that should return the "Number of bytes used to store a particular value (possibly compressed)"
lucifer
bitmap, atj: there is no index on track.name in prod afais
atj
lucifer: well that would explain it :P
bitmap was just checking that
do we have any indexed text / unlimited varchar fields?
actually, the insert would just fail
bitmap
we might
atj
so not a DoS
lotooo
Does your prod missing an index or does my setup have an extra one ?
bitmap
the error message doesn't indicate which index is triggering the error AFAICT
lucifer
lotooo: `\d musicbrainz.track` in your local database and share the output probably.
TOPIC: MetaBrainz Community and Development channel | MusicBrainz non-development: #musicbrainz | BookBrainz: #bookbrainz | Channel is logged and not empty as it is bridged to IRC; see https://musicbrainz.org/doc/IRC for details | Agenda: Reviews, Group Contacts on IRC (yvanzo)
using psql that is.
MonkeyPython
heh
yvanzo
MonkeyPython: your issue cannot be resolved beforehand ^
bitmap
it could be on a table other than track, I suppose, since we have some mirror triggers than run on track updates
liberastaff member was going ot poke some people about it
bitmap
lotooo: did you create this index yourself? because we used to have one with the same name, but it was dropped years ago.
MonkeyPython
so first we gotta figure out how to put reo in that list
bitmap
but I guess that may explain why nobody else has come forth to report replication issues (?)
reosarevok
yvanzo: I just asked a bit ago, yeah :) Still nothing coming out of it. Maybe I just need to send the email :)
lotooo
nope. we have a lot of mirrors at plex and we don't touch the database. We just grab the db from you guys and then replicate. But we built that setup years ago. We just update it to new versions but I usually don't rebuild nodes from scratch
last time I rebuilt everything was during the pg upgrade to pg12
yvanzo
reosarevok: my intend was to ask for volunteers to join the list that we would mail to the staff
reosarevok
Yes, that seems sensible too, anyway :)
MonkeyPython
++ that idea
bitmap
lotooo: strange, it's explicitly dropped in admin/sql/updates/20200518-pg12-before-upgrade.sql which should have been run during that upgrade
yvanzo
atj, bitmap: A similar title is set for the release, the track, and the recording.
MonkeyPython
the "public contacts" secton on /ChanServ INFO #metabrainz
yvanzo
yes
MonkeyPython
reo, and yvanzo, and possibly monkey seem like good candidates imho
yvanzo
they advise to have 2 or 4 contacts
MonkeyPython
good idea
yvanzo
volunteering is proposing yourself, not others :)
MonkeyPython
hahaha
lotooo
@bitmap : I just have to replay that script and call it a day ? (I'm planning on rebuilding our setup soon-ish anyway)
MonkeyPython
didn't want to seem too ostentatious
bitmap
lotooo: I wouldn't run that script by itself since it depends on steps in later scripts, but you can just drop the problematic index for now
lotooo
@bitmap which index specificatly ?
bitmap
DROP INDEX track_idx_txt;
lotooo
:thumbsup:
mmm that doesn't seem to have helped :thinking_face: triple_checking I am on the good server
bitmap
lotooo: my bad, track_idx_musicbrainz_collate too
lotooo
I see that track_idx_musicbrainz_collate is based on name as well
ah ok perfect
bitmap
yeah, it's alaso dropped in the script I mentioned
lotooo
Thanks that seemed to have helped ! Letting this server catch up and I'll execute the same on our other mirrors. Thanks for your help
bitmap
glad it's working š
yvanzo
bitmap: When you need it, you can try resuming the replication on wolf (it is only running at 03:00 UTC so it didnāt encounter this issue yet).
bitmap
yvanzo: thanks, I'll check
reosarevok
Should we then close the ticket and say that if anyone has similar issues they should drop those two indexes?
reosarevok: no because it is just a temporary workaround
reosarevok
Well, I thought the point was those were *not* supposed to exist anymore :)
From what bitmap said
yvanzo
I read āfor nowā
bitmap
yes, if the schema change was applied correctly they should not exist, although we may have other indexes on TEXT columns that exhibit this issue
sorry, "for now" was as opposed to "rebuilding our setup soon-ish"
yvanzo: I guess the mirror on wolf would only help checking that the upgrade scripts worked correctly, and only if it was originally built prior to the PG12 upgrade in 2020.
yvanzo
It isnāt that old.
yellowhatpro
hi @yvanzo, I'll be there!
Maxr1998_ joined the channel
Maxr1998 has quit
yvanzo
bitmap: any TEXT or VARCHAR actually, right?
bitmap
right
minimal joined the channel
reosarevok starts going back to Belgium
reosarevok
mayhem: I should be available for the meeting (waiting in the airport) but if I see something goes wrong I'll let you know so you can take over :)
mayhem
k.
reosarevok goes back to try to fit all this Asturian cheese in the bag š
MonkeyPython
:V that's what he said
yvanzo
reosarevok: Iāve been through the same hardshipsā¦ š§
bitmap: I checked that trying to enter such a title for a currently indexed field currently fails with error 500, it could be possibly improved but it isnāt at risk for mirrors. Are they other things I can help you with checking?
there
MonkeyPython
man there is literally no image of this bagpipe on commons, and image in the article, linked to wikidata etc. is just a regular Great Highlands....
any armenians that wanna take some image of a parkapzuk :DD
"We and ourĀ 699Ā partners store and/or access information on a device, such as unique IDs in cookies to process personal data."
Jesus, Discogs
bitmap
I just checked the JSON dump mirror on hendrix, and it's failing with the reported "index row ..." error
reosarevok
That's... a lot of partners
MonkeyPython
i'd usually not slutshame but
bitmap
however, this DB was already running PG12 in 2019, so I don't think I had actually run the developed upgrade scripts during the schema change
huhridge joined the channel
Sophist-UK has quit
Sophist-UK joined the channel
Lotheric_ joined the channel
kilos_8 joined the channel
Leftmost_ joined the channel
d4rk-ph0enix joined the channel
kepstin has quit
d4rkie has quit
Leftmost has quit
eharris has quit
kilos_ has quit
minimal has quit
Leftmost_ is now known as Leftmost
iconoclast_hero has quit
Lotheric has quit
kilos_8 is now known as kilos_
eharris_ joined the channel
minimal joined the channel
iconoclast_hero joined the channel
kepstin joined the channel
v6lur joined the channel
mayhem
reosarevok: should I be preparing for the meeting?
yvanzo
bitmap: I have a mirror with āThe collation in the database was created using version 153.88, but the operating system provides version 153.120.ā for testing MBS-13360.
