#metabrainz

/

      • ruaok
        zas where on taz are the ftp files?
      • 2016-05-24 14554, 2016

      • zas
        /media/ftp if i remember
      • 2016-05-24 14544, 2016

      • zas
        /media/ftp/root/pub/musicbrainz/
      • 2016-05-24 14514, 2016

      • zas
      • 2016-05-24 14545, 2016

      • chirlu
        Historical reasons? Better too many than too few.
      • 2016-05-24 14531, 2016

      • zas
        Better the right number. Fuzzy coding doesn't mix well with C.
      • 2016-05-24 14559, 2016

      • ruaok
        I didn't change that line, did i?
      • 2016-05-24 14523, 2016

      • zas
        nope, it was here before
      • 2016-05-24 14503, 2016

      • ruaok
        ok, I didn't hear any objections to #1. bitmap and I will proceed.
      • 2016-05-24 14506, 2016

      • ruaok
        old packets removed.
      • 2016-05-24 14546, 2016

      • ruaok
        sorry, bad packets removed.
      • 2016-05-24 14522, 2016

      • mihaitish has quit
      • 2016-05-24 14514, 2016

      • ruaok
        ok, we're setting the replication number of have a gap in it to make it clear.
      • 2016-05-24 14521, 2016

      • ruaok
        we're now dumping a new dump
      • 2016-05-24 14503, 2016

      • JesseW has quit
      • 2016-05-24 14531, 2016

      • Freso
        ruaok: Wait, so you're going with #1? You were leaning towards and #2 and zas said #2 was only realistic option and chirlu didn't comment... ? I think I missed a step somewhere. :)
      • 2016-05-24 14550, 2016

      • chirlu
        Modulo typo.
      • 2016-05-24 14513, 2016

      • ruaok
        yeah, chirlu is right. going with option 6.
      • 2016-05-24 14528, 2016

      • ruaok is operating with way too little sleep
      • 2016-05-24 14540, 2016

      • ruaok
        we're breaking replication.
      • 2016-05-24 14549, 2016

      • ruaok
        waiting for a dump now, then we can test an import.
      • 2016-05-24 14556, 2016

      • Freso
        Alright.
      • 2016-05-24 14528, 2016

      • ruaok
        zas: can we test an import without having to copy the thing out of DWNI?
      • 2016-05-24 14541, 2016

      • Mineo joined the channel
      • 2016-05-24 14509, 2016

      • zas
        what do you mean ?
      • 2016-05-24 14514, 2016

      • CatQuest
        m everyone!
      • 2016-05-24 14518, 2016

      • CatQuest
        !m everyone!
      • 2016-05-24 14518, 2016

      • BrainzBot
        You're doing good work, everyone!!
      • 2016-05-24 14522, 2016

      • ruaok
        copying the dump out will take very long time.
      • 2016-05-24 14528, 2016

      • ruaok
        +a
      • 2016-05-24 14532, 2016

      • ruaok
        maybe import onto lolo?
      • 2016-05-24 14539, 2016

      • ruaok
        drop the pg 9.5 db there and import?
      • 2016-05-24 14556, 2016

      • ruaok
        once the dump is done i want to get out of read only and start collecting more packets.
      • 2016-05-24 14502, 2016

      • zas
        yes, you may just copy over using nc or scp
      • 2016-05-24 14507, 2016

      • ruaok
        we need to test apply the packets before we unleash them on the world again
      • 2016-05-24 14513, 2016

      • ruaok
        ok, sounds good.
      • 2016-05-24 14529, 2016

      • ruaok
        the BCN team is then going to take a break for an hour while we wait for the dump to complete.
      • 2016-05-24 14548, 2016

      • CatQuest
        :(
      • 2016-05-24 14522, 2016

      • Freso
        ruaok: I think you should take the rest of the week off once this has been dealt with.
      • 2016-05-24 14526, 2016

      • Freso
        (Srsly.)
      • 2016-05-24 14541, 2016

      • ruaok
        I wish. armalcolite needs my attention.
      • 2016-05-24 14505, 2016

      • ruaok
        I'm off to play in italy next week. I'll take it then.
      • 2016-05-24 14514, 2016

