[02:01] <Rascal999> when i try and ssh into box with public key encryption auth.log says Error attempting to add filename encryption key to user session keyring; rc = [1]. What do I need to change?
[02:33] <billybigrigger> anyone aware of a decent app to graph cpu usage, network traffic and disk io?
[02:36] <jmarsden> billybigrigger: gkrellm if you want on screen graphs; mrtg or similar things using rrdtool if you want longer term trends and graphs on web pages...
[02:37] <jmarsden> billybigrigger: cricket or cacti too, for the web-based type of monitoring...
[02:38] <mathiaz> billybigrigger: munin - is in main since hardy
[02:39] <mathiaz> billybigrigger: cacti is in universe while cricket is dead  upstream
[03:02] <benedikt> billybigrigger: hobbit (now xymon)
[03:07] <billybigrigger> cacti looks nice, just a little big more than what i need
[03:21] <ScottK> mathiaz: Is cricket dead enough we want it removed?
[03:22] <mathiaz> ScottK: yeah - seems like it
[03:22] <ScottK> mathiaz: Would you please file a removal bug then?
[03:23] <mathiaz> ScottK: yeah - that's one of the next step
[03:23] <mathiaz> ScottK: first we discuss it
[03:23] <ScottK> Thanks.
[03:23] <ScottK> mathiaz: Didn't we just do that?
[03:23] <mathiaz> ScottK: and then we take all appropriate actions
[03:23] <ScottK> My threshold for removals is pretty low.
[03:23] <mathiaz> ScottK: we'll do a last round of requests for comments with the whole list
[03:24] <mathiaz> ScottK: I'm still working on potential packages to be moved out of main/universe
[03:24] <ScottK> OK.  Just finished my mail server spec.  Hopefully ivoks can get a chance to look at it soon.
[03:24] <mathiaz> ScottK: cricket is in the list anyway
[03:24] <ScottK> OK.  Good.
[03:24] <ScottK> I'm having a very hard time following that spec based on email.
[03:33] <whatchasay> !ops
[03:34] <Pici> whatchasay: Whats up?
[03:34] <whatchasay> !ops
[03:34] <whatchasay> peace dog
[03:34] <Pici> whatchasay: Why are you doing that?
[03:35] <MTecknology> lovely..
[03:45] <Eloff> ...
[07:58] <jetole> hey guys, I fubar'd my sudoers file on a remote server and I can no longer run sudo, I can't use su because there is no root password, does anyone know how else I can get a root shell?
[08:03] <kane_> jetole: can you reboot single user mode and access the serial console?
[08:10] <quizme> is there a way for me log into my friend's computer and have him watch the commands that i type into his terminal ?  cuz i want to show him some stuff.
[08:11] <arj> screen -x
[08:11] <kane_> quizme: shared screen session is the way to go, or use vnc for graphical sharing
[08:11] <quizme> shared screen session sounds good
[08:11] <quizme> how do i do that ?
[08:12] <quizme> what does he need to do
[08:12] <quizme> he is at home
[08:12] <kane_> quizme: ubuntuforums are great for this: http://ubuntuforums.org/showthread.php?t=299286
[08:13] <quizme> thanks
[09:04] <quizme> i'm trying to forward messages from port 9000 on my server to port 80 on my localhost.  Does anybody know how to construct that ssh command?
[09:04] <kane_> quizme: tried man ssh?
[09:04] <quizme> ssh -gNR 127.0.0.1:80:thirdreplicator.com:9000 dev@thirdreplicator.com  <---  i tried that
[09:04] <quizme> i tried that
[09:04] <kane_> you're missing a -L infront of the 127
[09:05] <quizme> oh
[09:05] <quizme> i need -L and -R ?
[09:05] <quizme> ssh -gNR -L 127.0.0.1:80:thirdreplicator.com:9000 dev@thirdreplicator.com
[09:05] <quizme> ?
[09:06] <kane_> you just want -N -L 80:thirdreplicator.com:9000
[09:06] <quizme> what about -g ?
