My Name is Joseph Crim. I would like to be a part of the Fedora team. I
have several years experience working for one of the largest web hosting
providers in North America. My strengths are in web and database
administration on CentOS 4 and Fedora Core 6 systems.
I feel the open-source community has given me so much, and I want to do my
part to give back. Please feel free to email me if you would like more
information or if you have any specific questions.
for two months there has been no progress on a security ticket:
https://hosted.fedoraproject.org/projects/fedora-infrastructure/ticket/88https://koji.fedoraproject.org spits out an strange certificate instead of one
signed by an well known CA, e.g. Equifax. Can maybe someone who reads here
and did not notice this Security Bug fix this? In case there is no money
available for this, then please use at least a certificate from cacerct.org
instead of this imho nearly complete useless certificate. Also it is not very
wise to educate users (Fedora maintainers) to accept bad certificates in
Fedora's Infrastructure, so that in case there is a Man-in-the-middle attack,
e.g. on an conference with free wifi, the regarding maintainers will be
Well, we haven't even had koji for a full release cycle yet and we're
already worried about storage:
There are plans to implement a koji purger that gets rid of some of our
builds but I've not seen any estimates as to how much space this will
get us. I doubt, highly, that it will be more than half recovery.
Given the worst case scenario we need to purchase additional storage
immediately. Thats what this email is about.
Right now we're storing our stuff on some pretty quick netapps, which
stinks considering that koji is only accessing a small percentage of
that storage at any point in time and even then... its over NFS. I'd
like to propose we move to cheaper storage that is dedicated to just
koji. There are a number of options to do this, including different
models of netapps. I'm putting a spreadsheet together that takes in to
consideration price, how much we get and figures out how much it costs /
G / month over a 4 year period. If you guys know of anything to add to
the list just let me know.
Mike is in the process of tracking down some issues with our database at
the moment, so there may be some intermittent outages with some of our
web services, including:
- accounts system
I'm still working on finding all of the details but we will likely have
an outage of xen1 and xen2 next week for the memory upgrades. This will
likely be during the day and will affect some services. Specifically:
The only real "uh-oh" there is releng1 and I think during the day is
really a good time for releng1 so please correct me if I'm wrong.
Hello,everyone,nice to meet you.I feeI sorry of my poor
I'm a new comer.A student of ShangHai University who major in computer
Happy to join in the fedora.I am interested in linux,device driver,embedded
I am familiar with C++&ASM&JAVA and good in C.
My name is Laith Juwaidah, I'm 17 years old, I'm taking an HND in
I started using Linux almost a year ago, I started with kubuntu, then I
shifted to Gentoo, and just a few days ago I shifted to Fedora.
I really liked it, it's not as so-easy-to-use as kubuntu, nor is it as
do-a-lot-of-things-yourself as Gentoo.
I've been looking for a project to contribute to since I first started
using Linux, and now I think I found one :)
The thing is that I don't know what I should do now :D
What do you think of changing the Fedora Apache Test Page? I've seen many
emails come in asking about domains. We need to make it clear that we don't
have anything to do with the site except for the OS.
16:00:57 -!- mmcgrath changed the topic of #fedora-meeting to: Fedora Infrastructure -- Role Call
16:00:59 < mmcgrath> who's here?
16:01:12 * ivazquez looks around
16:01:13 * ricky is.
16:01:14 < jcollie> hey, i'm actually here for once :)
16:01:43 < mmcgrath> abadger1999 dgilmore f13 ivazquez jcollie mdomsch paulobanon ricky skvidal +
anyone I've missed ping
16:01:54 < abadger1999> pong
16:02:37 < jcollie> ricky, what worked?
16:03:01 < ricky> The warren and jeremy not dropping on "/me "
16:03:10 < jcollie> ah
16:03:18 < mmcgrath> k, lets get started.
16:03:22 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Tickets
16:03:28 < mmcgrath> .tiny
16:03:28 < mmcgrathbot> mmcgrath: http://tinyurl.com/yth34b
16:03:46 < mmcgrath> First ticket is 154 - DNS.
16:04:03 < mmcgrath> Most of this is done in regards to the new setup
16:04:10 < mmcgrath> vpn that is
16:04:44 < jcollie> i've been out of it, what's being used for the vpn?
16:05:04 < mmcgrath> jcollie: in the near future just about everything, we've started decentralizing our
16:05:23 < jcollie> nono, i mean, openvpn or ipsec
16:05:23 < mmcgrath> This has caused our network design to get more complicated but in general it will
make systems administration easier.
16:05:28 < ricky> OpenVPN, I think.
16:06:12 < mmcgrath> <nod> openvpn
16:06:13 < jcollie> ok, i use openvpn myself to vpn back home
16:06:33 < mmcgrath> I've also added header information to our app and proxy servers so we can see which
servers were hit and how long it took.
