dpawlik | ttx: hey, so it's a good news! | 08:06 |
---|---|---|
ttx | dpawlik: It's better, but still slightly above what we can afford unfortunately. Back in October we were running at $133/day, then on Oct 18 it jumped to $155/day and with my adjustments we dropped to $141/day. Our donation is based on $134/day... so I would still look at options for disabling the Fargate logstash if it's actually not used :) | 09:30 |
ttx | Just running the OpenSearch service and friends would drop us to $105/day which would be super safe | 09:39 |
dpawlik | that would be very good for us | 10:23 |
dpawlik | Hope R eed will reply for an email that I send on Friday | 10:23 |
dpawlik | ttx Maybe I did not add you to cc | 10:23 |
dpawlik | tl;dr I asked R eed if he can remove logstash or confirm that we can do that | 10:24 |
dpawlik | I guess you have enough permissions to do it in AWS console in cloudformation service | 10:24 |
dpawlik | when I check it today, I don't see any relationship between OpenSearchstack and Lostash stack so it might be safe to remove. Hope he will reply soon so we will "unlock"' some money | 10:25 |
ttx | yes I have enough rights to do it if he confirms it's not going to kill us | 10:26 |
ttx | It certainly looks very idle from a resource consumption standpoint (now that it's not busy rebooting unhealthy hosts) | 10:27 |
ttx | (peaks at 1% CPU so at the very least could use a lot less powerful instances) | 10:28 |
scoopex | I don't quite understand how to deal with Gerrit :-) I created a fork of "ansible-kolla" on Github and created a change on Gerrit. (see https://review.opendev.org/c/openstack/kolla-ansible/+/900528) After receiving feedback, I modified the commit on my branch using "--amend". | 12:30 |
scoopex | Now I have two patch sets :-( How can I fix this? Where can I read how to do this better? | 12:30 |
scoopex | Just saw in the topic that #opendev is the right place. Have now asked the question there... | 12:39 |
fungi | scoopex: yeah, if you have questions about contributing to openstack then this is probably the better channel since openstack has its own nuanced policies. if it's generally about contributing to some other project hosted in opendev then #opendev is a better channel for that | 12:59 |
frickler | scoopex: getting a new patchset when you amend a commit is exactly the expected event, https://review.opendev.org/c/openstack/kolla-ansible/+/900528/2..3 looks just fine to me, what issue do you see with that? | 13:33 |
JayF | looks like google is indexing a wiki-upgrade-test.openstack.org | 15:43 |
JayF | which is now #2 in results for some openstack queries, and pointing to a bad cert | 15:43 |
JayF | e.g. https://www.google.com/search?q=nova+meeting+openstack&oq=nova+meeting+openstack shows https://wiki-upgrade-test.openstack.org/wiki/Meetings/Nova #2, directly beneath the actual wiki | 15:44 |
fungi | JayF: that's the name of the wiki server, for unfortunate reasons | 16:11 |
fungi | it was an in-progress upgrade test (circa... 2015?), but then the original server got compromised and we had to pivot | 16:12 |
fungi | we can probably just add a permanent redirect in the vhost for it. i'll try to look at that later today | 16:12 |
tonyb | fungi: Is that server actually managed by puppet as the vhost config suggests? | 22:39 |
fungi | it is not. it was on its way there and was provisioned by a partial puppet module | 22:49 |
fungi | but the module was never completed | 22:50 |
tonyb | Okay. | 22:50 |
tonyb | So I could add a new vhosts for wiki-upgrade-test.openstack.org:80 and wiki-upgrade-test.openstack.org:443 that redirect to wiki.openstack.org:443 | 22:51 |
fungi | or just set the catch-all vhost to redirect any url that doesn't have the expected hostname, i think the etherpad vhost may have an example of that? | 22:52 |
tonyb | Oh okay that'd work | 22:53 |
* tonyb will poke and report back | 22:53 | |
fungi | mmm, no we used a separate redirect.vhost.j2 in the etherpad role. i'll find a better example | 22:55 |
fungi | tonyb: https://opendev.org/opendev/system-config/src/branch/master/playbooks/roles/codesearch/templates/codesearch.vhost.j2 may be the most concise example | 23:07 |
fungi | and i was wrong, etherpad does also do the same thing (not sure why it also has a separate redirect vhost template in the role) | 23:09 |
fungi | RewriteCond %{HTTP_HOST} !{{ etherpad_vhost_name }} | 23:09 |
fungi | RewriteRule ^.*$ https://{{ etherpad_vhost_name }} [L,R=301] | 23:09 |