[09:07] <quizme> i'm typing this locally
[09:07] <quizme> not on the server
[09:07] <kane_> quizme: if you want to know more, you should really read a howto, like this: http://www.debianadmin.com/howto-use-ssh-local-and-remote-port-forwarding.html
[09:07] <kane_> google's your friend
[09:08] <quizme> ssh -N -L 80:thirdreplicator.com:9000 dev@thirdreplicator.com
[09:08] <quizme> Privileged ports can only be forwarded by root.
[09:08] <quizme> i have been googling for hours
[09:08] <quizme> and manning for hours
[09:08] <quizme> that's why i'm here
[09:08] <kane_> you didn't type 'ssh port forwarding howto' then
[09:08] <kane_> quizme: you can forward only non-privileged ports as a normal user
[09:08] <kane_> so, use 1080 instead for example
[09:08] <quizme> google is not my friend.
[09:09] <quizme> he is my enemy at this point
[09:09] <kane_> there's bing ;)
[09:09] <quizme> lol
[09:09] <kane_> take a look at the link i gave you, it walks you through step by step
[09:09] <kane_> if you still can't make it work, feel free to ask here
[09:10] <quizme> thanks
[09:10] <quizme> reading..
[09:10] <twb> quizme: "ssh -NL 8080:127.0.0.1:80 fs" connects 127.0.0.1:80 on fs to 8080 on the ssh client.
[09:11] <twb> quizme: I suspect you just got the 80 and 9000 the wrong way around
[09:12] <quizme> oh
[09:12] <quizme> yeah probably
[09:13] <quizme> and i don't know the difference between -L and -R
[09:13] <quizme> let me try
[09:14] <twb> quizme: -L connects a local port to a remote port, -R connects a remote port to a local port.
[09:14] <twb> quizme: you almost always want -L
[09:15] <quizme> are you assuming this command is run on the server or on my laptop ?
[09:15] <twb> quizme: laptop
[09:15] <quizme> ok
[09:15] <quizme> that's what i'm assuming
[09:16] <twb> e.g. ssh -fNL 8080:127.0.0.1:80 www.foo.com && sensible-browser http://localhost
[09:17] <quizme> maybe i should explain what i am trying to do
[09:17] <quizme> i'm trying to suck down messages sent to port 9000 on my server down to my localhost's port 80 (apache) server.
[09:18] <twb> quizme: for that, you want -R
[09:18] <quizme> ok.. hehe
[09:18] <jetole> kane, I already went to the data center and modified the kernel line to have init=/bin/bash
[09:18] <jetole> back already
[09:19] <twb> jetole: IME "single" or "rescue" is less hassle (if they work)
[09:19] <jetole> twb, yeah I didn't try rescue cause I was kinda under the suspicion that it would need a root passphrase which there isn't one
[09:20] <twb> jetole: Ubuntu doesn't ask
[09:20] <kwork> you can always get away with single
[09:20] <kwork> if you dont know the root pw
[09:20] <twb> kwork: no, on a normal system single will still require you to enter the root password
[09:20] <quizme> ssh -NR thirdreplicator.com:9000:127.0.0.1:80 dev@thirdreplicator.com
[09:20] <quizme> does that look right ?
[09:20] <jetole> twb well for my server thats cool
[09:21] <twb> kwork: Ubuntu is weird in that it doesn't
[09:21] <jetole> they are all in a locked rack
[09:21] <kwork> twb,  hmmmz im pretty sure you can get away without root on debian awell
[09:21] <twb> It annoys me, even though I know that if they have physical access you're screwed in other ways.
[09:21] <kwork> twb,  i remember doing smt like that
[09:21] <twb> kwork: if root's password is the null password (not the same as NO password), then Debian's single will say "hit enter for a maintenance shell" instead of "enter root password for a maintenance shell"
[09:22] <kwork> but maybe you needed cd for it
[09:22] <quizme> twb thanks!
[09:22] <jetole> does anyone know how I can setup open-iscsi to map to a specific disk on each boot or a static path to access it by?
