zas: hi! please let me know when you are around to deploy some oauth stuff?
pite has quit
dabeglavins has quit
aerozol[m]
Hey ansh worked a bit late but got the homepage mockups all ready for you to work on! I have left a few important notes as figma comments, let me know if you’d rather have the notes written out elsewhere, or if you would rather have a ticket with screenshots or something. Let me know if you have any questions
Link for the icons (use the linked ones, they are optimized for size) is at the top, in the yellow box
reosarevok[m]
<aerozol[m]> "Don’t know if it’s on purpose..." <- That's an artist credit so I guess it might show on hover there?
huhridge joined the channel
huhridge has quit
huhridge joined the channel
huhridge has quit
yvanzo: hey, is there any way I can help with the solr stuff today? :)
huhridge joined the channel
LupinIII
re recording alias
[12:29] reosarevok[m] The alternative I can see is showing it on hover only
I'd prefer that tbh
yvanzo[m]
Hi reosarevok: Sure, I’m currently stuck on backup creation.
LupinIII
oh i guess maybe wont show on mobile. but honestly an option for "alwaysshow/only with hover" where alwasy show is default would be stellar
lucifer[m]
monkey: hi! are you still unable to download your data exports? i tried it last week and today again. it works fine for me every time. i do see some errors in sentry so it is clearly broken for some users but unable to reproduce it on my end so far.
LupinIII
aerozol[m]: i *really* like the addition of the word "profile" in the sidebar. that's very usefull to a guy like me who from the lb frontpage just want to go to their listens and such
i always end up going to like "feed" first because idk the current doesn't make me think "my lsitens/myprofile"
<LupinIII> "oh i guess maybe wont show on..." <- For what it's worth, very very few recordings currently have aliases, so it'll be very rare at least at first anyway :)
I think we can move it once more aliases exist if it's bothering people?
(maybe it won't)
mayhem[m]
restarting the container fixed the problem, but the PR above is to mitigate this mysterious rollback required from happening again.
I thought that the backups were not working, I was just not connecting to the same cloud.
lucifer[m]
mayhem: i think that error is unrelated. the connection having problems in the error log is `ts_conn`, the connection in the code is `db_conn`
mayhem[m]
fun
lucifer[m]
i'll add the missing the rollback and move it to a finally block instead.
if you prefer to restart the container for these errors, i can leave that part in.
mayhem[m]
lucifer[m]: yes please. I think all of our containers should have that idea -- a catch all, that if we catch that, exit the container. who knows what went wrong, when this happens, so why rely on the state of a possibly broken container?
lucifer[m]
makes sense but then we don't need a catch all. the process/container will exit on an uncaught exception anyway.
atj: ^^ we have 2a01:4f8:c010:b00c::2 in DNS, node has inet6 2a01:4f8:c010:b00c::1/64 scope global, firewall issue?
Ah got it, just a misconfiguration; almost all our nodes use :2, but not solr cloud ones
I'll fix the DNS
mayhem[m]
<aerozol[m]> "Came up in my subscriptions..." <- thanks, responded.
reosarevok[m]
yvanzo: what exactly is it that I should be doing? Is it running the upgrade instructions on that file on a local install of solr?
(the main reason I haven't been working on solr stuff so far is because there seem to be no actionable instructions on what needs to be done and tested and how)
yvanzo[m]
reosarevok: I pointed you to the docs for setting up SSH.
reosarevok[m]
Yes, I saw that, but just setting up ssh isn't much of a test, so I expect you want me to do something with it too :)
yvanzo[m]
Read the backlogs, there are issues with SSH-ing.
Setting up ssh is a prerequisite to help further.
zas[m]
Not really an SSH issue, there were issues with DNS IPv6 addresses of those nodes, it should be fixed now (though TTL is 1h). They were set as ending by :2 instead of :1. ssh used IPv6 first and fallback on IPv4 if it fails, explaining the delay.
reosarevok[m]
I did ssh mb-solrcloud-1 with that config and it let me in immediately, so I guess zas fixed it
yvanzo[m]
Yes he just did
mayhem[m]
zas: I am writing a cron wrapper script that monitors the output of the cron job and if the cron job does not exit with status code 0, it posts the last X lines from the output to the LB telegram channel. I've got everything working, except for the "send to telegram" bit. Any thoughts on how to best do this?
ok, send it directly. I wondering if you had something else in mind. cool, I'll investigate
do you have the bot token and a chat id I can use for this purpose, zas ?
reosarevok[m]
yvanzo: ok, neat, what's the next step if any? :)
(if you're on top of things and don't need help that's of course fine, just say so!)
Maxr1998 has quit
Maxr1998 joined the channel
yvanzo[m]
reosarevok: Please let me know what time you are available today, to be efficient in Q/A.
reosarevok[m]
Well, I need to take 30 mins to eat at some point (although I can do that while stuff happens)
After that I'm available until 17 CET or so
yvanzo[m]
You need a MB Docker instance with sample data and search server set up.
I'm going for lunch, I'll be back for 14 CET.
monkey[m]
mayhem: might be worth storing those logs+message somewhere to retry in case the telegram sending fails (for example when we get a proper outage or loss of connectivity)
remember, these cron logs will still be on the server. we just won't get a notification for it
BrainzGit
[listenbrainz-server] 14amCap1712 merged pull request #3036 (03master…mitigate-background-task-rollback-error): Background tasks container stuck in a rollback loop https://github.com/metabrainz/listenbrainz-serv...
allen joined the channel
monkey[m]
mayhem: If you have some time, can you please look at LB#3042 ? It's short and we're getting a bunch of 500 errors on artist pages after the partial date fix
