[10:32] <oxide94> Hi all
[10:32] <oxide94> I'd like to add a new package to Ubuntu
[10:32] <oxide94> I'm reading https://wiki.ubuntu.com/UbuntuDevelopment/NewPackages
[10:33] <oxide94> and the 1st point is to post here, so I'm posting :)
[10:35] <oxide94> I am a member of MooseFS Team, and we'd love to add MooseFS to Ubuntu
[10:35] <oxide94> we already have our own repository with packages
[10:37] <oxide94> could somebody help me?
[11:22] <oxide94> anybody?
[11:35] <Rhonda> oxide94: Do you have started with it in a PPA already, or on your own site?
[11:36] <Rhonda> oxide94: And even better option would be to upload it to Debian and not Ubuntu, to make it available to even more people.  Ubuntu will sync from there then.
[11:39] <oxide94> Hi @Rhonda, now we don't have Launchpad PPA, but it is on roadmap to add
[11:40] <oxide94> about Debian - I know
[11:40] <oxide94> but there's a problem with adding MooseFS to Debian, 'cause there's  LizardFS available in Debian (a fork of MooseFS)
[11:41] <oxide94> and the LizardFS' maintainer objected to adding MooseFS, so probably it will be a long way
[11:41] <Rhonda> You can do a personal PPA, not a project PPA for a start.
[11:41] <Rhonda> They can object to it, but if you maintain it there's little they can do.
[11:41] <oxide94> http://ppa.moosefs.com/stable/
[11:41] <Rhonda> Debian is a do-ocracy -- who does stuff decides what's done.
[11:42] <Rhonda> Were the reasons to keep moosefs out valid?
[11:42] <oxide94> there were some things we're working on now, like public bugtracker
[11:42] <oxide94> or source code repository different than this https://moosefs.com/download/sources.html (e.g. GitHub)
[11:43] <oxide94> and we're working on it
[11:43] <oxide94> but our, let's say, goal
[11:43] <oxide94> is to add MooseFS before FeatureFreeze / DebianImportFreeze
[11:44] <Rhonda> Well, that's nice reasons, but no show-stoppers, there are other packages that don't have that neither.
[11:44] <oxide94> and problably it may be simple, like marking MooseFS' and LizardFS' packages conflicting to each other
[11:44] <Rhonda> You should work on that then rather earlier than later
[11:44] <oxide94> We're working on it, probably we'll be uploading sources into GitHub
[11:44] <oxide94> yes, we're doing it now ;)
[11:44] <Rhonda> You think they need to conflict?  Can't they co-exist (don't know anything about either, just wondering)?
[11:45] <oxide94> You know, both are filesystems
[11:45] <Rhonda> So?
[11:45] <Rhonda> ext4 doesn't conflict with brtfs neither.
[11:45] <oxide94> yes, yes
[11:45] <oxide94> let me eplain :)
[11:45] <Rhonda> cat /proc/filesystems.  There is no reason for them to conflict? :)
[11:46] <oxide94> they're both distributed filesytem, LizardsFS is a fork of an old, outdated and unsupported MooseFS version (1.6.x)
[11:46] <oxide94> they actively develop LizardFS, and we actively develop MooseFS
[11:46] <oxide94> in the mean time
[11:46] <oxide94> there were made some changes, like metadata format
[11:47] <oxide94> and it may be a security risk, to try to install them both on the same machines
[11:47] <Rhonda> Do they register as the same fs?
[11:47] <oxide94> I'm not sure what you're asking about
[11:48] <Rhonda> Ok, if they really conflict on that level then it's some kinda bad fork approach they took, not nice.  But seemingly nothing you can do.
[11:48] <Rhonda> If they go by the same fs name underneath, or are recognized as the same fs by the kernel.
[11:49] <oxide94> you know, LizardFS didn't changed their binaries' names
[11:49] <oxide94> like mfsmaster, mfschunkserver, mfsmount
[11:49] <oxide94> etc
[11:49] <oxide94> we have mfs(xxx), which is obvious, they should have lfs(xxx), which is also obvious
[11:50] <oxide94> but they don't want to change it
[11:50] <Rhonda> *sigh*
[11:50] <Rhonda> Well, yeah, I can see the pain/difficulty to get that resolved.
[12:01] <oxide94> exactly
[12:02] <oxide94> @Rhonda: if you have some time, please take a look at this https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=810822 [Debian ITP for MooseFS] and this https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=810853 [Debian RFS for MooseFS]
[12:03] <oxide94> I posted them some time ago
[12:14] <Rhonda> oxide94: The "undesirable overhead from having two similar software products" is a bit of moot.  As long as there are people willing to maintain the things there's nothing wrong with it.  There are lots of similar packages, mail clients/browsers, but especially libraries which are very much more similar in what they provide than different mail clients, so …  that can be safely ignored.
[12:17] <oxide94> exactly :)
[12:18] <oxide94> but you know
[12:18] <Rhonda> But if the transparency and following open source is a valid critic, you should work on that.