[09:22] <jetole> oh wait
[09:22] <twb> kwork: yes, you can also do "rescue" from a d-i CD
[09:22]  * jetole looks at uuid
[09:22] <twb> kwork: d-i rescue doesn't require a root password, on ubuntu or debian
[09:22] <kwork> twb,  basicaly if i can chroot into the system i should be able to change the pw right ?
[09:22] <quizme> twb: it worked!
[09:22] <twb> jetole: /dev/disk/by-uuid?
[09:23] <jetole> thats what I am thinking
[09:23] <twb> jetole: run vol_id on the device name you already know
[09:23] <jetole> I see the UUID there and I am assuming they consistantly map to the same iscsi drives on each boot
[09:23] <jetole> twb: I don't have vol_id on ubuntu server 9.10
[09:23] <twb> jetole: it's part of udev!
[09:24] <twb> jetole: even 8.04 has it
[09:25] <jetole> twb: is it a bin? volname is the only thing bash tab completes starting with vol and `which` comes back empty
[09:25]  * jetole googles it and looks for a man page so I at least know what it did
[09:26] <jetole> ah
[09:26] <twb> jetole: are you still in init=/bin/bash?
[09:26] <jetole> oh no
[09:26] <twb> jetole: vol_id should be in /sbin/
[09:26] <jetole> I went to the data center and am now back at my office
[09:27] <jetole> twb, bash won't tab complete it for "vol" and which doesn't return anything
[09:27] <twb> Shrug
[09:29] <jetole> oh shit
[09:29] <jetole> /dev/disk/by-path/
[09:29] <jetole> that shows the full iqn names
[09:30] <jetole> that is something better since I have vm booting off iscsi
[09:31] <twb> Shrug
[11:44] <xperia> hello to all. is here somebody with experince hot to install andconfigure red5
[11:55] <jussi01> Hi all, where are samba passwords usually managed?
[12:02] <RoyK> jussi01: that depends how you configure samba
[12:03] <jussi01> RoyK: hrm, well thats frustrating as Ive no idea how the person before me configured it.
[12:05] <RoyK> jussi01: just look in smb.conf
[12:05] <RoyK> it's pretty self-explainatory after having read through that and its comments
[12:07] <jussi01> right, ive just been reading man pages, I can change a logged in users password with smbpasswd but how do I list all of the users?
[12:11] <Daviey>  /3
[12:17] <jussi01> Nevermind, I got it figured.  :)
[13:01] <zul> morning
[13:42] <MTecknology> Is there any easy way to have all my logs in one central location?
[13:42] <kwork> syslog central server ?
[13:42] <kwork> and all servers sending syslog to that box
[13:43] <MTecknology> nifty - thanks
[13:51] <MTecknology> kwork: That looks amazingly simple - http://news.softpedia.com/news/Setting-Up-a-Central-Syslog-Server-44063.shtml
[13:51] <kwork> yeah followed the same tutorial some days ago
[13:51] <kwork> works like a charm :)
[13:52] <MTecknology> :D
[13:52] <kwork> okey actualy lies
[13:52] <kwork> or atleast karmic has rsyslogd
[13:52] <kwork> but its really similar
[13:58] <MTecknology> I love  ufw + apparmor + denyhosts + central logging
[13:58] <MTecknology> don't have the logging server just yet though
[13:58] <MTecknology> kwork: where will the logs be stored?
[14:02] <kwork> MTecknology,  you define file per host
[14:03] <jdstrand> (ufw in lucid will ship an rsyslog configuration)
[14:04] <MTecknology> ok, thanks
[14:06] <smoser> nijaba, or anyone else maybe can answer
[14:06] <smoser> i'm under the impression that backports do not get "official canonical support" for the lifetime of normal support
[14:06] <smoser> is that correct?
[14:06] <nijaba> smoser: correct on backport support
[14:07] <smoser> thanks
[14:11] <ScottK> smoser: Generally if there's a problem with a backport, we just update it with a newer backport to fix the problem.
