#theforeman

/

      • cartik has quit
      • Mandorath has quit
      • raphink has quit
      • raphink joined the channel
      • raphink has quit
      • raphink joined the channel
      • phileas has quit
      • KAM has quit
      • phileas joined the channel
      • swapab has quit
      • mmello joined the channel
      • jgarr joined the channel
      • chudler joined the channel
      • mmello has quit
      • annanee_ joined the channel
      • toordog joined the channel
      • toordog has quit
      • toordog joined the channel
      • mhulan has quit
      • toordog_ joined the channel
      • toordog_ has quit
      • annanee has quit
      • annanee_ has quit
      • aitrus
        Anyone have any ideas on how to resolve "Package does not match intended download" errors?
      • annanee joined the channel
      • There's a RH KB article for Sat 6 that says that it happens when the md5 of the package is different from what pulp thinks it should be
      • jsherrill
        aitrus: it could mean a couple different things
      • the most likely that i've seen is that there are two different packages with the same NVREA
      • in the repo
      • although that problem shouldn't be around anymore
      • i didn't think
      • aitrus: what version of katello?
      • aitrus
        3.2.2
      • jsherrill
        is this a repo you are syncing, or uploading rpms to (via hammer or the UI)
      • ?
      • aitrus
        I have both CentOS and OEL repos sync'd, if that matters
      • It's on packages that are being sync'd from upstream (CentOS in this case)
      • such as "sos"
      • jsherrill
        hrmm, okay, that issue shouldn't happen then, but lets double check it
      • can you find the primary.xml file in that repo locally on the foreman/katello server?
      • mmello joined the channel
      • aitrus
        Will it be part of the content view's directory structure?
      • jsherrill
        aitrus: yep
      • if your client is using a content view
      • pondrejk has quit
      • aitrus
        in the "master" directory or do I need to look in the "https" directory?
      • For example... /var/lib/pulp/published/yum/master/yum_distributor/TKC_Holdings-pnp-production-c7stable-centos-c7updates/1483135995.77/repodata
      • yum/master or yum/https
      • oh, those are symlinks to master
      • okay, found it :-D
      • kpease joined the channel
      • mhulan joined the channel
      • jsherrill: okay, so... in the primary.xml
      • there is a sha256sum
      • hprakash joined the channel
      • jsherrill
        do you see that rpm listed in there more than once?
      • TommyTheKid joined the channel
      • aitrus
        the package that the content view points to (the symlink in /varlib/pulp/content/units...") has a different sha256
      • jsherrill: negative on the multiple listings
      • however, if I pull the package from the repo directly - as in still from Katello but not from the content view - the sha256sum matches
      • jsherrill
        aitrus: interesting
      • TKersten has left the channel
      • toni_ has quit
      • Mandorath joined the channel
      • toni_ joined the channel
      • inecas_ joined the channel
      • lfisher
        jsherrill: I saw that you made some updates to katello to support Fedora 25, but I don't see any client repos for it, is that going to happen in the next release or is there other work to do?
      • jbberry joined the channel
      • jsherrill
        aitrus: you might try downloading the rpm and just placing it overtop of the one that is symlinked to in /var/lib/pulp/content/
      • tstrachota has quit
      • lfisher: it will be in the 3.3 release, we're working on the rc1 right now. It should appear in nightly, but our (and foreman's) nightly build is broken currently due to some packaging issues
      • inecas has quit
      • lfisher
        jsherrill: okay, sounds good, I'm looking to upgrade my desktops soon and wanted to know when to expect it, I'll keep an eye on nightly and the 3.3 release and if I don't see it in 3.3, I'll ping you again
      • inecas_ has quit
      • aitrus
        jsherrill: well, thing is that there are bunches of these, and I haven't figured out a good way to find them. :-\
      • jsherrill
        lfisher: the rpm itself is already built, if you want to grab it from koji: http://koji.katello.org/koji/buildinfo?buildID=...
      • inecas_ joined the channel
      • lfisher: since all the pulp deps are now in fedora, we don't actually carry anything else
      • dcaplan is now known as dcaplan_AFK
      • aitrus: hrmmm
      • lfisher
        jsherrill: cool, thanks
      • jsherrill
        aitrus: ohh, i missed the part about the checksum being correct outside the content view
      • can you republish the content view?
      • aitrus
        yes, but I have already :-(
      • that was Friday though
      • i'll create a new version again
      • maybe whatever caused this magically cleaned itself up over the weekend after i re-pub'd
      • teriyakichild has quit
      • nstrug_ joined the channel
      • nstrug has quit
      • oprazak has quit
      • akofink has quit
      • akofink joined the channel
      • annanee has quit
      • annanee joined the channel
      • nstrug__ joined the channel
      • dcaplan_AFK is now known as dcaplan
      • annanee has quit
      • nstrug_ has quit
      • teriyakichild joined the channel
      • smeyer has quit
      • phileas has quit
      • phileas joined the channel
      • jsherrill: no change after publishing the new CV :-(
      • jsherrill: actually, hold on, i thought this box was pointed at library but it isn't
      • jsherrill
        oh?
      • tamarin joined the channel
      • rcernin has quit
      • sjagtap joined the channel
      • attachmentgenie has quit
      • lero has quit
      • ungos_ has quit
      • tesseract has quit
      • amitKarsale joined the channel
      • tamarin has quit
      • outi has quit
      • aitrus
        jsherrill: okay, published the new CV version to the correct environment. still having the issue. however, i checked a different package (systemd) and it is listed twice in primary.xml
      • not sure if i just missed that last time or if i'm facing multiple issues
      • orabin has quit
      • jsherrill
        and you're sure its the same NVREA ?
      • name version release epoch and arch?
      • tamarin joined the channel
      • annanee joined the channel
      • pgagne is now known as pgagne|afk
      • dbrandon joined the channel
      • hprakash has quit
      • aitrus
        jsherrill: oh, not, it's not the same NVREA (different releases)
      • jsherrill
        k
      • aitrus: to clarify, the checksum in the Library repo is correct
      • in the primary.xml in the library repo
      • but in the content view its not?
      • aitrus
        jsherrill: let me double check since my eyes are crossed at this point :-D
      • jsherrill
        hehe okay
      • annanee has quit
      • toni_ has quit
      • jomitsch is now known as jomitsch-afk
      • aitrus
        jsherrill: something has changed. the "sos" package is now installing fine (it's the one that had different sha256sums in library vs cv)
      • jsherrill
        very strange
      • aitrus
        the systemd package, though, has the same sha256sum in both library and the CV, but primary.xml has a different sha256 in it for that package
      • amitKarsale has quit
      • jsherrill
        aitrus: are you using on_demand for these repos ?
      • as the download policy?
      • aitrus
        yes
      • right now the systemd package is being served by squid
      • i.e., the /var/lib/pulp symlink to that systemdpackage is broken, and requests for the package are being served by /streamer
      • jsherrill
        ahh, so you think maybe its broken while squid is serving it (for some reason)
      • but then once pulp 'imports' it
      • its correct?
      • that would match up with why the sos package magically started working
      • beav
        jsherrill: i am getting caught up still, but is the issue related at all to one rpm being signed and one unsigned (or different signature)?