16:06:40 < mmcgrath> for example:
16:06:57 < mmcgrath> wget -SO/dev/null https://hosted.fedoraproject.org/projects/smolt
16:07:15 < mmcgrath> or
16:07:19 < mmcgrath> .headers https://hosted.fedoraproject.org/projects/smolt/
16:07:20 < mmcgrathbot> mmcgrath: apptime: D=345385, content-length: 9641, set-cookie:
trac_session=f3754ce4b4ca65ece51f2ae2; expires=Wed, 09-Jan-2008 20:07:20 GMT;
Path=/projects/smolt;, proxytime: D=348694, expires: Fri, 01 Jan 1999 00:00:00
GMT, connection: close, server: Apache/2.2.3 (Red Hat), appserver:
app2.fedora.phx.redhat.com, proxyserver: (1 more message)
16:07:23 < ricky> Will we have .fedoraproject.org in the search path at some time? (Just a tiny quibble)
16:07:40 < ricky> .more
16:07:41 < mmcgrathbot> ricky: Error: You haven't asked me a command; perhaps you want to see someone
else's more. To do so, call this command with that person's nick.
16:07:43 < mmcgrath> ricky: once all the vpn stuff is done we'll have vpn.fedoraproject.org in the
16:07:46 < mmcgrath> .more
16:07:46 < mmcgrathbot> mmcgrath: proxy2.fedora.phx.redhat.com, cache-control: must-revalidate, date:
Thu, 11 Oct 2007 20:07:20 GMT, content-type: text/html;charset=utf-8
16:07:50 < mmcgrath> ricky: :)
16:07:54 < ricky> Aha, OK :)
16:08:04 < mmcgrath> I'll try to get those on two lines right by eachother.
16:08:09 < mmcgrath> .headers https://hosted.fedoraproject.org/projects/smolt/
16:08:10 < mmcgrathbot> mmcgrath: apptime: D=333311, content-length: 9641, set-cookie:
trac_session=ca704f123bbbdfe401559eba; expires=Wed, 09-Jan-2008 20:08:10 GMT;
Path=/projects/smolt;, proxytime: D=336297, expires: Fri, 01 Jan 1999 00:00:00
GMT, connection: close, server: Apache/2.2.3 (Red Hat), appserver:
app2.fedora.phx.redhat.com, proxyserver: (1 more message)
16:08:11 < mmcgrath> .more
16:08:12 < mmcgrathbot> mmcgrath: proxy1, cache-control: must-revalidate, date: Thu, 11 Oct 2007
20:08:09 GMT, content-type: text/html;charset=utf-8
16:09:04 < ivazquez> Why didn't that one give a FQDN for the proxy?
16:09:10 < abadger1999> Those headers will be a lifesaver the next time we have issues.
16:09:16 < abadger1999> very nice.
16:09:24 < mmcgrath> Ok, so that last request hit appserver app2, and proxy server proxy1. Total time
to generate (proxy time) was 336297ms. Of that, the app server was 333311ms was
the app server.
16:09:29 < mmcgrath> abadger1999: thats the idea :)
16:09:40 < mmcgrath> ivazquez: its using $HOSTNAME so that box probably is configured differently for
16:10:05 < ivazquez> Worth adding a ticket for?
16:10:11 < mmcgrath> anywho, I wanted to make sure to share that with everyone as it will make
16:10:24 < mmcgrath> ivazquez: up to you, its worth taking a look to make sure they're both configured
the same way.
16:11:05 < mmcgrath> This is also something we can use to have our users help us troubleshoot, I've got
a firefox plugin that displays the app server and proxy server in my status bar for
every fedora page.
16:11:19 < mmcgrath> Anyone have any questions on that or DNS? If not we'll move on.
16:11:37 < ricky> Cool at the Firefox plugin.
16:12:10 < mmcgrath> ricky: its a combo of Live HTTP Headers and Header Spy
16:12:27 < jcollie> i saw a bit on securing the DNS, is that something like
16:12:49 < mmcgrath> This will also give us some solid numbers as to how much time our proxy servers are
adding. Which, so far, has been very small.
16:12:58 -!- BhanuPatia1 [n=bhanu(a)dialpool-210-214-122-102.maa.sify.net] has quit "Leaving."
16:13:31 < mmcgrath> jcollie: could be handy. If you think its something we could use after a bit of
research let us know.
16:13:41 < mmcgrath> Ok, I'll move on. there's really not that much to talk about this meeting I think.
16:13:42 < ricky> Maybe I'm being silly, but isn't 336297ms >5 minutes?