[14:19] <xperia> hello to all ! anybody here how can help me with virtual hosting ?
[15:23] <MTecknology> xperia: the question seems to be missing from the backlogs
[15:26] <xperia> MTecknology: have solved it thank you a lot !
[16:20] <rags> Hello, i'm using racoon to establish ipsec tunnels..can some one tell me how to remove a SA manullly...
[16:20] <rags> I nknow I can use racoonctl to remove isakmp sa'...based on peer address...
[16:20] <zul> jjohansen: ping
[16:20] <rags> but this doesn't seem to work with SA's....
[16:21] <rags> Also, I've found it very difficult to find any documentation on ipsec-tools and racoon...am I using something tht is archaic?...wht do ppl usually go for?...OpenSwan?
[16:25] <rags> hmm..does anybody work on ipsec here?
[16:28] <CyberSnooP> What's the proper way to prevent "localhost [127.0.0.1]" mentionings in outgoing mail with postfix ?
[16:42] <ScottK> CyberSnooP: Why do you care?
[16:43] <CyberSnooP> Mail I'm sending gets immediately marked as spam.
[16:43] <CyberSnooP> (I've just started sending registration confirmation mails to users)
[16:44] <CyberSnooP> My SPF-records are okay, reverse DNS is checked and lot's of servers seem happy (like gmail). But Hotmail and some universities immediately throw away the message
[16:44] <CyberSnooP> (without anybody marking it as junk, as it's the first time we mail them.. new server, new website etc..)
[16:44] <CyberSnooP> So, I've read that spam-scanners care about "localhost" mentionings
[16:49] <CyberSnooP> ScottK: Does that make any sense or does it make clear that I shouldn't run a mail-server at all :P
[16:50] <ScottK> You can remove them with Postfix header checks, but be careful as it's done with regular expression matching, so it's easy remove more than you want
[16:50] <ScottK> I think it's unlikely to help, but I've heard other people give similar theories.
[16:50] <ScottK> People do weird stuff to try to filter spam, so who knows.
[16:54] <eagles0513875> !ufw
[16:54] <adurity> Is there a way to hold a package at a specific version, so that when apt checks for updates, it ignores updates to that particular package?
[16:55] <MTecknology> adurity: apt-get {hold,unhold}
[16:55] <adurity> thank you!
[16:55] <MTecknology> :)
[16:58] <eagles0513875> hey guys im just wondering
[16:58] <eagles0513875> with ufw is there a way i can allow access on a range of ports
[16:59] <MTecknology> eagles0513875: you'd need to specify each individually
[16:59] <adurity> MTecknology, even better, I found the Forbid option
[16:59] <jdstrand> MTecknology: actually it depends on the version
[17:00] <jdstrand> eagles0513875: what version of ufw?
[17:00] <eagles0513875> jdstrand: 0.29
[17:00] <eagles0513875> the server is running karmic
[17:00] <jdstrand> eagles0513875: yes, ufw supports multiport rules
[17:01] <MTecknology> jdstrand: oh.. nifty
[17:01] <eagles0513875> how do i do that im follwoing the ufw link the bot provides
[17:01] <jdstrand> eagles0513875: eg: ufw allow proto tcp from any to any port 80,443,8080:8090
[17:01] <eagles0513875> https://help.ubuntu.com/community/UFW?action=show&redirect=Uncomplicated_Firewall_ufw
[17:01] <jdstrand> eagles0513875: it's in the man page
[17:01] <eagles0513875> so basically sudo ufw allow 6000-7000
[17:01] <jdstrand> eagles0513875: man 8 ufw
[17:02] <jdstrand> eagles0513875: you must use the extended syntax
[17:02] <CyberSnooP> ScottK: header_checks succeded in removing the header. But at least hotmail still doesn't seem to like my e-mails (and they don't say why)
[17:02] <jdstrand> eagles0513875: eg:
[17:02] <jdstrand> ufw allow to any port 6000-7000
[17:03] <ScottK> CyberSnooP: Hotmail is very random.  No one outside Hotmail really knows how it works, AFAIK.