      • Freso
        Okay, good.
      • 2016-05-24 14504, 2016

      • regagain joined the channel
      • 2016-05-24 14557, 2016

      • Slurpee has quit
      • 2016-05-24 14509, 2016

      • kanha has quit
      • 2016-05-24 14557, 2016

      • ruaok
        and next monday I might be kinda scarce too. my late term adopted brother is in town. :)
      • 2016-05-24 14512, 2016

      • UmkaDK
        Guys, I know you've got your hands full, so just a quick question
      • 2016-05-24 14530, 2016

      • UmkaDK
        Has the new schema been officially released?
      • 2016-05-24 14539, 2016

      • UmkaDK
        I've seen the merger commits in the "production" branch but there hasn't been a post on the blog and I can't find if there is anything special we need to do for this schema upgrade beyond running upgrade.sh
      • 2016-05-24 14512, 2016

      • yeeeargh joined the channel
      • 2016-05-24 14529, 2016

      • adhawkins
        UmkaDK: I believe they're still working on completing the scheme release on the 'live' servers. Once it's all working there, they will get round to issuing upgrade instructions.
      • 2016-05-24 14533, 2016

      • UmkaDK
        Fair play. Good luck guys!!
      • 2016-05-24 14554, 2016

      • reosarevok
        Freso: looks good actually, just publish :)
      • 2016-05-24 14557, 2016

      • chirlu
        UmkaDK: You won’t be able to upgrade directly because some changes were lost (didn’t make it into replication packets). The upgrade path will be to import a fresh dump, then resume replication from that.
      • 2016-05-24 14542, 2016

      • UmkaDK
        thanks chirlu! I suspected that this will not be a simple upgrade... Looking forward to those upgrade instructions, when you're ready. :)
      • 2016-05-24 14500, 2016

      • Leo_Verto has quit
      • 2016-05-24 14549, 2016

      • Freso
        reosarevok: I'm going to keep it in the over until Monday, probably.
      • 2016-05-24 14539, 2016

      • Slurpee joined the channel
      • 2016-05-24 14547, 2016

      • Freso
        reosarevok: I think these last few days have been quite full of announcements of various sorts and with a schema change changelog thing going on the blog today (... or maybe tomorrow? :p), I'll let it sit and stir for a bit.
      • 2016-05-24 14558, 2016

      • reosarevok
        Fair
      • 2016-05-24 14509, 2016

      • Freso
        It's already... 8 months late.
      • 2016-05-24 14523, 2016

      • Freso
        So a few more days are not going to make that much of a difference. :)
      • 2016-05-24 14554, 2016

      • diana_olhovyk_ joined the channel
      • 2016-05-24 14548, 2016

      • ruaok
        Gentlecat, alastairp: dinner tonight? I think we can swing it.
      • 2016-05-24 14553, 2016

      • ruaok
        hellska doesn't seem to be around...
      • 2016-05-24 14529, 2016

      • Gentlecat
        yes, I could use some food
      • 2016-05-24 14509, 2016

      • ruaok
        do you know how to ping hellska?
      • 2016-05-24 14534, 2016

      • Gentlecat
        I have his watsapp
      • 2016-05-24 14542, 2016

      • ruaok
        ping him, plz.
      • 2016-05-24 14543, 2016

      • Gentlecat
        let me ask if he's around
      • 2016-05-24 14547, 2016

      • ruaok
        k
      • 2016-05-24 14555, 2016

      • ruaok
        bitmap is out for a ride, but you probably know that.
      • 2016-05-24 14545, 2016

      • Gentlecat
        yeah, he's downstairs waiting for me to pick up the bike
      • 2016-05-24 14521, 2016

      • kanha joined the channel
      • 2016-05-24 14503, 2016

      • Gentlecat
        ruaok: he has a flight tomorrow at 6 in the morning and needs to do some things today
      • 2016-05-24 14521, 2016

      • ruaok
        not gonna stop bitmap from coming to dinner. :)
      • 2016-05-24 14523, 2016

      • Gentlecat
        sent his greetings to everyone though
      • 2016-05-24 14529, 2016

