[00:44] <fighter> so, i just ran find . -size +0 -exec mv -i {} {}xxx \; and now i have a bunch of files that are named genericporntitle.avixxx, can anyone tell me how to get rid of the xxx's i just added to my files?
[01:03] <qman__> fighter:  ls > list; while read file; do mv "$file" "`echo "$file" | sed s/avixxx/avi/`"; done < "list"; rm list
[01:04] <qman__> there's probably a better way but that will work
[01:05] <Jagged> sed s/xxx// would work as well
[01:05] <qman__> yeah but the string "xxx" is more likely to be in the title than "avixxx"
[01:06] <fighter> I'm willing to loset he xxx in the title
[01:06] <Jagged> well
[01:06] <Jagged> in that case:
[01:06] <qman__> then change it to sed s/xxx//g
[01:06] <Jagged> sed s/xxx//g
[01:09] <fighter> i typed it but so far i'm still waiting, i'm checking the dir on my windows box and nothing has changed yet. any idea how long it should take?
[01:09] <fighter> 226gb in 461 files
[01:10] <qman__> should be changing them as it goes
[01:10] <qman__> unless it's still waiting on the "ls > list"
[01:11] <Jagged> ps ax | grep ls
[01:11] <Jagged> and see if ls is still running
[01:11] <qman__> and really it shouldn't take extra time based on size, since it's really just changing the name
[01:11] <qman__> no need to actually move the data
[01:11] <qman__> the time should be based on the number of files
[01:12] <fighter> ah there we go
[01:12] <fighter> got it now
[01:12] <fighter> thank you for you help :)
[01:12] <qman__> no problem
[01:57] <smoser> erichammond, ping
[02:08] <smoser> soren, mdz erichammond feedback on release criticalness of bug 458576 and my attached patch would be nice.
[02:08] <smoser> i can definitely see an argument for it being fairly release critical.
[02:16] <smoser> good night all.
[02:16] <smoser> ttx, soren, zul, erichammond you have mail from me. i will see you all on monday AM.
[02:16] <zul> smoser: have a good day off
[02:26] <erichammond> smoser: I don't have any specific examples of who/what has automated security checks of ssh host keys, how they are implemented, or how difficult it would be for them to adjust regular expressions to work with a different console output format.
[02:27] <erichammond> I'm sure this type of checking exists, but don't have any real insight.  So, I'd probably go with importance "Low" until somebody complains.  If it's going to be easier to fix now than after release, and you want to avoid complaints, then bump it up to a higher priority.
[02:27] <twb> erichammond: are you talking specifically about checking for host keys generated during the Debian ssh-keygen entropy fiasco?
[02:28] <erichammond> twb: checking for ssh host key fingerprints in the EC2 instance console ouptput.
[02:28] <erichammond> twb: I think I know what you're talking about (a while back) and that's something different.
[02:28] <twb> OK, no worries.
[02:29] <twb> I was gonna say, that kind of host key scanning ought to be done by openssh-blacklist or whatever the package is.
[02:29] <twb> Dunno if it's any help to RTFS that.
[04:10] <twb> !guide
[04:10] <twb> !rute
[04:11] <twb> !sag
[04:11] <twb> Grr, I need to teach ubottu my mnemonics for the admin guide.
[04:12] <twb> !learn sag is https://help.ubuntu.com/9.04/serverguide/C/index.html
[04:14] <poningru> need help with sshfs and fstab
[04:14] <poningru> here is my situation
[04:15] <poningru> hmm hold on
[04:15] <twb> You can't automount sshfs filesystems unless you're using passphraseless key-based authentication.
[04:15] <poningru> I do
[04:15] <twb> OK, carry on.
[04:16] <twb> (But I highly recommend you learn about passphrases and ssh-agent for most other ssh keys.)
[04:16] <poningru> I have server with username torrent that allows all the client to login
[04:17] <poningru> so on one of my client computers logged in as media I do: sshfs -o allow_other torrent@consus2:/home/torrent/data/music music/
[04:18] <poningru> and that mounts it properly
[04:18] <poningru> but if I try to do it with fstab it does not work
[04:18] <poningru> here is what I have on fstab
[04:19] <poningru> sshfs#torrent@consus2:/home/torrent/data/music /home/media/music fuse comment=sshfs,auto,users,exec,allow_others,reconnect 0 0
[04:19] <twb> Doing A/V files over sshfs is gonna be pretty sucky.  If you're on a LAN, have you considered using NFS or CIFS?
[04:19] <poningru> ...
[04:19] <poningru> I hate both of them
[04:19] <poningru> sshfs ftw
[04:20] <poningru> hmm
[04:20] <poningru> you know I wonder...
[04:23] <qman__> what's with the comment=
[04:23] <qman__> something fuse specific?
[04:23] <twb> qman__: probably ends up in mtab or something
[04:24] <twb> poningru: what do you "hate" about NFS/CIFS?
[04:25] <poningru> twb, too many issues
[04:25] <twb> poningru: such as?
[04:26] <qman__> NFS is difficult to secure, but that's usually not a problem on a LAN
[04:26] <qman__> can't see any problems with CIFS
[04:26] <poningru> ok nfs: issues with reconnecting, crashing windows/osx etc.
[04:27] <qman__> I would never use NFS with windows
[04:27] <qman__> CIFS is ideal for windows
[04:27] <poningru> but not for linux
[04:27] <qman__> it works fine on linux
[04:27] <poningru> trying to setup samba on lin or osx is worse than pulling teeth
[04:28] <qman__> but there's no reason you can't use both
[04:28] <poningru> anyway
[04:28] <qman__> not at all
[04:28] <poningru> qman__, well my experience says otherwise
[04:28] <twb> Surely if your machine can use sshfs well, it can use NFS well.
[04:28] <poningru> twb, its not a matter of processor power
[04:28] <qman__> the default configuration is pretty well set
[04:28] <poningru> oh god
[04:28] <qman__> adjusting it is easy
[04:28] <twb> I realize that.  I was speaking about the quality of the NFS implementation.
[04:29] <poningru> ok if you guys wanna argue the merits of nfs/cifs vs. sshfs ...
[04:29] <poningru> I just want help with sshfs
[04:29] <poningru> twb, look at the lkml discussions on nfs
[04:29] <twb> sshfs has its uses, but I think it's a poor choice for distributing large A/V files over a LAN.
[04:29] <poningru> well I would like to disagree with you
[04:29] <qman__> I have to agree with twb, too much overhead
[04:29] <poningru> not at all
[04:30] <poningru> todays modern cpus?
[04:30] <qman__> it's not just CPU power
[04:30] <qman__> besides, why waste it on the network mount
[04:30] <poningru> my 'old' pentium dual core with 1.46 I have maybe 1% overhead with sshfs
[04:30] <twb> It'd be like riding a bicycle on an interstate highway
[04:30]  * poningru rolls eyes
[04:30] <poningru> oh god
[04:30] <poningru> sorry I bothered you guys
[04:31] <twb> poningru: I don't think you actually describe the problem you're having with sshfs yet.
[04:31] <qman__> just saying, I wouldn't ever bother setting up sshfs that way
[04:31] <qman__> as such my advice is to not use sshfs for that purpose
[04:31] <poningru> twb, uh...
[04:31] <poningru> qman__, point taken
[04:32] <poningru> twb, <poningru> I have server with username torrent that allows all the client to login
 so on one of my client computers logged in as media I do: sshfs -o allow_other torrent@consus2:/home/torrent/data/music music/
 and that mounts it properly
 but if I try to do it with fstab it does not work
 here is what I have on fstab
 sshfs#torrent@consus2:/home/torrent/data/music /home/media/music fuse comment=sshfs,auto,users,exec,allow_others,reconnect 0 0
[04:32] <twb> poningru: ah, sorry, I missed the "not"
[04:32] <qman__> having never put an sshfs mount in the fstab, my first guess is that #
[04:32] <twb> poningru: the mount command is run as root, not as media.