fungi | go figure | 23:09 |
tonyb | Got it | 23:09 |
tonyb | fungi: ... and I don't have sudo on that server | 23:10 |
tonyb | so much for me taking $item off your hands, more I've "forced" you to lok at it on my timetable :/ | 23:11 |
fungi | tonyb: log out and back in again, then try once more | 23:14 |
fungi | i've manually added you to the sudo and admin groups on the server, i should have thought to do that when i created your account | 23:14 |
fungi | anyway, as i said, cribbing from the codesearch example is probably the easiest since it doesn't involve any jinja templating in the rewritecond statement | 23:15 |
tonyb | pefect thanks | 23:15 |
fungi | just change codesearch to wiki and opendev to openstack and it should simply work | 23:16 |
fungi | ouch, the RedirectMatch in the https vhost is kinda hilarious | 23:18 |
fungi | it's redirecting to http when is then going through the http vhost to redirect back to https again... | 23:18 |
* fungi sighs | 23:18 | |
tonyb | fungi: https://paste.opendev.org/show/bewAF6FW6wLBFZ6fz1qg/ look okay? | 23:19 |
fungi | lgtm, also while you're there the RedirectMatch destination can be corrected to https | 23:20 |
tonyb | https://paste.opendev.org/show/bS3rXoA9rNjqgYGJzOKU/ | 23:21 |
fungi | i think it dates from when we hadn't added an ssl cert | 23:21 |
fungi | and got missed in the conversion | 23:21 |
fungi | perfect | 23:21 |
fungi | oh, openstack not opendev | 23:21 |
fungi | tonyb: ^ | 23:21 |
fungi | i almost missed that | 23:21 |
tonyb | https://paste.opendev.org/show/bwpS2VToTHtsDi56BRwe/ | 23:22 |
fungi | yeah, that's got it now. thanks! | 23:23 |
tonyb | Looks good: https://paste.opendev.org/show/blMeiWx8kNAFEcZWUw1v/ | 23:25 |
tonyb | JayF: ^^ | 23:27 |
JayF | Any objection to me asking google to reindex? | 23:27 |
JayF | ah, you have to be the owner in the search console to do that now | 23:28 |
tonyb | JayF: not from me. | 23:28 |
JayF | which I am not, so I can't make it go faster :) | 23:28 |
fungi | i don't know that there is any search console for it | 23:28 |
fungi | if there is, i don't know who would have access to it | 23:29 |
tonyb | JayF: Ahh okay. I roughly understand hwo that should work in terms of validation but I'd need to work on getting access | 23:29 |
JayF | It'll get fixed in time anyway. You have to establish ownership via DNS or a file on the website, as you usually do with such things, but it doesn't matter. | 23:29 |
JayF | Used to be there was just a place, you put in a URL, it goes in a queue, no authentication | 23:29 |
tonyb | fungi: It uses domain/DNS validation to allow authorised, I assume, google accounts to log into the search console and do $stuff | 23:30 |
tonyb | fungi: It's kinda on my "nice to have list" so we can address some docs issues that were dicussed on openstack-discuss recently | 23:30 |
fungi | yeah, i'm vaguely aware of what it is and how it works, just no idea if someone set anything up for it. if they did, it was likely done a dozen years ago | 23:31 |
tonyb | fungi: Assuming we get consensus around it allow us to tell google to prefer docs.o.o/latest over all the old links *without* removing them | 23:31 |
fungi | i take it we can't do that via robots.txt or sitemap.xml? | 23:32 |
fungi | full disclosure: most of what i know about being a webmaster is fresh from the 1990s, and i still miss gopher and archie... | 23:33 |
tonyb | We can do some of it but not all of it and without access to the GSC we can't ask googlebot to look again | 23:33 |
* tonyb was thinking about gopher just the other day | 23:34 | |
JayF | The tools you get these days are more like | 23:34 |
JayF | "here is WHY we prefer X over Y" | 23:34 |
JayF | not "please prefer Y over X" | 23:34 |
JayF | but they are incredibly helpful, and with a few minor changes I cleared up any indexing issues a refactor of my website (oldos.org) had caused to my pagerank | 23:35 |
tonyb | and recalling when the sysadmins installed it into /usr/local/bin and removed /home/**/{bin/gopher,src/gopher*} and freed up $many MB of storage | 23:35 |
JayF | tonyb: literally like, a week before I was about to setup a gopher server for oldos (I had done the technical research on how to get hugo to output gopher-friendly content from oldos.org) is when firefox announced they dropped gopher support so I canned the whole project :( | 23:41 |
tonyb | JayF: :( | 23:41 |
tonyb | fungi: Looks like no-one has set it up. | 23:47 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!