[02:44:36] danmichaelo, you can use https://meta.wikimedia.beta.wmflabs.org/wiki/Special:OAuth now [02:44:38] I set up a valid, not self-signed, certificate for that domain [08:17:17] Looks like I have no route to gerrit.wikimedia.org :o [08:17:24] It worked until 5 min ago [08:43:50] Nemo_bis, still having issues? [08:44:32] I'm getting packet loss in Telia [08:47:42] and now it's stopped [08:49:13] Seems ok now [08:50:02] phabricator worked though, are they in different clusters/routes? [08:53:33] gerrit.wikimedia.org has address 208.80.154.85 [08:53:35] phabricator.wikimedia.org has address 91.198.174.217 [08:54:32] looks like that phab address is misc-addrs... so it's routed via esams for us [08:56:13] whereas gerrit has it's own public IP (due to git ssh on port 29418 I guess) and so you go all the way to eqiad before getting to the wikimedia network [08:58:13] yeah they're entirely different for these purposes Nemo_bis [08:59:25] makes sense [08:59:40] pity that eqiad is so poorly connected [09:35:57] eqiad isn't poorly connected at all [10:22:23] well, poorly connected to every location in Europe I happen to connect from ever [10:53:34] Nemo_bis, well, we have esams :) [10:58:30] Sometimes we do, yes [10:59:15] (Cf. https://phabricator.wikimedia.org/T122917#2444213 ) [10:59:52] you're making a lot of invalid assumptions along the way :) [11:00:11] anyway, please provide the ip addresses of the locations you're having throughput issues from and I can have a look [11:33:50] mark: which ones qualify as having thoughput issues? [11:34:12] ...the ones where you think they're slow and could be faster? :) [11:34:21] That's all of them, even labs :) [11:34:38] no, i'm talking about -your- ips/boxes [11:34:46] Define "your"? [11:34:57] the ones where you were experiencing poor throughput from? [11:35:05] yes, that includes labs [11:35:20] I'm still waiting for comments on the test I made in December under faidon's direction https://phabricator.wikimedia.org/T120425#1858664 [11:35:55] unlikely you will get any now unless you provide ips again [11:36:55] You can get most of those IPs with a ping of the specified domain... [11:37:10] what domain? [11:37:29] Those mentioned in the comment [11:37:59] Though I think only koti.kapsi.fi is still the same [11:39:39] At home the only thing I care about is saturating my bandwidth, since it's not much (10 Mb/s). Usually I do bigger transfers from the servers though. [11:40:08] Or from university network, do you need some IPs of a university network? [11:42:01] An example would be 131.175.28.130; I never manage to get a decent download rate there (decent, on that network, is at very least 5 MiB/s; more often 30) [11:42:06] ok [11:42:35] that route is only one hop telia, then hops over to Cogent [11:49:34] On the bright side, IPv6 on one of my servers makes the download 4 times as fast [11:50:18] But only when one doesn't need it (from esams) [15:15:51] mark: hey, Can I be added as a member to ops-l? I requested but probably it needs approval. If it requires a dedicated access request (phab card, etc.) Tell me and I will make one [15:16:08] Ticket number? :) [15:17:03] andre__: I don't get it. I'm probably wasn't clear. https://lists.wikimedia.org/mailman/listinfo/ops [15:17:10] I want to be member of this mailing list [15:18:22] Amir1: is there a Phab ticket ID? [15:18:37] there won't be [15:18:41] these requests don't go through phab [15:18:46] ah. [15:18:58] Amir is referring to a mailing list [15:18:59] I thought it might also be an #Ops-Access-Request maybe :) [15:19:02] yeah [15:19:13] ops-access-requests are for server groups [15:19:19] I see... thanks! [15:20:13] per the page mark or robh (and also Daniel who is not around atm) can approve [15:22:25] so ops-l requires nda or staff rights [15:22:36] I have signed nda [15:22:42] robh: ^ [15:22:59] i'd suggest creating the ops-access-request for it though just so there is an audit trail. was it the nda on phab? [15:23:09] that way we can document that it was checked, etc. [15:24:22] I didn't have to do that [15:25:14] Krenair: you are staff so its less confusing [15:25:15] robh: https://phabricator.wikimedia.org/T134651 [15:25:30] I am a contractor [15:25:35] https://github.com/wikimedia/operations-puppet/blob/production/modules/admin/data/data.yaml [15:25:40] It's confusing in a different way [15:25:40] Krenair: ok, same thing to me =] [15:25:48] (search for ladsgroup) [15:26:22] ops-l access should've been given along with deployment access [15:26:46] agreed but i had no idea amir1=ladsgroup until just this second. [15:26:53] so its less confusing now ;] [15:27:02] :D [15:27:22] and i just approved a bunch of your posts to the list anyhow [15:27:22] I wrote my IRC nick in my user page every where possible. [15:27:30] no worries, im adding you now [15:27:50] thanks robh :) [15:27:56] well, im lookign for the deployment access task to append to it [15:27:58] and then add you [15:29:43] probably it should a subtask [15:30:03] i happen to agree with Krenair on this one, i think anytime we give someone deployment, they should be added to ops-l =] [15:30:14] since its where we announce major deployment windows and the like [15:30:48] yeah, agree [15:31:30] i think that T136406 is the most relevant but is also pretty old, so meh. just adding you to the list now =] [15:31:30] T136406: Requesting deployment access (for deploying to scb) for Ladsgroup - https://phabricator.wikimedia.org/T136406 [15:32:12] Amir1: you want it to the gmail address you posted to the lsit recently with right? [15:32:31] robh: yeah [15:32:36] ladsgroup at gmail [15:33:20] Amir1: subscribed [15:33:34] thanks robh! [15:33:35] side benefit, no more auto-moderation of your messages ;] [15:33:41] :D [15:34:23] drawback=gerritspam, welcome to gerrit spam hell. [15:34:41] i have half a dozen filters for ops-l messages. [15:34:58] :))) [15:35:03] let's wait and see [15:35:45] well, half of ops is on vacation, it'll be a light week of commit messages compared to most weeks =] [15:35:55] (maybe not half, but a third it seems) [15:38:43] so, you *read* gerrit messages? ;) [15:39:45] not all of them, i skim but they are auto-archived out of inbox [15:39:52] i only inbox the private repo commits [19:29:46] sorry, just looking, [20:18:49] huh missed the chance to say hi [20:20:49] it was quite a slim chance :o [20:22:11] 14 seconds... [20:47:18] I woul dhave had to be watching for it