[04:32] <qman__> second guess is the comment= on the options
[04:33] <twb> poningru: therefore, *root*'s .ssh/config must be set up correctly.
[04:33] <qman__> but I could easily be wrong about it
[04:33] <twb> qman__: the # means it's a fuse entry
[04:33] <qman__> ah
[04:33] <poningru> twb, right but thats what the uid and gid is for
[04:33] <twb> poningru: what uid and gid?
[04:33] <poningru> oh sorry that is old one
[04:34] <qman__> poningru, he's right about that
[04:34] <qman__> root's keys need to be set up too
[04:34] <twb> qman__: well, just tell /root/.ssh/config to use ~media/.ssh/id_%h or whatever.
[04:34] <poningru> well add on uid=1001,gid=1001
[04:35] <poningru> right but I remember setting this up a while ago
[04:35] <poningru> and putting it into root is the easy way out
[04:35] <poningru> I have it working like that
[04:36] <twb> poningru: that doesn't change who the operation runs as.
[04:36] <poningru> twb, oh hmm
[04:36] <qman__> mounting a systemwide file share as root isn't exactly risky, considering that all your other filesystems are mounted as root
[04:36] <poningru> how does one do that then?
[04:36] <poningru> qman__, I know
[04:36] <twb> poningru: you can't.
[04:36] <twb> poningru: mount -a always runs as root.
[04:37] <twb> poningru: if you want to mount a filesystem as some other user, you need to put -o noauto in its fstab entry, then add an init script or an entry in /etc/rc.local.
[04:37] <twb> As qman__ says, it's pretty silly to try to restrict that, generally.  What are you guarding against?
[04:38] <poningru> I'm not
[04:38] <poningru> I'm trying to figure it out
[04:38] <poningru> curiosity
[04:40] <poningru> twb, there has to be a way for a user to not have to share his private key with the root
[04:40] <poningru> I mean technically the root can just view it...
[04:40] <poningru> but still there has to be a way
[04:41] <qman__> you generate and trust another key
[04:41] <qman__> but using the key you already have set up is easier
[04:41] <KurtKraut> Is there a Unified Monitoring System that doesn't rely on SNMP to gather data? I find SNMP so hard to configure.
[04:42] <poningru> KurtKraut, what are you trying to configure?
[04:42] <poningru> err monitor*
[04:43] <KurtKraut> poningru, the essencial is: if the host is up or down (by ping) and what services are reachable by the central node (httpd, irc).
[04:43] <poningru> eh go with something simple like bigbrother
[04:43] <qman__> KurtKraut, you could configure a script on the monitoring host with nmap
[04:43] <twb> 14:40 <poningru> twb, there has to be a way for a user to not have to share his private key with the root
[04:43] <twb> poningru: if you can't trust root, you are fucked.
[04:44] <poningru> twb, read what I said immediately after that
[04:44] <KurtKraut> qman__, that is possible indeed. By I think this is so essential for small scenarios that I'm impressed no one has already did it.
[04:46] <qman__> KurtKraut, I've been too lazy to set one up myself
[04:46] <qman__> nagios does this, but it may be a little bigger scope than what you're looking for
[04:47] <poningru> KurtKraut, bigbrother
[04:47] <KurtKraut> poningru, is it a package name?
[04:48] <poningru> http://www.bb4.org/
[04:48] <poningru> no clue
[04:48] <qman__> flash? really?
[04:48] <qman__> I wouldn't use it on that basis alone
[04:49] <poningru> qman__, yeah you can turn that shit off
[04:49] <poningru> http://www.zabbix.com/
[04:49] <poningru> btw another cool/simple one
[04:50] <crohakon> what is that?
[04:51] <KurtKraut> Zabbix rely on SNMP. This single fact makes me think this cannot be pointed out as simple.
[04:51] <poningru> KurtKraut, you dont have to rely on snmp iirc with that
[04:51] <poningru> I remember setting it up to do just ping/nmap tests
[04:52] <poningru> crohakon, network monitoring tool
[04:52] <KurtKraut> poningru, so I'll give it a try
[04:52] <crohakon> for datacenters or something?
[04:52] <qman__> still seems like overkill
[04:52] <poningru> crohakon, eh sure
[04:52] <poningru> crohakon, you can use it for your home network too
[04:52] <qman__> nagios is pretty much exactly like those
[04:53] <crohakon> hmmm something to play with.. =)
[04:53] <qman__> and is in the ubuntu packages
[04:53] <poningru> qman__, absolutely nagios is a good one too
[04:53] <poningru> imho a little to much for us sys admins though
[04:53] <poningru> not point and click install
[04:53] <qman__> dont' know if it requires snmp to work or not
[04:53] <qman__> I know it can use it
[04:53] <qman__> cacti is snmp based
[04:54] <poningru> right
[04:55] <qman__> still overkill for my taste, which is most of why I don't have one set up
[04:56] <qman__> eventually I'll write something up in php that parses a simple nmap output
[04:58] <crohakon> no ubuntu package for zabbix?
[04:58] <poningru> crohakon, it should be there iirc
[04:58] <poningru> qman__, bb is for you then
[04:59] <crohakon> iirc?
[04:59] <poningru> if I recall correctly
[04:59] <qman__> I see 6 zabbix packages in my apt-cache
[04:59] <poningru> yeah same here
[04:59] <qman__> so yes, though it may be universe/multiverse/canonical
[05:00] <crohakon> So can I install this on my server, and view it from my laptop?
[05:00] <qman__> looks like you install zabbix-server-??sql and zabbix-php-frontend on the server
[05:00] <sub> Zabbix has a web interface, yeah
[05:00] <qman__> and zabbix-agent on the systems you want to monitor
[05:00] <qman__> then you just browse to the web site
[05:01] <crohakon> wow, lots of little pieces.
[05:01] <poningru> sub, yo
[05:01] <poningru> qman__, yep
[05:03] <poningru> crohakon, if you dont like snmp you may wanna try out spong
[05:03] <poningru> though I have never used it
[05:04] <crohakon> <--- just getting into this stuff. Not really sure what the benefits or disadvantages are of snmp.
[05:04] <qman__> I'm not a fan of snmp
[05:04] <qman__> it's difficult to configure, and a cleartext protocol to boot
[05:04] <qman__> at least, the version everything supports is
[05:04] <poningru> crohakon, eh its a little complicated unneccesarily and hard to configure
[05:05] <qman__> supposedly newer versions can be encrypted but it's just a mess
[05:05] <poningru> qman__, whats wrong with cleartext?
[05:05] <qman__> nmap is a better option
[05:05] <poningru> just put it over ssh
[05:05] <poningru> qman__, nmap isnt super reliable though
[05:06] <sub> hey poningru
[05:06] <poningru> didnt know you hung out here
[05:06] <sub> yep
[05:06] <sub> here and #ubuntu-virt
[05:07] <poningru> oh cool
[05:07] <sub> i do some light contributing to Ubuntu VMBuilder and triag bugs in server when I have the spare time
[05:07] <twb> I heartily support protocols that use cleartext, and a separate encryption layer.
[05:07] <twb> It makes debugging far less painful.
[05:07] <poningru> twb++
[05:07] <poningru> sub, I've been meaning to get into vmbuilder
[05:08] <poningru> I wanna get into getting debian/rhel/suse build be as easy as ubuntu
[05:08] <sub> it's a cool little piece of code, although about to go through a heavy refactoring or rewrite
[05:08] <crohakon> I want a static IP so I can host my own websites. But thats not going to happen. =)
[05:09] <poningru> sub, really?
[05:09] <sub> yeah
[05:09] <poningru> didnt soren here write that code?
[05:09] <poningru> sub, link?