16:13:56 -!- BhanuPatia1 [n=bhanu(a)dialpool-210-214-122-102.maa.sify.net] has joined #fedora-meeting
16:14:02 < jcollie> yeah i don't follow that word for word but it helps with locking things down
16:14:19 < ivazquez> I think those might be in ns.
16:14:27 < ricky> Aha, OK :)
16:14:35 < ivazquez> Or us, rather.
16:14:55 < ricky> That's much better.
16:15:00 < ivazquez> (I hope they're not jiffies...)
16:15:19 < mmcgrath> ivazquez: looks like microseconds.
16:15:28 < mmcgrath> http://www.csgnetwork.com/timemath.html
16:15:37 < mmcgrath> Ok, next ticket is #170
16:15:44 < mmcgrath> mizmo has been working on a good mockup for us.
16:15:58 < mmcgrath> f13: whats the lastest on the games spin? Is it approved, still waiting for
16:16:41 -!- fab_a [n=fabian_a(a)84-75-167-142.dclient.hispeed.ch] has joined #fedora-meeting
16:16:55 < mmcgrath> jeremy: ^^ ?
16:16:59 < EvilBob> When can more spins be added to the queue and what are the submission requirements.
16:17:00 < jcollie> looks like a cool plan... are the spins going to be bittorrent only?
16:17:02 -!- fabian_a [n=fabian_a(a)84-75-167-142.dclient.hispeed.ch] has joined #fedora-meeting
16:17:04 -!- fab_a [n=fabian_a(a)84-75-167-142.dclient.hispeed.ch] has quit Read error: 104 (Connection
reset by peer)
16:17:20 * mmcgrath gets link
16:17:56 < mmcgrath> EvilBob: this isn't completely set in stone yet but is what we're following right
now - http://fedoraproject.org/wiki/Infrastructure/CustomSpins
16:17:59 < jeremy> mmcgrath: I think largely we just need to figure out some of the bits around
spins.fedoraproject.org and it'll happen. I didn't see any no's, just "need to
finish fleshing out the details for hosting, etc"
16:18:19 * dgilmore is here
16:18:20 < EvilBob> mmcgrath: Thanks
16:18:27 < mmcgrath> jeremy: I didn't realize it had been approved yet. I'll have something ready by
the end of the day.
16:19:21 < mmcgrath> Anyone have anything else regarding the custom spins?
16:20:02 -!- BhanuPatial [n=bhanu(a)dialpool-210-214-122-2.maa.sify.net] has quit Read error: 110
(Connection timed out)
16:20:07 < mmcgrath> k, moving on to the schedules apge
16:20:13 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Schedule
16:20:59 < mmcgrath> http://fedoraproject.org/wiki/Infrastructure/Schedule
16:21:32 < mmcgrath> On the topic of corporate sponsorship I had another meeting with the people from
peer 1 on monday, nothing new. I'm pretty hopeful but they're working out some
technical and marketing details.
16:21:45 < jwb_gone> peer 1?
16:21:58 < mmcgrath> I'm really hoping they will be able to provide us with torrent and hosted.fp.o
16:22:01 < jwb_gone> nevermind, ignore me
16:22:17 < mmcgrath> :)
16:22:29 < spoleeba> jwb_gone, you were thinking pier 1
16:22:33 < jwb_gone> yes
16:22:49 < mmcgrath> Also we've gotten funding a server which I've put a PO request in to be delivered
16:22:54 < mmcgrath> I don't have shipping information yet though.
16:23:18 < mmcgrath> Anyone have any questions regarding the corporate sponsorship thats been going on?
16:23:53 < mmcgrath> k
16:23:58 < mmcgrath> nothing new on architectural documentation
16:24:37 < mmcgrath> And as far as SOP's go I've added one on our netapp setup -
16:24:51 * dgilmore is around now
16:25:01 < mmcgrath> as always you can get a list of all the SOP's we've got at the bottom of the
16:25:08 < mmcgrath> Thats all I've got so I'll open the floor
16:25:14 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Open Floor
16:25:25 < mmcgrath> Anyone have anything they'd like to discuss?
16:25:55 < abadger1999> wwoods has been doing some good work on a python-bugzilla module.
16:26:10 < mmcgrath> abadger1999: such as? :)
16:26:11 < abadger1999> I'm goingto add it to the fi repo until he gets it into Fedora
16:26:28 < abadger1999> and start using it with the packagedb to link to bugs on packages.
16:27:06 < abadger1999> http://wwoods.fedorapeople.org/python-bugzilla/
16:27:42 < mdomsch> abadger1999, skvidal's been hacking something together using that too
16:27:48 < mdomsch> packages.fp.o
16:27:56 < abadger1999> Yeah -- we've been talking about what needs to go where.
