Morning, yvanzo (and other interested parties). Just wanted to let you know that I successfully completed a reindex overnight. It took about 16 hours, up from around 12. It looks to me that artist indexing is significantly slower than it used to be. But that's just a sample size of one ;)
2025-05-21 14138, 2025
lucifer[m]
mglubb: there are a couple of unreleased PRs that improve the indexing time.
2025-05-21 14157, 2025
lucifer[m]
especially, artist core.
2025-05-21 14144, 2025
lucifer[m]
for us, it halves the indexing time for artist core from ~20 mins to 10 mins.
2025-05-21 14145, 2025
mglubb[m]
Thanks, lucifer . I'll wait for them to be merged into main then give it a whirl. As it happens, I'm seeing similar slowness in ongoing indexing. Hopefully they'll help that?
2025-05-21 14124, 2025
lucifer[m]
yes, i think the live indexing should be faster too with that.
2025-05-21 14156, 2025
Maxr1998 joined the channel
2025-05-21 14120, 2025
Maxr1998_ has quit
2025-05-21 14151, 2025
mthax has quit
2025-05-21 14108, 2025
mthax joined the channel
2025-05-21 14119, 2025
yvanzo[m]
Hi lucifer, yes.
2025-05-21 14155, 2025
yvanzo[m]
It took ~5h.
2025-05-21 14114, 2025
lucifer[m]
the mb docker numbers i have for it are 7-8h. so seems a across the board improvement.
Also I stopped the live indexer for reindexing the whole collections.
2025-05-21 14117, 2025
yvanzo[m]
So there are some messages to be processed now.
2025-05-21 14113, 2025
yvanzo[m]
lucifer: I just left a comment to your last PR for the retry strategy.
2025-05-21 14157, 2025
outsidecontext[m
mayhem, zas: I'd like to tag a libdiscid release. Can one of you help me with getting the release files copied to the MB file servers afterwards?
2025-05-21 14105, 2025
outsidecontext[m
There is not much changes, but a CMake fix that is waiting to be released for two years now, and also the binaries for Windows and macOS get code-signed now.
2025-05-21 14120, 2025
mayhem[m]
not a problem.
2025-05-21 14143, 2025
outsidecontext[m
ok, great. then I'll tag the release and ping you when its ready.
2025-05-21 14123, 2025
outsidecontext[m
looks like I'm accidental de-facto maintainer of libdiscid since a few years. Not sure how this happened
2025-05-21 14101, 2025
lucifer[m]
yvanzo: done.
2025-05-21 14121, 2025
lucifer[m]
yvanzo: is it fine i disable the sir jenkins jobs?
[@_discord_410549040392962058:chatbrainz.org](https://matrix.to/#/@_discord_410549040392962058:chatbrainz.org) when did you send the messages? I'll check.
2025-05-21 14159, 2025
adhawkins[m]
10:33:41 (BST) and 11:38:35 (again BST)
2025-05-21 14109, 2025
adhawkins[m]
s/BST/GMT + 1/, s/BST/GMT + 1/
2025-05-21 14131, 2025
lucifer[m]
adhawkins[m]: [@bitmap:chatbrainz.org](https://matrix.to/#/@bitmap:chatbrainz.org) and [@yvanzo:chatbrainz.org](https://matrix.to/#/@yvanzo:chatbrainz.org) got the same issue yesterday. while waiting for their response you can check the chatlogs from last night/early morning.
2025-05-21 14158, 2025
adhawkins[m]
Will do. Thanks.
2025-05-21 14154, 2025
rayyan_seliya123 has quit
2025-05-21 14144, 2025
adhawkins[m]
There is talk of rebuilding images and the like. Wonder if that's been done?
2025-05-21 14120, 2025
adhawkins[m]
I'll wait for bitmap or yvanzo to respond. For now, lunch time!
2025-05-21 14143, 2025
yvanzo[m]
adhawkins: Apparently, an additional command can be required (depending on your Docker Compose version?), see the updated step 4.
2025-05-21 14104, 2025
adhawkins[m]
Thanks, will double check.
2025-05-21 14114, 2025
yvanzo[m]
(the build command)
2025-05-21 14130, 2025
yvanzo[m]
Patched live indexing sir container for production SolrCloud 9 cluster.
2025-05-21 14153, 2025
texke has quit
2025-05-21 14139, 2025
petitminion has quit
2025-05-21 14110, 2025
petitminion joined the channel
2025-05-21 14155, 2025
lucifer_ joined the channel
2025-05-21 14102, 2025
lucifer_
testing irc bridge.