[17:03] <jdstrand> MTecknology, eagles0513875: the ufw features list can be seen here: https://wiki.ubuntu.com/UncomplicatedFirewall#Features
[17:03] <CyberSnooP> Yeah, well I hope other mail-servers will start to work due to this change at least.
[17:04] <CyberSnooP> Thanks for the tip anyway :)
[17:06] <eagles0513875> jdstrand: i did like the example you gave me but its saying bad syntax in regards to 6000-7000
[17:07] <jdstrand> eagles0513875: oops
[17:07] <jdstrand> ufw allow to any port 6000:7000 proto tcp
[17:07] <jdstrand> eagles0513875: ^
[17:07] <eagles0513875> jdstrand: ahhh ok ty
[17:07] <jdstrand> eagles0513875: assuming you want tcp...
[17:07] <eagles0513875> sry for the 20 questions im green when it comes to ufw
[17:07] <eagles0513875> ya
[17:41] <mgpcoe> Having trouble getting Postfix+Dovecot to listen on port 25.. SSL and TLS say hello fine, but basic SMTP is totally nonresponsive... anybody run into this before?
[17:42] <lamont> could it be that your ISP is blocking it?
[17:42] <lamont> mgpcoe: that's usually the issue
[17:43] <mgpcoe> lamont: Thought that, but I tried SSHing to an outside server and going in that way.
[17:43] <mgpcoe> lamont: Same result, and I'm sure I've been able to get in on port 25 from that server in the past.
[17:44] <mgpcoe> Oh, FFS, when I tried doing it from the server, using the public IP I got right in.. this just became a whole other problem, I think.
[17:46] <mgpcoe> Does Hardy do any filtering on 25 by default?
[17:46] <lamont> ufw might, but installing postfix should open that
[17:47] <lamont> what does lsof -ni :25 have to tell us (as root)
[17:48] <mgpcoe> lamont: Gives me two lines, one on IPv4, one on IPv6, both NODE: TCP and NAME: *:smtp (LISTEN)
[17:48] <mgpcoe> lamont: And they're both running as root, if that helps; command is `master'
[17:56] <billybigrigger> master = postfix
[17:57] <billybigrigger> mgpcoe, iptables blocking it?
[17:59] <lamont> and (just for completeness...) does the IP you're trying and failing to connect to exist on the machine, or somewhere else  (I hate it when I do that)
[17:59] <mgpcoe> billybigrigger: I'm not sure; how would I find out? Last time I tried listing anything with iptables, I couldn't make heads or tails of the information.
[18:00] <mgpcoe> lamont: Yeah, usually I just enter the domain and let it resolve itself. IP matches all the way across the board.
[18:01] <lamont> and "ip route get $IP" points where you think it should?  (lo, I expect..)
[18:01] <lamont> mgpcoe: iptables -t nat -nvL; iptables -nvL
[18:02] <lamont> and simplest to toss that output into paste.ubuntu.com
[18:02] <mgpcoe> lamont: ip route tells me "local IP.IP.IP.IP dev lo  src IP.IP.IP.IP"
[18:04] <mgpcoe> lamont: http://paste.ubuntu.com/333984/
[18:06] <lamont> mgpcoe: so what that says is you have no iptables rules, and are using the default (accept) for everything
[18:06] <lamont> which then gets to "what command are you using to talk to it?"
[18:07] <mgpcoe> Right now, just telnet x.x.x.x 25
[18:08] <VaineDragon> I just did a fresh insta and configure of pure-ftpd and am unable to logon, here is the output: http://pastebin.com/d4eb60235
[18:08] <mgpcoe> lamont: which merrily reports "trying x.x.x.x..." and never gets there. I'm tailing /var/log/mail.log on the server and it never even reports the attempt.
[18:09] <lamont> mgpcoe: so... "tcpdump -ni lo port 25" as root and then telnet to the IP...