      • ruaok
        though his flight is at 10am.
      • 2016-05-24 14533, 2016

      • ruaok
        k, thanks for asking. :)
      • 2016-05-24 14541, 2016

      • Gentlecat
        and he doesn't have to move stuff :)
      • 2016-05-24 14557, 2016

      • Leo_Verto joined the channel
      • 2016-05-24 14504, 2016

      • MajorLurker has quit
      • 2016-05-24 14548, 2016

      • MajorLurker joined the channel
      • 2016-05-24 14536, 2016

      • Slurpee has quit
      • 2016-05-24 14501, 2016

      • diana_olhovyk__ joined the channel
      • 2016-05-24 14511, 2016

      • alastairp
        ruaok: sorry, already at black lab with a upf thing
      • 2016-05-24 14541, 2016

      • diana_olhovyk_ has quit
      • 2016-05-24 14528, 2016

      • Zastai joined the channel
      • 2016-05-24 14519, 2016

      • Slurpee joined the channel
      • 2016-05-24 14548, 2016

      • Mineo has quit
      • 2016-05-24 14559, 2016

      • derobert_work
        (read through the logs) Thanks for all the work! shmall/shmmax are just limits, having them too high won't break anything (well, unless something actually allocates that shared memory). You can check with ipcs -u (for a summary) or ipcs -m (for a list). In fact, the default on 4.4 on my desktop is almost 16 exabytes...
      • 2016-05-24 14533, 2016

      • diana_olhovyk__ has quit
      • 2016-05-24 14535, 2016

      • derobert_work
        You can set the core dump location with kernel.core_pattern — that way you can find the dump
      • 2016-05-24 14506, 2016

      • derobert_work
        to get a dump, you'll probably have to change ulimit -c, and maybe also set fs.suid_dumpable = 1 (not sure if pg is suid)
      • 2016-05-24 14550, 2016

      • derobert_work
        kernel.core_pattern = /var/crash/core.%u.%t.%p_%e
      • 2016-05-24 14550, 2016

      • derobert_work
        fs.suid_dumpable = 1
      • 2016-05-24 14524, 2016

      • derobert_work
        ... that's what I have on old box here. Docs for that are in the kernel docs, in sysctl/kernel.txt.gz — appears to still be valid as of 4.4.
      • 2016-05-24 14554, 2016

      • derobert_work
        though as of 4.4, there is a suid_dumpable = 2 which is preferable. Obviously /var/crash should be mode 0700.
      • 2016-05-24 14557, 2016

      • mamocs joined the channel
      • 2016-05-24 14545, 2016

      • zas
        it wasn't about kernel core dumps ;) but postgresql ones
      • 2016-05-24 14546, 2016

      • mamocs
        Trying to understand how url-rels maps to the database schema.. specifically I'd like to be able to query for an Artist and get the relevant url-rels via the tables but I think I must be missing a step or two..
      • 2016-05-24 14544, 2016

      • mamocs
        In the `url` table, I see gid.. I'm assuming there's some way to use the artist gid to get the set of url gids related to that artist?
      • 2016-05-24 14512, 2016

      • LordSputnik
        mamocs: generally the integer IDs are used for foreign keys in the DB
      • 2016-05-24 14552, 2016

      • LordSputnik
        mamocs: in this case, you can use the l_artist_url table to join the artist.id and url.id
      • 2016-05-24 14506, 2016

      • mamocs
        LordSputnik: Ah - got it. So basically it would look like select artist_id from l_artist_url and then use that to select into url (e.g. entity1)
      • 2016-05-24 14515, 2016

      • LordSputnik
        mamocs: yup, or something like "SELECT u.* FROM l_artist_url AS r LEFT JOIN url AS u ON u.id = r.entity1 LEFT JOIN artist AS a ON a.id = r.entity0 WHERE a.gid = 'uuidhere';"
      • 2016-05-24 14508, 2016

      • mamocs
        LordSputnik: Super helpful. Thanks!
      • 2016-05-24 14510, 2016