[05:09] <sub> yes, and he's the one who wants to rewrite it, hehe
[05:09] <sub> let me see if i can find the email
[05:10] <qman__> crohakon, dynamic DNS works pretty well for me, though my ISP filters ports 21, 25, 53, 80, and 445 for security reasons
[05:10] <sub> oh it's lind of long, i'll see if i can find an archived one i can link you to
[05:10] <sub> poningru: https://lists.launchpad.net/vmbuilder/msg00003.html
[05:11] <crohakon> yeah, charter blocks those ports as well. I sent an email asking what I need to do to get a static IP and those ports unblocked. They want me to pay $99/m for a 10mb down/1mb up connection with a static IP.... not going to happen.
[05:12] <qman__> hah, I have charter too
[05:12] <qman__> total rip
[05:12] <qman__> unfortunately there's no competition, no other options
[05:13] <crohakon> 3 years ago I had a business account with verizon, and although DSL is much slower, I had a static IP for like 34/m and no ports blocked. Where I live now, however, charter is my only option.
[05:13] <poningru> crohakon, that sucks
[05:14] <qman__> my 5M/512k service is $50 a month on top of the "bundle" discount for getting TV with them too
[05:15] <qman__> the only other option is a T1 at $400 a month
[05:16] <crohakon> Currently with charter I have a 8down/1up connection with dynamic IP and ports blocked for 34.99/m... to get the same up speed and a static ip with no port blockage they want $65 extra a month.
[05:16] <poningru> jeezus
[05:16]  * crohakon his hand on head
[05:16] <poningru> crohakon, work for cablevision
[05:16] <poningru> with no vid
[05:16] <crohakon> cablevision?
[05:17] <poningru> 30down and 5up is about 65USD thats with a free domain, hosting, all ports open, static ip etc.
[05:17] <poningru> a cable company that services ny,nj,parts of ct etc.
[05:17] <crohakon> damn... I need to move.
[05:18] <twb> http://www.internode.on.net/residential/broadband/adsl/nakedextreme/pricing/ (annex M)
[05:18] <crohakon> Sadly, I live in the rust belt... Michigan.
[05:18] <Sorrell> is anyonw having static ip issues with 9.10 ?
[05:18] <crohakon> No one has work, and everything is expensive.
[05:18] <poningru> Sorrell, my vm isnt
[05:18] <poningru> server, install from iso
[05:18] <qman__> monthly download quota? that wouldn't work
[05:19] <Sorrell> :( dang, just me then.
[05:19] <qman__> I exceed 150GB every month
[05:19] <Sorrell> bandwidth cap = suck
[05:19] <twb> And in .au, all domestic lines are rate limited (or subject to hefty excess fees) after exceeding a monthly quota.  I get the impression that's uncommon in other nation-states.
[05:19] <poningru> yeah its sucks
[05:19] <poningru> qman__, comcast has a technical limit of 250gig
[05:19] <poningru> per month
[05:20] <poningru> I do exceed that time to time
[05:20] <poningru> but...
[05:20] <poningru> they dont care
[05:20] <twb> Theoretically our wonderful new center-left government will put in FTTH (FTTP) in the next ten years.
[05:20] <qman__> I used 450GB in august
[05:20] <poningru> qman__, hehe
[05:20] <qman__> though that's not typical
[05:21] <twb> http://en.wikipedia.org/wiki/National_Broadband_Network
[05:21] <poningru> qman__, downloading ubuntu isos ;)
[05:21] <qman__> let's go with that
[05:21] <twb> "Download speeds of 100 mbps for 90% of Australia homes and businesses"
[05:21] <poningru> cool
[05:21] <poningru> http://www.phoronix.com/scan.php?page=news_item&px=NzYzMA
[05:21] <qman__> wow
[05:21] <twb> Of course, that won't apply to anything hosted OUTSIDE .au
[05:21] <twb> And that's assuming they actually manage to build it, on time and on budget (har har har)
[05:22] <qman__> yeah
[05:22] <qman__> AT&T promised the US 100 meg lines to every home over a decade ago
[05:22] <twb> The government said "the telco's bids don't offer value for money, so we're going to do it ourselves, for cheaper" (har har har)
[05:23] <qman__> in return for tax breaks
[05:23] <poningru> qman__, THANK YOU
[05:23] <qman__> not only have they not delivered on that, their service hasn't improved in years
[05:23] <poningru> bunch of dirty bastards
[05:23] <poningru> qman__, and now everytime net neutrality comes around
[05:24] <poningru> they prop up the argument that 'we own the network we can do what we like with it'
[05:24] <poningru> no you dont you shit heads
[05:24] <poningru> tax payers essentially paid for it with tax cuts
[05:25] <twb> Have you ever gotten sick of hearing AT&T take credit for things that they didn't invent? You will.  -- Dave Hamilton
[05:25] <crohakon> I wish I had the resources to get a dedicated line just for fun.
[05:25] <qman__> I've seriously considered getting a T1 despite the ridiculous cost
[05:25] <qman__> charter has terrible service
[05:26] <twb> qman__: you could just get three separate DSL connections from separate providers, then multiplex them
[05:26] <qman__> can't get DSL
[05:26] <qman__> if I could, I would
[05:26] <crohakon> In this area it is hard to get DSL from one provider... let alone three!
[05:27] <twb> So you're saying that in .us, you can get fibre in places you can't get DSL?
[05:27] <twb> WTF, are your exchanges so far apart, or just lacking DSLAMs?
[05:27] <qman__> you can get a T1 anywhere if you're willing to pay
[05:27] <qman__> but you can't get FIOS in most places
[05:27] <twb> T1's a wire protocol.  What does it run on?
[05:27] <twb> Existing copper lines?
[05:28] <qman__> new copper, I'm pretty sure
[05:28] <qman__> it's a dedicated line
[05:28] <twb> I feel sorry for the kind of people who would lay new copper in this day and age
[05:29] <crohakon> Americans like to live far apart from each other... which makes DSL difficult.
[05:29] <qman__> oh, the US is far from that
[05:29] <qman__> copper's not going anywhere
[05:29] <twb> crohakon: you'll get that in ANYTHING that uses copper as a transport.
[05:30] <qman__> my choices are 21k dialup, 512k satellite, charter cable, or a leased line (T1 or better)
[05:30] <twb> crohakon: it's just that DSL uses only the high frequencies (to leave room for voice), so it attenuates quickly.
[05:30] <qman__> I have no other possible service
[05:30] <crohakon> Maybe I should get a lease line and try to get my neighbors to use me as an ISP. =)
[05:30] <qman__> AT&T is too cheap to fix the phone lines
[05:31] <qman__> there's no way they'd lay fiber here
[05:31] <twb> qman__: not so much cheap as lacking competition (and antimonopoloy laws), I suspect
[05:32] <twb> *monopoly
[05:32] <qman__> yeah, they're the only phone provider in my specific area
[05:32] <qman__> same goes for charter cable
[05:32] <qman__> only cable provider
[05:32] <crohakon> So, how much more bandwidth do you get with a basic T1 connection?
[05:33] <qman__> less download, more upload, 1.544mbps full duplex
[05:33] <crohakon> as compared to a basic cable connection?
[05:33] <qman__> but, it's got a 99.999% uptime guarantee
[05:33] <twb> My arse it does
[05:33] <qman__> as opposed to charter, which in my area is about 90%
[05:34] <twb> That's like two minutes downtime per year.
[05:34] <qman__> that's what they advertise
[05:34] <twb> advertising isn't the same as the contract
[05:34] <crohakon> My only problem with charter in my area is that it starts to slow down after midnight (when I normally am trying to watch something on hulu or netflix
[05:34] <twb> Maybe they're only giving five nines that e.g. the cable won't explode, and not that e.g. the repeaters won't shit themselves.