2025-05-21 14136, 2025
lucifer[m]
adhawkins: the irc bridge seems to be working, can you check you sent the message to right channel on the right network?
2025-05-21 14131, 2025
petitminion has quit
2025-05-21 14121, 2025
adhawkins[m] uploaded an image: (87KiB) < https://matrix.chatbrainz.org/_matrix/media/v3/download/chatbrainz.org/acQOtOwHvWSzvMCvwYYNWoOj/image.png >
2025-05-21 14153, 2025
adhawkins[m]
Done that, and now the upgrade seems to be progressing. Thanks.
fwiw we are on pg 16 in production so you can use gen_random_uuid without uuid-ossp too. i intend to do that in LB once timescale is migrated to 16 or 17.
2025-05-21 14102, 2025
lucifer_ has left the channel
2025-05-21 14113, 2025
fettuccinae[m]
lucifer[m]: ill do that.
2025-05-21 14142, 2025
fettuccinae[m]
<lucifer[m]> "change the column type to uuid" <- i was type casting it back to test locally `uuid_generate_v4()::text` and was still getting the same error.
I keep seeing the following in my replication logs. Should I be concerned?... (full message at <https://matrix.chatbrainz.org/_matrix/media/v3/download/chatbrainz.org/IiXhwEQdYzlnZPFswnCtpYWf>)
2025-05-21 14138, 2025
outsidecontext[m
An MB API questions: In release results mediums have a field track-offset. What exactly does this indicate, and in which cases is it something other than 0?
outsidecontext: I believe this is used for continuous numbering across mediums. So for example if you have songs 1 - 10 spread on two mediums, you will get tracks 1-5 on each medium.
2025-05-21 14144, 2025
monkey[m]
track offset allows you to define that medium 2 track 1 is actually track 6
2025-05-21 14109, 2025
monkey[m]
(although please correct me if I'm wrong)
2025-05-21 14150, 2025
mshahnawaz1[m] has quit
2025-05-21 14110, 2025
outsidecontext[m
monkey: I'd assume something like this. But can this actually be set for releases on MB? I had seen this field before, but never questioned it :)
2025-05-21 14149, 2025
monkey[m]
Didn't realize you can't set it on the website
2025-05-21 14146, 2025
outsidecontext[m
Actually it looks to me as if this field is actually unused. I guess the intention was as you described. But in practice it isn't in use. Something like the case with different numbering would just have the track "number" strings set differently from the position. E.g. on https://musicbrainz.org/ws/2/release/805cc5b2-1d5…
<lucifer[m]> "@fettuccinae:matrix.org: mamanul..." <- Sorry internet issue due to bad weather!
2025-05-21 14134, 2025
monkey[m]
<outsidecontext[m> "Actually it looks to me as if..." <- I suppose you're right. Perhaps a leftover from the days of yore when one could not set arbitrary numbering for tracks?
2025-05-21 14103, 2025
yvanzo[m]
The MusicBrainz indexed search is up-to-date and refreshed in real-time again.
2025-05-21 14130, 2025
yvanzo[m]
A backup of SolrCloud 9 collections will be available in a few hours.
2025-05-21 14115, 2025
julian45[m]
<lucifer[m]> "monkey: mac or linux?" <- if i may add my $0.02 on this: IMO macs provide a considerably nice (and, if you're willing to pay for it, beefy) setup that "just works" out of the box, but main downsides are cost & (sometimes) less ability to tinker w/ things you might want to tinker with.
2025-05-21 14115, 2025
julian45[m]
on the other hand, linuxes of various sorts are getting closer and closer to that nice "just works" state out-of-box and you have a lot more freedom to tinker, but generally there can be a lot more *need* to tinker to reach desired working state (e.g., peripheral drivers, firmware, desktop environment, etc.) vs macs
2025-05-21 14128, 2025
julian45[m] uploaded an image: (9KiB) < https://matrix.chatbrainz.org/_matrix/media/v3/download/julian45.net/M5Xw5KaXSMzY76HO6Bi2N9m1tpWIWXzt/2025-05-21_12-07-56.png >
2025-05-21 14132, 2025
julian45[m]
also, a nice coincidence as i found myself in the GSoC welcome meeting a little bit ago ^
2025-05-21 14102, 2025
lucifer[m]
julian45: yes makes sense. i asked because monkey used to use a mac and i use linux and usually its me having the a/v issues.