[18:10] <lamont> should show the normal SYN SYN+ACK ACK - and at that point, I suspect that postfix is trying to resolve your source IP and is failing miserably - lets look at /var/log/mail.log, and /var/spool/postfix/etc/resolv.conf and see if they tell us anything useful
[18:10] <lamont> and, (seriously), see if leaving it alone for 5 minutes makes a difference at all
[18:11] <mgpcoe> lamont: Doesn't show a thing... I'm starting to wonder if the external server I'm using to access port 25 is crippled on that port too..
[18:12] <lamont> well, if you're talking on lo, you should see yoursefl
[18:12] <mgpcoe> Aha, righto. One second.
[18:12] <mgpcoe> There it is.
[18:13] <mgpcoe> lamont: When I connect from the server itself, it doesn't have an issue
[18:14] <mgpcoe> lamont: Problem is I need to be able to connect to this thing from the cloud, with smtp auth.. and Evolution and Thunderbird won't even connect.
[18:14] <ScottK> mgpcoe: Use Port 587 (submission)
[18:18] <mgpcoe> ScottK: I do, and SMTPS for dumber clients, but when Evolution can't even figure out what auth mechanisms are supported, even I know there's something Wrong.
[18:19] <ScottK> You're mixing different problems then.
[18:19] <ScottK> Do you need port 25 or do you need MUA's to be able to submit?
[18:21] <lamont> the cloud blocks port 25, so that spammers don't spam from the cloud.
[18:21] <mgpcoe> ScottK: Well, I'm trying to set up an SMTP server for a business client so that they can send email from their domain. The server's a virtual host somewhere in, I don't know, Texas, so in order to let my client use the server for their outgoing mail, but not let just anyone do it, I'm trying to set up authenticated SMTP. I can get into it with 465 and 587, and they appropriately bitch about it, but it looks like 
[18:22] <lamont> like most ISPs, outbound port 25 gets blocked... enable submission (587) and use that
[18:22] <mgpcoe> (and by bitch about it, I mean they give me the relay access denied message, as they should)
[18:25] <MTecknology> If I start opening ssh to certain IP's with ufw; am I safe to assume anything not in the ip ranges given will be blocked; or like hand written iptables, do I need to specify a default reject ?
[18:26] <jmarsden|work> MTecknology: man ufw.  You get to choose.  sudo ufw default deny incoming    # is probably what you want.
[18:26] <BeardedChimp> exit
[18:26] <BeardedChimp> oops :P
[18:27] <MTecknology> jmarsden|work: thanks :)
[18:28] <jmarsden|work> MTecknology: You're welcome... but try to get in the habit of reading the man page *before* asking on IRC :)
[18:29] <zul> jjohansen: ping when you are around I got a kernel package question for you
[18:29] <BeardedChimp> I have a ppp0 device connected to a server. When running iptables -t nat -A PREROUTING -i ppp0 -j accept ; it gives no hits for packets accepted even though I can see packets coming in through tcpdump -i ppp0. This is very confusing because I'm trying to dnat the incoming packets but the nat table doesnt seem to see them
[18:30] <mgpcoe> lamont: I'm starting to wonder if it might be worth it to just gun all the email set it up and do it again from scratch. I've been working from, like, six different tutorials because nothing's been able to get it completely set up...
[18:30] <jjohansen> zul: whats the question?
[18:31] <zul> jjohansen: im suppose to be reviewing an asterisk MIR and it depends on a couple of kernel-packages such as zaptel is there a linux-restricted-modules package for lucid anymore or is all dkms?
[18:31] <lamont> mgpcoe: delivering to port 25 from the cloud is a non-starter.  figuring out whether or not the ISP blocks outbound to port 25 is the first step of any such attempt
[18:31] <jjohansen> zul: all dkms
[18:32] <zul> jjohansen: crap
[18:32] <zul> jjohansen: ok thanks ;)
[18:32] <jjohansen> yeah, its a pain
[18:33] <jjohansen> have you tried asterisk at all?  Do we need to up the kernel Hz for it?