[05:35] <qman__> probably
[05:35] <qman__> but they will fix the problem
[05:35] <qman__> unlike charter, who waits around for a few hours first
[05:36] <crohakon> So... if I wanted to host a few websites... would a T1 be any better then a 10down/1up cable connection?
[05:36] <qman__> company policy, they won't fix the problem unless at least five people make it to tier 3 and complain
[05:36] <qman__> even if they know for a fact that it's down
[05:36] <poningru> heh wow
[05:37] <twb> Here, if you report a problem to the ISP, and it's a problem with the physical line, they wait until the end of the day to queue up all the fault requests and then send them as a batch to telstra (who own the copper itself).
[05:37] <twb> And then telstra take days to get back to the ISP
[05:37] <qman__> charter's DNS is also totally worthless
[05:37] <qman__> I've had to run my own for years
[05:37] <Sorell> is there a special way to set the static ip in ubuntu server?
[05:37] <Sorell> 9.10
[05:38] <twb> Sorell: /etc/network/interfaces (see "man interfaces")
[05:38] <qman__> Sorell, /etc/network/interfaces
[05:38] <Sorell> :/ that's what I have been doing
[05:38] <twb> Sorell: purge NetworkManager
[05:38] <Sorell> go in and change auto to static.
[05:39] <qman__> you don't change auto
[05:39] <twb> There are copious examples in the manpage
[05:39] <Sorell> maybe it's dynamic then
[05:40] <Sorell> I was wondering exactly for 9.10
[05:40] <Sorell> I had this working in 8
[05:40] <qman__> it's the same as for every previous version I'm aware of
[05:40] <Sorell> .
[05:41] <qman__> huh
[05:42] <qman__> error upgrading, corrupted archive
[05:42] <qman__> wouldn't mind except that's the kernel
[05:43] <crohakon> omg... I so desire an OC-48...
[05:44] <Sorell> just sign here in blood
[05:44] <crohakon> lol
[05:44] <Sorell> :)
[05:45] <twb> I don't know why, really.
[05:45] <twb> Ultimately the bandwidth all goes to bittorrent, which never has anything good on it
[05:45] <crohakon> =)
[05:46] <twb> Even if you wanted to watch something by Joss Wedhon, it's probably cheaper in .au to walk to the store and buy a boxed DVD set than to rent an ADSL2+ line.
[05:47] <crohakon> Anyone know a good co-location host?
[05:48] <twb> What are your metrics for "good"?
[05:48] <twb> Fast, cheap, unaccountable, ... ?
[05:48] <crohakon> yes.
[05:48] <crohakon> lol
[05:48] <crohakon> fast and cheap mostly
[05:49] <crohakon> cheap... mostly.
[05:49] <crohakon> lol
[05:49] <qman__> I'm starting to regret the decision to upgrade my desktop to the RC
[05:50] <qman__> corrupt packages, dbus errors abound
[05:51] <twb> Fortunately, you're gonna report those bugs so they're fixed for everyone else
[05:53] <crohakon> Okay, so the real reason I have been playing around with ubuntu server is that my employer wants to host his website(s) on sight. What kind of connection speed do you need to host 3-4 websites that get anywhere between 100-200 hits a day?
[05:54] <qman__> depends entirely on the content
[05:54] <crohakon> No video
[05:55] <qman__> a few basic pages, half a meg ought to do fine
[05:55] <qman__> probably 384k
[05:55] <crohakon> mostly text and images. database heavy, though.
[05:56] <crohakon> Yeah, I did not think it would take much.
[05:56] <qman__> 100-200 hits per day suggests that you wouldn't have more than a dozen people loading pages at once
[05:57] <crohakon> correct
[05:57] <qman__> assuming your page downloads are small, a half meg should be able to keep up
[05:58] <crohakon> Now, this connection would also handle are merchant services, quickbooks services, etc..
[05:58] <crohakon> would that have any effect? I don't see where it would.
[05:59] <qman__> well, if you have more than a couple users I'd suggest you throw in another 256k
[05:59] <qman__> nothing serious
[06:00] <qman__> "Could not install the upgrades
[06:00] <qman__> The upgrade is now aborted. Your system could be in an unusable state. A recovery will run now (dpkg --configure -a)."
[06:00] <qman__> bleeding edge ftw
[06:03] <crohakon> and having the mysql server and the apache server on the same machine should not be a problem?
[06:04] <qman__> not for that small of a load
[06:04] <qman__> you can always create another mysql server and move the databases if it becomes a problem, too
[06:06] <crohakon> true
[06:07] <qman__> ok, I think I got all the broken packages sorted
[06:07] <qman__> time to cross fingers and reboot
[06:07] <qman__> bbl
[06:08] <crohakon> good luck
[06:14] <qman__> well, it booted up
[06:14] <qman__> but firefox overrode all my settings with defaults, lost my bookmarks and everything
[06:14] <crohakon> at least it booted...
[06:18] <qman__> hmm
[06:19] <qman__> I can get my 3.0 settings back, but I was using 3.5 before
[06:20] <twb> One more reason to drop firefox
[06:20] <crohakon> And what do you suggest instead of firefox?
[06:20] <twb> Well, I use a combination of emacs-w3m, w3m, html2ps, midori and opera.
[06:20] <twb> But obviously that is not for everyone
[06:21] <qman__> I was trying to get uzbl to work, but no dice
[06:21] <twb> Webkit is a featureful engine, but the downside is that it's a featureful engine
[06:21] <qman__> well, I only switched to 3.5 from 3.0 a few weeks agos
[06:21] <twb> Just like xulrunner, it takes like 3 hours to compile on typical hardware, and wants to use gobs or resources at runtime
[06:21] <qman__> so not much lost
[06:22] <crohakon> Anyone have desire for chrome to be released for linux?
[06:23] <twb> Nope.
[06:25] <qman__> not really
[06:25] <crohakon> I have been using firefox for so long the thought of using something else has not even crossed my mind...
[06:25] <qman__> I was really unhappy with firefox 3
[06:26] <qman__> but 3.5 is much better
[06:26] <qman__> though still not ideal, it works well enough
[06:27] <crohakon> firefox has never not met my needs.
[06:28] <qman__> 3 was really, really slow
[06:28] <qman__> I had whole second UI delays
[06:28] <twb> qman__: you think that's bad?
[06:28] <qman__> and when flash would die, as it tended to do
[06:28] <twb> qman__: try ssh -Xf fs firefox
[06:28] <qman__> I basically had to kill it, wait five minutes, open it up, and wait five more minutes for it to reload my tabs
[06:28] <twb> openoffice is even more hilarious to run remotely
[06:28] <crohakon> I guess I just did not notice... most likely at the time I blamed it on my internet connection.
[06:28] <qman__> hah
[06:29] <twb> Both of them basically abuse the fuck out of pixmap objects to fake their own "themed" widgets -- even more so than GTK2 does
[06:30] <jussi01> !language | twb
[06:31] <qman__> it appears afraid.org's DNS servers are down
[06:32] <qman__> first time that's happened since I started using them...probably 5 years ago, now
[06:32] <twb> Linguistic apartheid!
[06:33] <poningru> NOOOOOOOOO
[06:34] <poningru> what do you get from poningru.close.com?
[06:34] <poningru> can you ping it?
[06:34] <twb> downforeveryoneorjustme.com
[06:35] <poningru> twb, I dont have any web sites here
[06:35] <poningru> just mail
[06:35] <poningru> can you ping poningru.close.com?
[06:36] <qman__> oh, looks like the domain my subdomain is on got pulled for some reason
[06:36] <qman__> guess I just have to wait it out
[06:36] <poningru> oh ok cool
[06:37] <qman__> nice site
[06:38] <qman__> man, if this domain doesn't come back, I have a lot of work to do
[06:38] <qman__> I have certificates to replace and everything
[06:38] <poningru> oh shit
[06:40] <qman__> might even have to trash my entire VPN setup
[06:40] <poningru> man that sucks
[06:42] <jussi01> poningru: Ive just asked twb to watch the language, if you could do same that would be great :)
[06:42] <poningru> arr?