      • LordSputnik
        mamocs: https://github.com/metabrainz/musicbrainz-server/… is the actual table definition in case that's helpful for you :)
      • 2016-05-24 14545, 2016

      • UmkaDK has quit
      • 2016-05-24 14514, 2016

      • mihaitish joined the channel
      • 2016-05-24 14520, 2016

      • ariscop joined the channel
      • 2016-05-24 14557, 2016

      • derobert_work
        zas: the kernel makes core dumps, regardless of which process is crashing. So that's how you control where postgres core dumps go
      • 2016-05-24 14553, 2016

      • diana_olhovyk_ joined the channel
      • 2016-05-24 14507, 2016

      • ariscop has quit
      • 2016-05-24 14522, 2016

      • bitmap
      • 2016-05-24 14517, 2016

      • bitmap
        importing 20160524-160045 on git checkout c2f5a94bf60870e87677d3c77d4a16220dfdb3ee
      • 2016-05-24 14525, 2016

      • yeeeargh has quit
      • 2016-05-24 14542, 2016

      • bitmap
        ok, 20160512-mbs-8719.sql wasn't added to upgrade.json
      • 2016-05-24 14538, 2016

      • bitmap
        ran that manually on totoro
      • 2016-05-24 14520, 2016

      • regagain has quit
      • 2016-05-24 14513, 2016

      • zas
        so, what are the plans now ?
      • 2016-05-24 14510, 2016

      • bitmap
        doing a new dump now, then we'll test importing it again
      • 2016-05-24 14519, 2016

      • Freso
        reosarevok: "una ciudad Catalana" or "una ciudad catalana"?
      • 2016-05-24 14534, 2016

      • diana_olhovyk_ has quit
      • 2016-05-24 14500, 2016

      • ariscop joined the channel
      • 2016-05-24 14534, 2016

      • ruaok
        ok, who wants to help test something real quick?
      • 2016-05-24 14551, 2016

      • derobert_work
        I can try
      • 2016-05-24 14537, 2016

      • ruaok
        ok, hang on.
      • 2016-05-24 14552, 2016

      • mihaitish has quit
      • 2016-05-24 14505, 2016

      • ruaok
        we've taken beta out of read only to test edits: https://beta.musicbrainz.org
      • 2016-05-24 14517, 2016

      • ruaok
        if you've got an edit you've been meaning to make, now is the time. :)
      • 2016-05-24 14523, 2016

      • ruaok
        remember, this is the live database!
      • 2016-05-24 14534, 2016

      • derobert_work
        ok, I have some
      • 2016-05-24 14503, 2016

      • MajorLurker
        Want more added in? I can add some if you want
      • 2016-05-24 14528, 2016

      • Leo_Verto
        Great, logging in right now!
      • 2016-05-24 14554, 2016

      • derobert_work
        ruaok: tried to add cover art, got "Error creating edit"
      • 2016-05-24 14500, 2016

      • derobert_work
      • 2016-05-24 14542, 2016

      • derobert_work
        should I try an album edit as well?
      • 2016-05-24 14552, 2016

      • ruaok
        try making a non album are edit, please.
      • 2016-05-24 14545, 2016

      • ruaok
        MajorLurker: go.
      • 2016-05-24 14506, 2016

      • Leo_Verto
        successfully added a group
      • 2016-05-24 14554, 2016

      • derobert_work
      • 2016-05-24 14526, 2016

      • ruaok
        woah. you guys had some pent up edits. :)
      • 2016-05-24 14532, 2016

      • ruaok
        looking promising so far.
      • 2016-05-24 14534, 2016

      • Freso
        *arthttps://beta.musicbrainz.org/search/edits?conditions.0.args.1=38897068&conditions.0.operator=BETWEEN&conditions.0.field=id&conditions.0.args.0=38897067
      • 2016-05-24 14542, 2016

      • Freso
      • 2016-05-24 14554, 2016

      • ruaok
        we need to wait for the last full dump to finish dumping before we can complete the test.
      • 2016-05-24 14525, 2016

      • MajorLurker
      • 2016-05-24 14549, 2016

      • derobert_work
        although it looks to be displaying a little funny (not sure if that's the way it's always been, though)