[12:18] <oxide94> Rhonda: I'm of course aware of the advantages of putting the package to Debian at first
[12:18] <oxide94> yes, we are
[12:19] <oxide94> ok, could you tell me more about this "transparency"?
[12:19] <oxide94> I'm not sure what does it mean exactly
[12:26] <oxide94> Rhonda: ok, let's assume, that we'll improve some things like public bug tracker, packages (lintian checks etc.), sources on GitHub
[12:27] <oxide94> what should we do next?
[12:28] <oxide94> I'd like to put packages into Ubu before FeatureFreeze date
[12:43] <Rhonda> When is that?
[12:44] <Rhonda> Given that for Ubuntu it's much earlier than for Debian, I'm uncertain on that grounds.  I've always got my packages through Debian into Ubuntu (and it's also the suggested way).  About time difficulties I'm not the right person to suggest you something, sorry.
[13:05] <rbasak> oxide94: interesting. Your situation is similar to MySQL/MariaDB, but not the same.
[13:05] <rbasak> MySQL and MariaDB work together within the same maintenance team in Debian and work to make sure their packages interact on Debian well.
[13:06] <rbasak> oxide94: on the other hand you've been accused of not having Debian packaging in shape.
[13:07] <rbasak> oxide94: your difficulty I think is that you aren't already a Debian or Ubuntu developer able to handle any needed packaging fixes. Unfortunately that takes a ton of time and experience, especially in the forked situation you have.
[13:08] <rbasak> oxide94: it seems to me that volunteer sponsors aren't going to be able to spare the time you need to make your deadline.
[13:09] <rbasak> oxide94: you could engage an existing Debian developer or Ubuntu developer to help you. Even then it's tight.
[13:09] <rbasak> That's how it looks to me, anyway. I hope that's helpful.
[13:14] <oxide94> rbasak: yes, it is, thanks
[13:17] <oxide94> @Rhonda: I mean Ubuntu FeatureFreeze, Feb 18th
[13:18] <oxide94> rbasak: about debian packaging "in shape" - I checked our packages using lintian and it issued, let's say, minor warnings which are easy to fix (like outdated Free Software Foundation postal address ;)
[13:19] <rbasak> oxide94: there are plenty of things that lintian is not capable of picking up.
[13:20] <oxide94> of course there are many more things, and some of them are more important
[13:20] <oxide94> Ok, I see
[13:20] <oxide94> So... should we try to do everything to go through Debian and be imported automatically to Ubuntu?
[13:21] <rbasak> I haven't looked at your packaging, so I cannot judge. But I do deal with packaging created by non-Debian developers quite often, and I have very rarely seen any that isn't somehow flawed.
[13:21] <oxide94> ok
[13:22] <rbasak> I'm biased. I work for Canonical. I'd say that the only chance you have is to engage Canonical. I should point out that this isn't a requirement; for Ubuntu all you need is competent packaging, consensus amongst Ubuntu devs (not Canonical) that it is good, an existing Ubuntu developer to upload, and an Ubuntu archive admin to accept. None of this requires Canonical.
[13:23] <rbasak> But for you to achieve that is potentially a very tall order.
[13:23] <rbasak> I have undertaken this kind of work before, and it has taken months and multiple cycles.
[13:23] <rbasak> In time you can become an Ubuntu developer and handle most of this yourself.
[13:24] <rbasak> The issue is your deadline.
[13:24] <oxide94> Yes...
[13:24] <oxide94> and especially that this is LTS release...
[13:24] <rbasak> OTOH, you can seek to get it into Debian, and in that case it'll automatically get synced to Ubuntu if it's in before Ubuntu's feature freeze. But that seems even more unlikely to happen in time.
[13:24] <oxide94> exactly.
[13:24] <rbasak> (to your deadline)
[13:27] <oxide94> at the moment I managed to add MooseFS to FreeBSD :)
[13:27] <oxide94> of course, there were some problems etc.
[13:27] <Rhonda> oxide94: Putting it into a personal PPA (doesn't need a group PPA) would be a big start.  That way you can already ask people to look at it.
[13:27] <oxide94> Ok, great, I'll start with that
[13:28] <oxide94> and then... post here one more time, or maybe there's something like ITPs or RFSes like in Debian?
[13:36] <rbasak> You can file a [needs-packaging] against Ubuntu (no particular package) and tag it needs-packaging, Wishlist. That's the equivalent of an ITP, but isn't a requirement. It's also a place to track an RFS.
[13:36] <rbasak> If you do create one, please make sure to link to the Debian bugs, as the first question any sponsor will ask is "why can this not be done in Debian?"
[13:37] <rbasak> If you want to put a proposed upload into the sponsorship queue, see https://wiki.ubuntu.com/SponsorshipProcess
[13:56] <oxide94> ok, many thanks!
[18:46] <Punkoivan> guys, I moving to devops engineer, can anyone give me good book or maybe online resource to learn about it?
[18:48] <Punkoivan> Which config manager easy to learn for novice? I read that it's ansible, doesn't?
[22:32] <thebwt> ... this is probably not the right channel for that