2025-05-21 14150, 2025
monkey[m]
Linux this time...
2025-05-21 14120, 2025
julian45[m]
personally i have a bit of all three majors: my main laptop is an m2 macbook air from a few years ago, but my gaming/dev setup is a windows 11 desktop i built myself and use a lot of wsl/containers - then my homelab environment is quite linux-y
2025-05-21 14141, 2025
monkey[m]
I have this very reassuring physical switch to turn off my microphone and webcam each separately.
2025-05-21 14141, 2025
monkey[m]
Turns out Zoom didn't like me turning on the hardware after starting the call.
2025-05-21 14100, 2025
monkey[m]
The usual have-you-tried-turning-it-off-and-on-again prevails.
2025-05-21 14140, 2025
julian45[m]
as for linux distros... started & grew my *nix skills with ubuntu and don't mind it at all, but after i started using RHEL for $dayjob i started to get a lot more interested in that ecosystem, incl upstream fedora... (for one thing, os upgrades tend to Just Work [or if it won't work, it'll block you from upgrading, tell you why, and roll you back] thanks to leapp.) also partial to arch
2025-05-21 14143, 2025
Jade[m] joined the channel
2025-05-21 14143, 2025
Jade[m]
My experience recently has been that windows has been abysmal, and that mac and linux have both been pretty good, with different failings
2025-05-21 14124, 2025
Jade[m]
I won a mac recently and I've been using it pretty much exclusively, because of the hardware, but macos has been much more reliable than windows was during the time I had to use it
2025-05-21 14101, 2025
Jade[m]
And stuff like built-in dictation is very nice (when you had to pay for a I think £700 licence for dragon on windows, and it's just not available on linux)
2025-05-21 14157, 2025
petitminion joined the channel
2025-05-21 14107, 2025
julian45[m]
yeah. even from an enterprise standpoint even though windows has a pretty staying presence as The enterprise OS and microsoft has put out lots of tools to manage the complete vertical of their ecosystem (intune for device management, entra id for identity/access management, "good" old active directory, etc.), windows unreliable overall and microsoft seems to be actively trying to make it worse over time, which is impressive in a way
2025-05-21 14125, 2025
julian45[m]
and if you're an org that needs to manage more than a handful of mac endpoints, there's pretty respectable tooling out there for doing that, even if it doesn't come straight from apple. (heck, there are now tools for doing similar quality work for linux endpoint fleets!)
2025-05-21 14116, 2025
julian45[m]
at $dayjob our userbase on college-issued computers is something like 70% macs vs 30% windows these days
2025-05-21 14129, 2025
julian45[m]
* these days, working out to well over a thousand macs
2025-05-21 14130, 2025
julian45[m]
* at $dayjob our userbase on workplace-issued computers is something like 70% macs vs 30% windows these days, working out to well over a thousand macs
2025-05-21 14135, 2025
bitmap[m]
<adhawkins[m]> "I keep seeing the following in..." <- hi adhawkins, assuming you're using MB docker, this is likely due to the icu library being upgraded in the `db` container when it was recreated. can you try the following command and see if it resolves the issue?... (full message at <https://matrix.chatbrainz.org/_matrix/media/v3/download/chatbrainz.org/FDsXKlCiILvYPvUYeJtNNaYM>)
2025-05-21 14138, 2025
bitmap[m]
lucifer: one issue that I did notice in my IRC logs recently: all of the "matrix.chatbrainz.org" links (for uploaded images and text snippets) are 404s
2025-05-21 14110, 2025
yvanzo[m]
bitmap[m]: We should add it to the release notes if needed, and check icu versions for further releases.
2025-05-21 14110, 2025
lucifer[m]
bitmap: yes that's an unfortunate result of matrix authenticated media where the url is not enough to access the media item.
2025-05-21 14154, 2025
lucifer[m]
i think Jade's is handling that in the matrix archiver project but she can confirm.
2025-05-21 14104, 2025
lucifer[m]
s/'s//
2025-05-21 14146, 2025
Jade[m]
Yeah you ahve to be logged into a matrix account on the server t oview media nowadays, 'authenticated media'
2025-05-21 14108, 2025
Jade[m]
Mainly to avoid your server being used to distribute illegal media / gore / etc
2025-05-21 14149, 2025
Jade[m]
Discord made the same change for the same reason a bit before matrix did