16:28:00 < mdomsch> cool
16:28:06 < mmcgrath> that will be handy.
16:28:08 < abadger1999> I'll be taking his bug page and incorporating it into the pkgdb.
16:28:23 < abadger1999> He'll get repoview running on packages.fp.o and link to the pkgdb bug page.
16:28:24 < mmcgrath> It seems like decent integration of a lot of our services is just on the horizon.
16:29:08 < mdomsch> mmcgrath, how are we on storage?
16:29:28 -!- tibbs [i=tibbs@fedora/tibbs] has quit "Konversation terminated!"
16:29:32 < mdomsch> or should I not ask ? :-)
16:29:55 < mmcgrath> mdomsch: we're about to hit 85% on the netapp which will put us in "warning" I've
been asking repeatedly for the koji cleanup to happen but it just hasn't yet.
16:30:20 < mmcgrath> The thing I'm worried about is if the cleanup doesn't actually clean much up. By
the time we get funding and a netapp installed it might be too late.
16:30:49 < mmcgrath> The other thing I've been realizing is we've got some pretty expensive storage for
RPMS that just sit there, I'm wondering if we shouldn't invest in a different
medium for the koji storage.
16:30:54 -!- rdieter [n=rdieter(a)sting.unl.edu] has quit Remote closed the connection
16:31:44 < mdomsch> amen
16:32:31 < mmcgrath> I think I'll start a thread about that on fedora-infrastructure. I've done some
research on price/G. netapps are freaking awesome but we pay out the nose for them.
16:32:40 < mmcgrath> I wonder if we just should get a powervault and load it up.
16:32:55 < mmcgrath> mdomsch: whats the largest powervault you guys have?
16:33:54 < mdomsch> MD3000i iSCSI attached
16:33:59 < jcollie> 353TB?
16:34:14 < jcollie> according to this:
16:34:31 < mmcgrath> <nod> I'll start a thread on the list.
16:34:53 < mdomsch> 15 400GB drives
16:35:23 < mdomsch> in one enclosure
16:35:32 < jcollie> http://www.dell.com/content/products/results.aspx/pvaul_md3000i?~ck=anav&...
16:35:34 < mdomsch> plus 2 more enclosures
16:35:38 < jcollie> says 18TB
16:35:40 < mdomsch> so, 45 400GB disks
16:36:00 < mmcgrath> thanks, I'll do some research.
16:36:19 < mmcgrath> Ok, anyone else have anything to discuss?
16:36:32 < mmcgrath> If not I'll end the meeting in 30.
16:36:53 < mmcgrath> 15
16:36:59 < mmcgrath> OH one thing.
16:37:14 < abadger1999> stop the presses! :-)
16:37:26 < mmcgrath> I wanted to thank mizmo, ricky and ivazquez for working on the website stuff.
They've been doing some good work.
16:37:29 < mmcgrath> ok.
16:37:31 < mmcgrath> 10
16:37:34 < mdomsch> hear hear
16:37:39 < mmcgrath> 5
16:37:47 -!- mmcgrath changed the topic of #fedora-meeting to: Infrastructure -- Meeting End
16:37:51 < mmcgrath> thanks for coming everyone
Apologies for dropping off the radar for the past few weeks, but things
at work got busier rather than slacking off after the semester started
(I got assigned several new high-priority projects - such is life when
working in an understaffed IT shop) and I vastly underestimated how busy
things would be in the evenings and weekends shuttling my kids back and
forth to various activities.
Anyway, things still are very busy with non-Fedora related stuff, but I
thought that I owed it to the community to post this and get the
discussion started again.
It was my impression that there was very little consensus on what (if
anything) should be done. There seemed to be three camps:
1) Stick with CVS as is.
2) Switch to another VCS, but the data in the repository would remain
similar to what we currently use (RPM spec plus patches).
3) Switch to another VCS, but the data in the repository would radically
change to an "expanded source" style repository.
All three camps have vocal proponents as well as vocal opponents. My
suspicion is that the "silent majority" is in the "stick with CVS" camp.
My personal choice would be to switch to Git for the VCS but keep the
repository data the same (spec file plus patches). I feel that
switching to expanded source-style repositories is too radical of a
change - we give up the notion of pristine source plus patches. Also,
using an expanded source-style repository would mean that packagers
would have to become much more familiar with the VCS since they would
need to maintain various branches (vendor branch, branches for various
In any case, it's unlikely that I'll have significant amounts of time to
work on a VCS conversion anyway so until someone else steps up to take
charge of this project we'll have to live with what we have.
P.S. I'll be in Chicago next week for a training class so if anyone in
the Chicagoland area would like to get together some evening for drinks
ping me on IRC or off-list. There's probably a couple of you out there
that I should probably buy a beer so this could be your chance to