[06:42] <poningru> sucks?
[06:43] <poningru> oh shit? sorry
[06:49] <altf2o> just wondering if anyone's ever used gitosis + gitweb. I have it running fine on my test virtual server, however i have everything pointed to /home/git/repositories/ , where all my repos including 'gitosis-admin.git' reside. The problem is when viewing: http://virtual_server/gitweb/ , it does show my test repos but also 'gitosis-admin.git' which is fully browsable and public keys downloadable. Surely this isn't desired, any ideas?
[06:50] <altf2o> One link said to add:  [repo REPONAME] gitweb = no  , tried that, restarted apache2, no dice.
[07:08] <nijaba> good morning
[07:11] <crohakon> good night
[07:16]  * altf2o incase anyone else wonders: In /etc/gitweb.conf , setting: $project_list = "/home/git/gitosis/projects.list";  $export_ok = ""; $strict_export = "true";  , will only allow those repos listed in /projects.list to show up. In your gitosis.conf file: [repo gitosis-admin] gitweb = no  , DOES then control if it's visible or not. Setting it to: gitweb = yes  , will make it world viewable again.
[07:49] <maxagaz> hi
[07:49] <maxagaz> I have a chinese video capture card for which lspci returns "02:00.0 Multimedia video controller: Conexant Systems, Inc. Device 8200", what module should I load to make it work ?
[07:49] <maxagaz> Is there a module for this card ?
[07:51] <twb> maxagaz: google that string
[07:51] <twb> The part after the colon
[07:51] <maxagaz> twb, I did it...
[07:52] <maxagaz> twb, but I go nothing
[07:52] <maxagaz> got
[07:55] <twb> Well, I got lots of hits
[07:55] <twb> I didn't look to see how useful they were
[08:00] <Boohbah> twb: No results found for "Conexant Systems, Inc. Device 8200".
[08:00] <Boohbah> Results for Conexant Systems, Inc. Device 8200 (without quotes):
[08:00] <Boohbah> twb: useless
[08:00] <twb> http://www.google.com/search?q="Conexant+Systems,+Inc.+Device+8200"8
[08:00] <twb> The first three hits are ubuntu ones, for me.
[08:00] <Boohbah> Your search - "Conexant Systems, Inc. Device 8200"8 - did not match any documents.
[08:01] <twb> Oops, no 8
[08:01] <twb> Bad copy-and-paste job.
[08:01] <Boohbah> ooh, second link looks possibly good. video controller: Conexant Systems, Inc. Device 888
[08:01] <twb> Boohbah: are you in China?
[08:01] <Boohbah> Device 8880
[08:01] <Boohbah> twb: nope
[08:02] <twb> Maybe you're logged into google?
[08:02] <twb> I don't know why else google would be filtering your search results
[08:02] <Boohbah> am i getting chinese results? i was googling for chinese pages yesterday
[08:02] <Boohbah> but i am in the US
[08:02] <twb> Boohbah: Google filters its results within China
[08:02] <maxagaz> i don't think they would filter results for such things
[08:03] <twb> Shrug.
[08:03] <maxagaz> anywway, i'm wondering if i can make this card work on ubuntu
[08:03] <maxagaz> how would you try to make it work ?
[08:04] <maxagaz> at least, dmesg returns no error at the boot
[08:04] <twb> I would generally make it work by waiting for the revolution to put the MPAA against a wall and shoot it, since HDMI has more DRM than DVDs
[08:05] <maxagaz> twb, do you mean that there are no way to make it work easily ?
[08:05] <qman__> maxagaz, not likely
[08:06] <qman__> if it doesn't work out of the box, and google is no help, there probably isn't a driver for it
[08:06] <twb> maxagaz: the whole point of DRM is to make it impossible for FOSS projects unable to use hardware that you theoretically "own".
[08:06] <qman__> did you try to see if it works?
[08:06] <twb> qman__: good question!
[08:07] <maxagaz> qman__, no, I didn't
[08:08] <qman__> my tuner is pretty much dead, but I used to use tvtime with it
[08:54] <kaushal> hi
[08:54] <kaushal> is there a way to know in MySQL DB Server which query consumes more time ?
[08:59] <_ruben> there's a slow query log config option
[11:30] <kaushal> pwd
[11:48] <maxagaz> how can can i check the number of kB uploaded and downloaded on a machine ? Is there a command to check the activity of my NIC ?
[11:49] <soren> maxagaz: ifconfig
[11:50] <maxagaz> soren, ifconfig ? how to check it with ifconfig ?
[11:50] <maxagaz> soren, i'd need a tool like htop
[11:50] <soren> maxagaz: By looking at it?
[11:50] <soren> maxagaz: It's right there in the standard ifconfig output.
[11:51] <soren>           RX bytes:3908829145 (3.9 GB)  TX bytes:69635931 (69.6 MB)
[11:53] <maxagaz> what RX and TX stand for ?
[11:54] <soren> One is receive, the other transmit.
[11:54] <maxagaz> and X ?
[11:55] <soren> Nothing, really.
[11:55] <soren> It's frm the old telegraph days.
[11:56] <soren> They abbreviated transmit-out as TX, and receive-in as RX.
[11:56] <soren> It's /very/ commonly used.
[11:58] <maxagaz> interesting
[11:58] <maxagaz> soren, ifconfig isn't very convenient to monitor the nic activity
[11:59] <soren> You never said anything about monitoring it.
[11:59] <soren> If you want to keep track of it, look at munin.
[12:20] <maxagaz> soren, yes munin...
[12:20] <maxagaz> soren, I just tried ntop, it looks good
[12:22] <soren> maxagaz: Well, if that's what you need, just stick with that.
[12:26] <maxagaz> what i'm wondering now is, can i monitor the data flow coming the NIC of a machine and leaving the server from once of its NICs ?
[12:30] <soren> What do you mean?
[12:46] <alvin> I have a problem with a fresh installation of Karmic RC1. (On same server, nog problem with Jaunty):
[12:46] <alvin> Sometimes, the server boots fine, but most of the time the boot process halts when it can not mount the nfs /home because the network isn't there yet.
[12:46] <alvin> That's not a big problem, but the network service doesn't even try to connect after that. Is this a known problem? Shouldn't it retry connecting?
[12:52] <alvin> I like the fact that you try to boot ubuntu faster, but actually, I don't find it very important on a server. Starting services one by one is ok with me. even waiting 5 minutes until the network comes online. But this is a showstopper.
[13:01] <alvin> ok, another question: is there (experimental) boot logging yet?
[13:06] <pmatulis> alvin: what error messages do you see?
[13:44] <kRocKodile> hello
[13:44] <kRocKodile> how many years of update packages has the ubuntu server?
[13:45] <kRocKodile> (my english sucks)
[13:54] <soren> 8.04: 5 years. 8.10, 9.04, and 9.10: 1½ years. 10.04: 5 years.
[13:54] <soren> kRocKodile: ^
[13:58] <aubre> when I try to attach a volume to an instance using SC I get bad input params to vnetAttachTunnels() followed by failed to attach tunnels for vlan 10 during maintainNetworkState() followed by network state maintainance failed - what am I doing wrong?
[13:58] <aubre> that's in the cc.log btw
[13:59] <aubre> reinstalling didn't fix my problem
[13:59] <aubre> euca-attach-volume -i i-4F8E099C -d sdb vol-329904A4 VOLUME	vol-329904A4
[14:02] <aubre> I can get everything else to work, and I am anxious to show the system to the director, but this is a show-stopper for me
[14:19] <alvin> I'm going to reinstall Jaunty, because the network service starts before idmap/nfs. Is there anyone who will stop me and tell I can use Karmic, but have to change the configuration a little?
