ChanServ changed the topic of #freedesktop to: https://www.freedesktop.org infrastructure and online services || for questions about freedesktop.org projects, please see each project's contact || for discussions about specifications, please use https://gitlab.freedesktop.org/xdg or xdg@lists.freedesktop.org
Consolatis_ has joined #freedesktop
Consolatis has quit [Ping timeout: 480 seconds]
Consolatis_ is now known as Consolatis
ybogdano has quit [Ping timeout: 480 seconds]
ngcortes has joined #freedesktop
ximion has quit []
Rainer_Bielefeld_away has joined #freedesktop
ngcortes has quit [Remote host closed the connection]
pohly has joined #freedesktop
<bluepenquin> I got bit by that when submitting an issue on mesa/mesa, for whatever reason the captcha won't load so I had to reload the page and somehow lost the entire filled out template.
eroux has joined #freedesktop
danvet has joined #freedesktop
Rainer_Bielefeld_away has quit [Remote host closed the connection]
alanc has quit [Remote host closed the connection]
alanc has joined #freedesktop
MajorBiscuit has joined #freedesktop
eroux_ has joined #freedesktop
eroux has quit [Read error: Connection reset by peer]
chaim has joined #freedesktop
MrCooper has quit [Remote host closed the connection]
vbenes has quit []
vbenes has joined #freedesktop
pjakobsson has quit [Ping timeout: 480 seconds]
MajorBiscuit has quit [Ping timeout: 480 seconds]
<bentiss> \o/ after spending quite a lot of time, I can say that users with *only* snippets are spammers, and I can safely detect and delete them (got 2 exceptions, but worse case, they'll come back at us)
<bentiss> one of the exceptions pasted the snippet in other snippets services, so safe to delete
<bentiss> the other is... probably not so much important (autogen.sh which says s/m4_copy/m4_copy_force/ )
MajorBiscuit has joined #freedesktop
MrCooper has joined #freedesktop
Thaodan is now known as foobar
foobar is now known as Guest1642
Guest1642 has quit []
Thaodan has joined #freedesktop
___nick___ has joined #freedesktop
MajorBiscuit has quit [Ping timeout: 480 seconds]
MajorBiscuit has joined #freedesktop
<daniels> bentiss: oooh yeah, that's a really good metric
<bentiss> daniels: I am tempted to delete all 18000 users with no activity in the past 60 days, no project, no group, no notifications and no snippet
<daniels> prob not a bad idea
<daniels> you can also add coinbase + binance to your filters btw
<bentiss> good point
<bentiss> actually, only 12000 without notifications
<emersion> please make sure you don't delete users with just one comment in an issue
<bentiss> emersion: yeah, I am checking that they have no participation in any thread/comment/issue/MR
<emersion> cool
<bentiss> well... hopefully :)
MajorBiscuit has quit [Ping timeout: 480 seconds]
<emersion> select distinct(author_id) from issues
<emersion> i'm not sure this selects issue commenters?
<emersion> maybe there's a separate comment for MR/issue comments?
<daniels> emersion: yeah that's notes
<emersion> oh
<daniels> issues + MRs are their own thing at the root, then each comment is a note
<emersion> what does the "destroy user" function does?
<bentiss> I checked that comments and thread are covered by "notes"
<emersion> does it delete everything the user created?
<bentiss> emersion: yes, everything
<emersion> or do the comments go to a "ghost" user like on GitHub?
<bentiss> not so sure
<bentiss> probably a ghost
<bentiss> let me check
<bentiss> so yes, ghost is then used :)
thaller is now known as Guest1644
thaller has joined #freedesktop
<emersion> cool!
<emersion> that'll make me sleep a bit better :)
<daniels> right, delete user drops the user and makes everything authored by ghost; delete user & contributions removes all trace of them ever having been there
<emersion> so: do we need to explicitly destroy the snippets?
