn't thrown [23:51:41] you could change the !== 1 conditional just to test that it gets to the log [23:52:52] I have seen the error message about invalid data appear somewhere, I think on the command line when the test was broken... but I'm not positive that it goes to all the right places, I guess [23:52:54] ok cool that worked [23:53:04] cool! :) [23:54:14] i think this looks good if gerrit would behave [23:56:09] i guess the only thing is the log messages have a empty [] after them [23:56:23] but thats probably a log issue? [23:56:49] (PS4) Eileen: Attempt to unravel pre-create [wikimedia/fundraising/crm] - https://gerrit.wikimedia.org/r/684537 [23:58:10] cstone: ah yeah I think that's a thing with the new monolog stuff, where you can send in an array with the values of variables you want to log, or something? [23:58:27] ah nice [23:58:48] dunno if there's a different value we can send in to prevent the []? I think it's an argument we're not sending [23:59:53] yeah its fine for now