[14:20] <soren> alvin: If there's problems with NFS in Karmic, we need to fix them. Can you please file a bug, so we can look at it?
[14:21] <Bilge> Are services meant to be started and stopped by calling init.d scripts directly or is there a utility that's intended to be used instead?
[14:21] <alvin> soren: I don't know if NFS is the problem. If I don't mount any drives, the network service will not always start either. The NFS just halts the whole boot process if there is no network available at that time. Sometimes (rare), the server does boot.
[14:23] <alvin> soren: Can I perform some tests? Any suggestions?
[14:25] <alvin> Hmm, now the console says: 'Gave up waiting for root device'. That happens from time to time too, but only in Karmic.
[14:25] <alvin> The question is: is bringing up the network and the rootdelay related?
[14:27] <alvin> soren: I'd like to file a bug, but I don't know against what. Sometimes, root is not found, and sometimes the NFS server (shared home) is not found because the network isn't there.
[14:37] <pmatulis> alvin: take nfs out of the picture - use a regular home and test
[14:37] <alvin> after a lot of 'giving up waiting for root device' and reboots: If the NFS server is not found, you can escape to a rescue shell. If you wait a bit, the network will come up and the server can be seen. Then you can mount by hand and continue the boot process. If you just continue, mountall stops responding and nothing can be done.
[14:38] <alvin> I will try with a regular home. (please stand by)
[14:39] <alvin> done, rebooting. Another thing: when booting, mountall says that the swap is busy every time the home can not be mounted. Swap is just a local volume.
[14:55] <soren> alvin: File it against mountall. It may not be correct, but it's better to have it filed and then we can move it afterwards.
[15:07] <heath|work> apt-get seems to be struck. I ctl+c'd it and now I can't run it. How can I get it back?
[15:15] <pmatulis> heath|work: use the ps command to ensure it's no longer running
[15:15] <pmatulis> heath|work: kill it if it is
[15:15] <heath|work> pmatulis: thanks I got it
[15:22] <alvin> soren: I will. It's certainly network related. If I omit all NFS mounts and just boots and I can log in. I can then start a ping to the fileserver. First, there is no network, then the ping is unanswered, and after a while, the ping will start responding. So, it's really the network. (side note: it is a static network)
[15:26] <alvin> Hmm, my problem looks related to bug 431248, but that is reported to be fixed and I have all latest updates.
[15:31] <nijaba> smoser: hello.  is the current karmic aki/ari to use on ec2 published somewhere?
[15:43] <zul> nijaba: he is away today
[15:44] <nijaba> hey zul
[15:44] <zul> hi nijaba
[15:44] <nijaba> zul: would you have any clue where I could find that?
[15:44] <zul> yeah i think giime a sec
[15:46] <alvin> ok, reported as bug 459134. Thanks for the help. Work on this will continue on monday.  I have permission to test this a little while further before going back to Jaunty for production use.
[15:47] <nijaba> zul:duh  I look in this exact directory, did not think of looking at this file
[15:49] <zul> nijaba: heh
[15:57] <aubre> I figured out my SC problem, my SC had registered itself (during the automatic UEC install) to use my public IP and interface, when I switched to a private IP and private interface, and subsequently created a new volume, everything works. Then the question becomes why did the UEC disk install register itself with my public interface instead of my private interface?
[15:57] <aubre> could it be because it doesn't know about the private interface until the end? I changed my SC config in the web interface btw
[16:04] <zul> ttx: ping
[16:04] <ttx> zul: pong
[16:05] <zul> ttx: do you want me to upload the fix for 458576 if its not done already?
[16:05] <ttx> bug 458576
[16:06] <ttx> zul: we have another required fix in that area
[16:06] <zul> ttx: k
[16:06] <ttx> that may or may not also be in ec2-init
[16:06] <ttx> I'd better make a single change with both...
[16:07] <zul> ttx: agreed
[16:08] <zul> ttx: let me know when you need my expert services
[16:24] <eaman> Hello, anyone is aware of some problems with pam_mkhomedir.so (LDAP) script shipped in Karmik? Here it fails the UID of the user while creating the /home/dir (on NFS) at the first login (it goes to NOBODY).
[16:38] <ahasenack> eaman: pam_mkhomedir runs as root when creating the dir. On NFS, root on the client gets translated to nobody on the server by default
[16:38] <ahasenack> eaman: unless you have no_root_squash turned on
[16:38] <ahasenack> eaman: which is not generally a good idea
[16:38] <ahasenack> eaman: so, you shouldn't be using pam_mkhomedir with nfs mounted home directories. The home directories should be created on the server, not on the client
[16:39] <aubre> ttx: since I figured out what was causing my problem in bug https://bugs.launchpad.net/eucalyptus/+bug/452185 , but it could present problems for people who set it up in the future, should I close that bug and open a new one about the interface binding issue?
[16:40] <aubre> ttx: because while my system is working now I guarantee someone in the future will have the same thing happen to them
[16:41] <ttx> aubre: just a sec
[16:43] <heath|work> I'm trying to use checkinstall to install git but the git docs need to be installed also. In the INSTALL file it states to use make prefix=/usr/local install install-doc. How do I tell checkinstall to do that?
[16:44] <ttx> aubre: the fact that you might need to adjust registration of components in the case of multiple NICs on your cluster controller is, I think, a known issue
[16:45] <ttx> aubre: but yes, feel free to invalidate the current bug and create another one
[16:45] <ttx> at the very minimum it will be a documentation wishlist
[16:45] <aubre> aubre: ok. That's what I will make it
[16:45] <aubre> ttx: ok that's what I will make it
[16:45] <ttx> aubre: did you see the latest doc ?
[16:46] <ttx> https://help.ubuntu.com/community/UEC
[16:46] <ttx> feel free to fix it based on your experience
[16:50] <EtienneG> ttx, mathiaz, kirkland: the text of the eucalyptus/publicips debconf template is wrong; you cannot specify range in the form 192.168.1.1-192.168.1.20 for the value of VNET_PUBLICIPS
[16:51] <EtienneG> either that, or there has to be some logic in the postinst to expand that range into a list
[16:51] <EtienneG> or even, Eucalyptus should take it but don't
[16:51] <EtienneG> I am not sure
[16:51] <EtienneG> gotta report the bug, I guess it is a pretty important one
[16:51] <EtienneG> as it result in a non-working Eucalyptus setup
[16:56] <eaman> ahasenac: Thanks, this make sense.
[16:57] <eaman> ahasenack: And of course I'm not willing to use root squash on the nfs server
[16:58] <ahasenack> *no* root squash, you probably mean ;)
[16:58] <eaman> yes sorry
[16:59] <eaman> So now the problems is: finding a nice way to create the user on the server, as the user creation is trigged by a PHP script running on Apache
[16:59] <aubre> ttx: ok
[17:00] <ahasenack> seems like you need to do one more thing in that trigger
[17:00] <eaman> It comes to my mind making Apache (www-data) take some sudo priviledge in order to cast a mkdir and a chown on the new dir
[17:01] <ahasenack> and copy /etc/skel/
[17:01] <eaman> Well I could use a skell population as well I guess
[17:01] <eaman> Yes :)
[17:01] <ahasenack> sounds reasonable, if done carefully
[17:02] <eaman> Well it could be the only way
[17:07] <mdz> soren: ping
[17:07] <mdz> ttx: thanks for testing the appliance, looks like we have some work to do
[17:07] <kaushal> hi
[17:07] <ttx> mdz: yes...
[17:07] <kaushal> can i change catalina.out to catalina.log or something else. if its possible then how is it done ?
[17:08] <kaushal> I am talking about tomcat Application server
[17:09] <eaman> Weird things is that pam_mkhomedir.so is able to cast the right GID: it goes to 2000 here.
[17:10] <ttx> mdz: the other remaining issue is the UEC image / ephemeral disk automount one
[17:10] <eaman> And UID goes to NOBODY, which makes sense
[17:10] <mdz> ttx: right
[17:10] <mdz> ttx: smoser is on that, yes?