<daniels> so the latter would e.g. delete their snippets too
<bentiss> destroy also deletes all snippets and repos FWIW
<bentiss> I tested that while removing spam snippets
<emersion> ah, that sounds perfect for our use-case
<bentiss> yep
<emersion> and makes sense in a way
<bentiss> too bad this is not embedded in gitlab directly and that we have to run a job for it
<emersion> yeah…
<bentiss> daniels: any idea if VACUUM is run on our db from time to time?
* bentiss will start the big deletion of users this afternoon, once the current backup finished
MajorBiscuit has joined #freedesktop
<daniels> bentiss: ttbomk no
<bentiss> k, I'll try to run it this afternoon too after the "big cleanup"
Guest1644 has quit [Ping timeout: 480 seconds]
<daniels> ty :)
<bentiss> well, no guarantees I'll manage to do it
<emersion> bentiss: have you seen "[abuse #14939980] [Gandi] Your website has been compromised"?
<emersion> "compromised" is a bit of an overstatement, but oh well
<emersion> it's quite timely
<bentiss> emersion: nope, where is this?
MajorBiscuit has quit [Ping timeout: 480 seconds]
<emersion> bentiss: sitewranglers@
<bentiss> emersion: I must confess I am not subscribed to this one :)
<emersion> basically just a request to delete a few GitLab accounts
<emersion> oh
<bentiss> can you forward it to me (and also maybe add me to the list)
<emersion> i'll just delete the accounts then
<bentiss> ?
<bentiss> sure
<emersion> bentiss: sent you an invitation to sitewranglers@
<bentiss> thanks, received and applied for it
MajorBiscuit has joined #freedesktop
pjakobsson has joined #freedesktop
<bentiss> emersion: the gandi email doesn't show up in the archives
<emersion> hm
<MrCooper> bentiss: it was sent on February 28th FWIW
<bentiss> ouch :)
<emersion> i just received it… :S
<bentiss> I thought the "timely" meant that it came through now
<MrCooper> guess it was stuck in the moderation queue until today
<bentiss> ouch, we have *a lot* of binance accounts
MajorBiscuit has quit [Ping timeout: 480 seconds]
pjakobsson_ has joined #freedesktop
pjakobsson has quit [Ping timeout: 480 seconds]
eroux_ has quit []
eroux has joined #freedesktop
ximion has joined #freedesktop
eroux has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
eroux has joined #freedesktop
<bentiss> daniels: FYI, I am looking at https://www.datadoghq.com/blog/postgresql-vacuum-monitoring/ and autovacuum is enabled in the db -> "SELECT relname, last_vacuum, last_autovacuum FROM pg_stat_user_tables WHERE last_autovacuum IS NOT NULL;" returns 29 rows
<bentiss> the interesting bit is the disk size of tables
<bentiss> the interesting bit is the disk size of tables: ci_builds_metadata is 33 GB and ci_builds is 29 GB
<daniels> bentiss: mm yeah, vacuum doesn't actually shrink the available space - it just orphans rows and leaves their previous storage available for reuse
<bentiss> daniels: unless it frees an entire page, which I hoped would be the case when bulk-deleting users
<bentiss> but the user table is 21 MB, so nothing will be stripped out :(
<daniels> bentiss: ttbomk it doesn't make files sparse; the only case in which it reduces the size of the disk storage is when you lose a bunch of rows from the end
<emersion> logs should be very compressible, is there a way to enable compression on the builds table?
