<aerozol[m]> "Hemang Mishra: I reviewed your..." <- Thanks!
2025-05-18 13816, 2025
HemangMishra[m]
I'll make the updates and get back to u asap.
2025-05-18 13813, 2025
HemangMishra[m]
HemangMishra[m]: jasje: would love reviews from your side as well !!
2025-05-18 13812, 2025
zas[m]
I restarted mb-mail-service-prod on selda, it was apparently doing nothing but using 100% cpu
2025-05-18 13839, 2025
zas[m]
Also keydb on selda is still very busy.
2025-05-18 13845, 2025
aerozol[m]
<zas[m]> "I restarted mb-mail-service-prod..." <- Me on some days
2025-05-18 13831, 2025
lucifer[m]
zas: same for the mail service container on yamaoka and its also currently marked as unhealthy. restarted it.
2025-05-18 13814, 2025
Maxr1998 joined the channel
2025-05-18 13837, 2025
Maxr1998_ has quit
2025-05-18 13838, 2025
zas[m]
bitmap: about keydb on selda, since it is used, it should be running on 2 different hosts, and also I noticed it is using only one thread (https://github.com/metabrainz/docker-server-confi…) I guess using 4 would make more sense
2025-05-18 13829, 2025
Kladky joined the channel
2025-05-18 13817, 2025
ajhalili2006 has quit
2025-05-18 13826, 2025
ajhalili2006 joined the channel
2025-05-18 13813, 2025
mayhem[m] has quit
2025-05-18 13807, 2025
_BrainzGit
[mb-rngpy] 14amCap1712 opened pull request #11 (03master…actions-publishing): Add GitHub Actions workflow for PyPI publishing and simply update.sh script https://github.com/metabrainz/mb-rngpy/pull/11
2025-05-18 13858, 2025
leonardo has quit
2025-05-18 13820, 2025
leonardo joined the channel
2025-05-18 13854, 2025
julian45[m] has quit
2025-05-18 13830, 2025
bitmap[m] has quit
2025-05-18 13845, 2025
bitmap[m] joined the channel
2025-05-18 13845, 2025
bitmap[m]
<zas[m]> "bitmap: about keydb on selda..." <- makes sense, I don't think I set those up but it was probably configured to use fewer resources when it was on aretha and never updated after the move
valkey seems to be the most followed and reliable fork of redis so far
2025-05-18 13817, 2025
tux0r
i genuinely wonder why people even want to have redis forks
2025-05-18 13833, 2025
tux0r
or redis, in the first place (postgresql covers all use cases pretty well IMO)
2025-05-18 13835, 2025
yvanzo[m]
there was a licensing issue IIRC
2025-05-18 13851, 2025
Jade[m] joined the channel
2025-05-18 13851, 2025
Jade[m]
<zas[m]> "I restarted mb-mail-service-prod..." <- Ooh, interesting. I wonder what caused that
2025-05-18 13855, 2025
tux0r
oh? haven't seen that
2025-05-18 13824, 2025
Jade[m]
Jade[m]: Probably to do with html2text tbh
2025-05-18 13830, 2025
tux0r
re "mb-mail-service-prod", sounds like mb-mail-no-service-prod ;o)
2025-05-18 13800, 2025
yano joined the channel
2025-05-18 13801, 2025
yvanzo[m]
:-?
2025-05-18 13852, 2025
julian45[m] joined the channel
2025-05-18 13852, 2025
julian45[m]
<yvanzo[m]> "there was a licensing issue IIRC" <- yes, there was a big one; in march of last year, redis relicensed its core code from BSD to two non-free licenses, at which time the linux foundation forked the last BSD-licensed version to valkey
2025-05-18 13835, 2025
julian45[m]
but with the release of redis 8 some or all of the core code is now under the AGPL
2025-05-18 13805, 2025
julian45[m]
valkey is still prominent in the key-value DB space though, and redis & valkey are probably truly competitive (or at least closer to it) at both of their respective latest versions now
2025-05-18 13833, 2025
julian45[m]
* but with the release of redis 8 at the beginning of this month, the core code is now under the AGPL
nah, valkey still is. for example redis still has proprietary extensions and it's still sponsored/developed by a private company
2025-05-18 13812, 2025
Jade[m]
forks with different licencing situations are inevitably gonna diverge over time
2025-05-18 13836, 2025
julian45[m]
from a risk standpoint, if there are no features/functionalities MeB relies on specific to keydb and unavailable in other key-value stores, sticking with keydb long-term does not seem likely to be a sustainable choice. of the remaining options, valkey is probably the best candidate due to its development and sponsorship model (nonprofit foundation), as well as its performance & general quality
2025-05-18 13823, 2025
julian45[m]
julian45[m]: along this line, snapchat was (perhaps still is, but unclear) the apparent sponsor of keydb
2025-05-18 13842, 2025
Jade[m]
Jade[m]: Both are going to develop features as needed by their customers/community and it won't necessarily be obvious what the differences are
[sir] 14yvanzo merged pull request #160 (03master…SEARCH-738): SEARCH-738 (idx), SEARCH-739 (idx), SEARCH-740 (idx): Support and search by medium MBIDs https://github.com/metabrainz/sir/pull/160