[17:10] <ttx> mdz: yes, though he wasn't working today
[17:11] <mdz> ah, right, he's off
[17:11] <ttx> mdz: I still think he is in the best position to find the right way to fix it in the less-intrusive manner
[17:11] <ttx> that would increase our chances of having it pass release management
[17:16] <EtienneG> bug #459204
[17:16] <EtienneG> ttx, if you have a minute, you should look at the above
[17:17] <ttx> I don't have a minute, but I can have a look
[17:17] <EtienneG> ttx, sorry, man!
[17:17] <ttx> EtienneG: yes you can !
[17:17] <ttx> EtienneG: what makes you think you can't ?
[17:18] <EtienneG> ttx, huh?  I am afraid you lost me here.
[17:18] <ttx> EtienneG: specify 192.168.1.1-192.168.1.20
[17:18] <EtienneG> ttx, you can, it just won't work
[17:18] <EtienneG> ttx, easy enough to test without reinstalling
[17:19] <ttx> hmm? works here, as far as euca-describe-addresses goes
[17:19] <ttx> (or whatever the exact command is called)
[17:19] <EtienneG> ttx, change VNET_PUBLICIPS to 92.168.1.1-192.168.1.20", restart eucalyptus
[17:19] <EtienneG> ttx, ah, ok, then it has to be something else
[17:20] <ttx> EtienneG: wht evidence do you have its *not* working ?
[17:20] <ttx> doesn't show the addresses in euca-describe-addresses ?
[17:20] <EtienneG> ttx, evidence?  the output of euca-run-instances, perhaps?  :)
[17:21] <ttx> EtienneG: you mean its not affecting any of the addresses you configured ?
[17:21] <EtienneG> ttx, and yes, the addresses do not show in the the output of euca-describe-addresses either
[17:21] <ttx> hmmm... works here
[17:21] <ttx> (last time I looked)
[17:21] <EtienneG> ttx, which Euca version?
[17:21] <EtienneG> and which VNET_MODE?
[17:22] <EtienneG> ttx, fsck it, mark it Invalid
[17:22] <EtienneG> my bad
[17:23] <EtienneG> VNET_PUBLICIPS did not correspond to a range actually in use by the machine
[17:23] <EtienneG> ttx, that was a user error
[17:23] <EtienneG> ttx, sorry for wasting a few minutes of yours you did not had!
[17:23] <ttx> :P
[17:28] <kaushal> checking in again for my query ?
[17:29] <ruben23> hi can i further optimize ubuntu..? for network- have anyone tried it..?
[17:37] <ruben23> anyone..?
[17:37] <ruben23> have tried it..?
[17:41] <aubre> ttx: how's this ? https://help.ubuntu.com/community/UEC/StorageController
[17:45] <aubre> ttx: I need to make the command line stuff brown
[17:55] <SyL> ruben23: what kinda of optimization are you looking for?
[18:02] <gamla_kossan> hi people
[18:02] <gamla_kossan> I seem to have isntalled some package which has broken my samba server...
[18:02] <gamla_kossan> and the samba server won't start by running /etc/init.d/samba start
[18:02] <gamla_kossan> (nothing happens)
[18:03] <gamla_kossan> can anyone help me troubleshoot?
[18:23] <ruben23> SyL:network and the system itself
[18:30] <zul> mathiaz: ping any reason why autofs would conflict with itself?
[18:31] <mathiaz> zul: with itself or an *earlier* version of itself?
[18:31] <zul> with itself
[18:33] <mathiaz> zul: hm - that's a good question - I don'tknow
[18:33] <zul> debian unstable still has it as well
[18:33] <mathiaz> zul: I'd ask in #ubuntu-devel about it
[18:53] <ninjah> is there a difference between "apt-get dist-upgrade" and "apt-get upgrade"?
[18:53] <Pici> !dist-upgrade | ninjah
[18:54] <nijaba> soren: if you are around I'd love to have your point of view on the following.  I've been trying to upload your mediawiki image to ec2 using http://pastebin.ubuntu.com/299939/ but ec2-register send me the error you'll see there as well....  any idea?
[19:11] <nijaba> soren: nm, sorted
[19:44] <stephank> Hello! I installed 9.04 64-bit on an machine with an onboard 1GB/s intel network interface, and a broadcom 1GB/s pci express card. I bonded these two interfaces in /etc/network/interfaces, created VLANs on top of that and it all works peachy. But when I reboot, the interfaces don't come up saying "VLANs not supported on bond0".
[19:44] <stephank> When the machines it booted, and I restart networking, it all works fine again. I pasted my config here: http://www.pastie.org/667009 . I also added 8021q and bonding to /etc/modules. Is there anything else I'm missing?
[19:50] <qman__> stephank, sounds like a bug having to do with boot order to me, you could probably work around it by adding something to /etc/rc.local as a temporary measure
[19:54] <stephank> qman__: hehe, neat. I'll add a networking restart there. Let's see if this works
[20:00] <EtienneG> smoser, re: uec official image, the bug about waiting for /dev/sdb, is that just a matter of fixing /etc/fstab?
[20:09] <stephank> qman__: trick seems to fool it. Thanks :)
[20:13] <soren> nijaba: Oh, good, because I had /no/ clue :)
[20:13] <soren> mdz: Yes?
[20:20] <nijaba> soren: yeah, your mediawiki img is now instanciated in ec2 :P
[20:20] <soren> nijaba: Working?
[20:20] <nijaba> soren: quite well
[20:20] <soren> nijaba: Are you subscribed to ubuntu-devel?
[20:21] <soren> nijaba: ttx reported a failure, but I suspect he may have unintentionally have used an earlier version of the image.
[20:21] <nijaba> soren: I am, but not reading assiduously
[20:21] <nijaba> soren: I'll reply
[20:21] <soren> nijaba: I'm about to follow up there, mentioning the new version. Once I've done that, can you please post your results?
[20:21] <nijaba> soren: sure
[20:23] <nijaba> I even have it managed with Landscape :)
[20:43] <SyL> is there a way to turn off ipv6 in ubuntu?
[20:43] <genii> !ipv6
[20:44] <genii> SyL: Second link from the bot
[20:45] <SyL> thanks
[20:50]  * nijaba calls it an evening
[20:50] <Zodling> anyone know if the iscsi install works in 9.10-rc ?
[20:57] <ScottK> Zodling: I know there is at least one significant bug about that.  I don't know if it affects all iscsi or just some.
[20:57] <BrixSat> My ubuntu wont start so many error's, i went to a live cd and i cant copy some files because they are under root! How can i copy root files?
[20:58] <ScottK> BrixSat: sudo cp
[20:58] <BrixSat> even on a live cd?
[20:58] <Zodling> i managed to get it to install but.. it dont look to be booting right .. *i have rdp runing if anyone wants to look*
[21:02] <ScottK> BrixSat: I think so.
[21:04] <BrixSat> worked :) thanks
[21:44] <toddobryan> Does anyone have a how-to for OpenLDAP with TLS/SSL support that actually works?
[21:48] <bventura> DNS ?: is it OK to have two A records that with the same IP?   When I google this it says it's not recommended but nothing about why
[21:51] <jpds> bventura: Two different domains?
[21:52] <bventura> jpds: no same domain
[21:52] <jpds> That seems a bit redundant.
[21:52] <maswan> bventura: yes, perfectly fine
[21:53] <bventura> ok cool maswan
[21:53] <maswan> the PTR record should probably only be to one of these though
[21:53] <bventura> right
[21:53] <guntbert> bventura: of there are CNAME records too, but don't use one for your mail server
[21:53] <guntbert> *of course
[21:54] <maswan> But "kermit A 10.0.01" + "mail A 10.0.0.1" + "smtp A 10.0.0.1" + [in the reverse zone file] "1 PTR kermit" is perfectly fine
[21:55] <maswan> In fact, there is sufficient weird corner cases for CNAMEs that I wouldn't really recommend them for most places
[21:55] <bventura> right guntbert, that's what got me here, I can't use CNAME to map my mail server
[21:55] <maswan> Since an IP can have lots and lots of A records.