<bentiss> yep, so we are going to have to keep on an ever groing db, except when we need to bump the version :(
<bentiss> emersion: logs are not even stored in the db, it's just the CI build IDs, project/user/etc informations that are stored
<emersion> ugh
<emersion> that's quite a lot of data…
rgallaispou has joined #freedesktop
* bentiss starts the removal of > 12000 accounts
MajorBiscuit has joined #freedesktop
ximion has quit []
<bentiss> OK, found 3700 more useless account which never logged in while being created more than 60 days ago
ppascher has joined #freedesktop
<emersion> sounds like users like this one would be supposed to be caught by your script? https://gitlab.freedesktop.org/psmj921108
<emersion> they logged in via google it seems
<bentiss> emersion: yes, but it's less than 60 days ago, so not detected currently
<bentiss> and yes, OAuth bypasses admin validation, so the script
<emersion> ah
<bentiss> emersion: zlkitit seems to be part of the previous batch, except it was not. I'll further dig once I am done purging the ones I am doing
<bentiss> just FWIW, dashboard was giving ~50000 users this morning, we are at 33479 now (but I am not sure the count is correct because I removed less users than the difference)
<bentiss> emersion: zlkitit is part of the "last activity == never" batch, so it'll get destroyed in the nmxt hour or so
<emersion> ack!
<bentiss> mupuf: I saw that mupuf-api user was removed, I hoped it's not an issue for you :)
eroux has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
MrCooper has quit [Quit: Leaving]
MrCooper has joined #freedesktop
ybogdano has joined #freedesktop
ximion has joined #freedesktop
<mupuf> bentiss: we'll see :D I think I don't need it anymore
ximion has quit []
ybogdano has quit [Ping timeout: 480 seconds]
ybogdano has joined #freedesktop
MajorBiscuit has quit [Ping timeout: 480 seconds]
<bentiss> I think I've seen almost all now... They can be quite creative and not sure how we can detect all spam but:
<bentiss> - some are creating a project and just put a description (usually .vn people)
<emersion> am i the only one to experience very delayed GitLab emails?
<bentiss> - some are even trickier as there is nothing in the project but they create one or 2 issues with their spam (only .vn AFAICT)
<emersion> it's started a few days ago
<bentiss> - and some aredumb enough to put their "spam" project as private :)
<bentiss> emersion: I don't see any email task getting canceled, so I guess it might come from the smtp server
<emersion> delivery temporarily suspended: connect to mail.protonmail.ch[185.70.42.128]:25: Connec tion timed out
<emersion> hmmm
<emersion> guess it's on my side then!
<bentiss> emersion: I was about to say that I was receiving emails ASAP, and et voila :)
kem has quit [Ping timeout: 480 seconds]
kem has joined #freedesktop
jstein has joined #freedesktop
<Lyude> bentiss: so if a user confirms their email, does their account request just automatically go through?
ybogdano has quit [Read error: Connection reset by peer]
ybogdano has joined #freedesktop
ybogdano is now known as Guest1674
ybogdano has joined #freedesktop
Guest1674 has quit [Ping timeout: 480 seconds]
Haaninjo has joined #freedesktop
jstein has quit []
ybogdano has quit [Ping timeout: 480 seconds]
ybogdano has joined #freedesktop
immibis has quit [Ping timeout: 480 seconds]
immibis has joined #freedesktop
___nick___ has quit [Ping timeout: 480 seconds]
<bentiss> Lyude: yes, unless the user has one of the denylist keyword in the email, username and description
Consolatis is now known as Guest1682
Consolatis has joined #freedesktop
Guest1682 has quit []
ybogdano has quit [Ping timeout: 480 seconds]
pohly has quit []
jarthur has joined #freedesktop
danvet has quit [Ping timeout: 480 seconds]
danvet has joined #freedesktop
ybogdano has joined #freedesktop
jarthur has quit [Read error: Connection reset by peer]
chaim has quit [Quit: Konversation terminated!]
kchibisov_ has quit [Ping timeout: 480 seconds]
danvet has quit [Ping timeout: 480 seconds]
kchibisov has joined #freedesktop
Haaninjo has quit [Quit: Ex-Chat]
ngcortes has joined #freedesktop
Consolatis_ has joined #freedesktop
Consolatis is now known as Guest1690
Consolatis_ is now known as Consolatis