[21:55] <bventura> ok
[21:56] <guntbert> bventura: but as maswan already pointed out: use two differnt A records
[21:56] <bventura> i just wasn't sure when I google it, 2nd result says "When the domain name system was designed it was recommended that no two A records refer to the same IP address."
[21:57] <maswan> the second paragraph of that is probably "but these days, ..."
[21:59] <guntbert> bventura: that must have been before mail admins decided not to accept CNAMEs any more
[22:23] <blizzkid> Lo all. does anyone have experience in taking exams for a linux course?
[22:38] <bventura> OK i think I've put all this DNS stuff together , would anyone be kind enough to critique my config, it is here: http://ubuntu.pastebin.com/m1d75d683
[22:43] <stimble> bventura: r u from Ventura?
[22:43] <stimble> I live in a town named ventura
[22:43] <bventura> cool in CA?
[22:43] <stimble> yeah
[22:43] <bventura> I used to live down there but ithat's actually my last name
[22:43] <bventura> i gre up in newbury park
[22:43] <stimble> oh
[22:43] <stimble> oh, right next door
[22:43] <stimble> cool
[22:44] <stimble> small world on the internets :-)
[22:44] <bventura> ah i miss that southern california weather tho, i'm in sonoma now
[22:44] <bventura> :D
[22:44] <stimble> well, good wine and near San Fran at least (2 good things in my book)
[22:44] <bventura> oh yeah
[22:45] <bventura> that other recreational substance is readily available too hehee
[22:45] <stimble> ha
[22:45] <stimble> btw, whats your dns config for, just an example?
[22:46] <bventura> at this point yes, I'm just trrying to figure out how it's supposed to work but eventually I have to replace my company's primary DNS and mail server
[22:47] <bventura> we have an anicent xserve running OS X server and I hate it.. I want ubuntu
[22:47] <stimble> ah
[22:47] <stimble> yeah, i have been considering getting an automated dns setup for my ec2 deployments
[22:47] <stimble> but i hate dealing with bind, soo... havnt done it
[22:48] <bventura> yeah it's a bit challenging
[22:48] <guntbert> bventura: just a reminder: use a LTS for that purpose
[22:48] <bventura> guntbert - I have been wondering what you just mentioned can you tell me why?
[22:48] <bventura> not trying to argue I just dont think I get it
[22:49] <bventura> LTS = more secure?
[22:49] <ScottK> No, LTS = having to upgrade less often.
[22:50] <ScottK> Actually we've introduced a number of hardening features after Hardy, so the newer releases are 'more' secure.
[22:50] <alex_joni> LTS = Long Term Support .. so you get upgrades for longer
[22:50] <guntbert> bventura: In my experience such servers will be put into service and the "forgotten"  - so .. ^^^
[22:50] <guntbert> *then
[22:50] <bventura> i'm not sure what you mean by 'get' i thought they all were free?
[22:51] <bventura> 'get upgrades' *
[22:51] <guntbert> !eol | bventura
[22:51] <bventura> eureka!  OK I gotcha now I get it
[22:52] <bventura> so 8.04 LTS is the way to go i suppose
[22:53] <JanC> if you want to run BIND for serious purposes, buying a book about BIND might be useful...
[22:53] <stimble> or, perhaps run tinydns
[22:53] <JanC> or at least read all the docs
[22:53] <bventura> I have been chewing thru the docs on isc.org
[22:53] <JanC> stimble: still, reading about how DNS works is very useful  ツ
[22:53] <stimble> or if you dont need all the bind power, i have found dnsmasq to fill most of my LAN needs
[22:54] <stimble> yes, true
[22:54] <bventura> i just wanted to get a 2nd pair of eyes on what I have so far
[22:54] <bventura> i'm in a 1-man IT department over here ;p
[22:55] <stimble> I have a fresh 6 nodes UEC cloud/cluster and am getting internal IPs of 0.0.0.2.  Anyone have hints to troubleshoot this?
[22:55] <JanC> for a LAN, dnsmasq is absolutely great (it just does what you need without all the complicated stuff)
[22:57] <mathiaz> stimble: internal IPs? do you refer to NC private IPs or instances IPs?
[22:57] <stimble> JanC:  yeah, thats what I have found.  And it even does tftp and dns and dhcp together, which makes for a pretty nice integration, as in nice to have dhcp update your dns etc
[22:57] <stimble> instance ip
[22:57] <stimble> im using managed mode
[22:57] <JanC> yep, I used it for that (and so do many "home routers")
[22:57] <mathiaz> stimble: MANAGED-NOVLAN?
[22:57] <stimble> no
[22:57] <stimble> MANAGED
[22:58] <mathiaz> stimble: is your network VLAN aware?
[22:58] <stimble> yes
[22:59] <stimble> i guess i can switch to novlan and try it again
[22:59] <mathiaz> stimble: not necessarly
[22:59] <mathiaz> stimble: just making sure of your configuration
[22:59] <stimble> but i checked it for vlan support a while back and was pretty certain it came up positive
[22:59] <mathiaz> stimble: what does euca-describe-addresses say?
[22:59] <stimble> i have VNET_SUBNET="192.168.6.0"
[23:00] <stimble> it lists the public_ips i have in my eucalyptus.conf
[23:00] <stimble> 192.168.2.100 - 110
[23:00] <stimble> 192.168.2.100 is assigned to the one running instance
[23:00] <mathiaz> stimble: is there another dhcp server running on the network?
[23:00] <stimble> but i cnat ping it
[23:01] <stimble> yes
[23:01] <mathiaz> right - is the dhcp server on the same vlan as the NC?
[23:01] <stimble> well, it is plugged into the same switch
[23:02] <stimble> but, i would think it should not be on the same vlan
[23:02] <stimble> its a pfsense box
[23:02] <mathiaz> stimble: have you checked that you have multiple VLANs and set the switch correclty?
[23:02] <stimble> its an unmanaged switch
[23:03] <mathiaz> stimble: if the dhcp server is on the same vlan as the UEC nodes, then there is a chance that the instances got their IP from the external dhcp server rather than the UEC dhcp server
[23:03] <stimble> yeah, but then it should have been in the 192.168.2.0/24 network
[23:03] <stimble> its 0.0.0.2, which just seems bonkers
[23:04] <mathiaz> stimble: how did you notice that the instances had 0.0.0.2?
[23:04] <stimble> euca-describe-instances
[23:06] <mathiaz> stimble: could you pastebin the output of the command?
[23:06] <stimble> doing that now
[23:06] <mathiaz> stimble: command == euca-describe-instances
[23:07] <stimble_> mathiaz: http://pastie.org/667272
[23:07] <stimble_> im working from 2 machines (hence the 2 irc handles)
[23:07] <Burebista> hello
[23:08] <stimble> command output is at the bottom
[23:08] <stimble> hello Burebista
[23:08] <Burebista> i try to install a kvm server. the host server are done, but how can i install a virt machine/
[23:09] <Burebista> -topic
[23:09] <mathiaz> stimble: could you paste the output of: grep ^VNET /etc/eucalyptus/eucalyptus.conf
[23:10] <stimble_> updated pastie
[23:10] <stimble_> http://pastie.org/667272
[23:12] <mathiaz> stimble_: hm - well - I don't know - sorry
[23:12] <stimble_> just updated again with the full uncommented eucalyptus.conf
[23:35] <stimble> when UEC installs a eucalyptus node, it never asks to create a username, password.  Anybody know how to still log into a sudo capable account on a node controller?