[18:33] <zul> no i havent i dont have a landline
[18:33] <jjohansen> hehe me neither
[18:34] <mgpcoe> lamont: I'm 99% certain that in one of the business partners' case it does, but I don't know what ISP the other uses. I figure, if Gmail and my old university can get it set up that I can use their outgoing servers instead my ISP's, it's not out of the realm of possibility for me to set the same thing up... Apparently it's just Really Exceedingly Difficult, or I'm missing something.
[18:35] <lamont> that's what god invented the submission port for.  If you're an MTA, you use port 25, if you're submitting mail outbound, then you use submission.
[18:37] <mgpcoe> lamont: So, users from the Tubes would be connecting to submission rather than smtp, right? Do I have to specify to the clients that they need to use a different port?
[18:37] <lamont> yes, and yes.
[18:38] <lamont> at least, that's my expectation
[18:38] <lamont> and experience
[18:38] <lamont> afk
[18:38] <mgpcoe> lamont: Oh, I was afraid of that... so Evolution's probably trying to make the auth mechanisms connection over smtp, because I never see that connection.
[18:41] <smoser> jjohansen, fyi
[18:41] <smoser> IMAGE aki-b8de3cd1   ubuntu-kernels-testing-us/ubuntu-lucid-i386-linux-image-2.6.31-302-ec2-v-2.6.31-302.7-kernel.img.manifest.xml
[18:41] <smoser> IMAGE aki-9436d4fd   ubuntu-kernels-testing-us/ubuntu-lucid-i386-linux-image-2.6.32-300-ec2-v-2.6.32-300.1-kernel.img.manifest.xml
[18:42] <smoser> those went up last night, and the latest amis have them as their kenrel. just verified that
[18:42] <smoser> ami-4037d529   ubuntu-images-testing-us/ubuntu-lucid-daily-i386-server-20091203.manifest.xml
[18:42] <smoser> boots fine
[18:42] <jjohansen> nice :)
[18:44] <smoser> hm.. my aki pastes above were supposed to be i386 and amd64, not old i386 and new i386, but you get the picture. the -2.6.32-300 kernels are up there now.
[18:44] <erichammond> smoser: The S3 "location" is "us-west-1" for AMIs in the "us-west-1" EC2 region.  This means we chose a suboptimal (inconsistent) naming scheme for the us-east-1 buckets, but life goes one.
[18:45] <smoser> this is released information ? link ?
[18:45] <erichammond> Yep, very little fanfare
[18:45] <smoser> does suck that i chose sub-optimal naming :-(
[18:45] <mathiaz> kirkland: whody!
[18:45] <zul> come si come ca
[18:46] <mathiaz> kirkland: did you write up a wiki page with minimal configuration for UEC systems?
[18:46] <erichammond> smoser: Oh, by the way, boot from EBS also launched :)
[18:46] <smoser> where do you see such things?
[18:47] <erichammond> http://developer.amazonwebservices.com/connect/ann.jspa?annID=537
[18:47] <erichammond> http://aws.typepad.com/aws/2009/12/expanding-the-aws-footprint.html
[18:48] <smoser> http://aws.amazon.com/about-aws/whats-new/
[18:49] <smoser> thanks. hmm... now, i wonder what we do about naming
[18:49] <smoser> should i say all new stuff goes into newly named buckets? or new region stuff goes into newly named buckets.
[18:54] <kirkland> mathiaz: what do you mean by minimal configuration?
[18:54] <mathiaz> kirkland: things like RAM, CPU, disk sapce
[18:54] <mathiaz> kirkland: *space*
[18:54] <kirkland> mathiaz: yeah!
[18:55] <Aison> hmm, with ldap, everything is so complicated now :( how do I add a new schema?
[18:55] <kirkland> mathiaz: all of them are under https://help.ubuntu.com/community/UEC
[18:55] <Aison> on 9.10 it's quite crazy
[18:55] <kirkland> mathiaz: you're looking for https://help.ubuntu.com/community/UEC/SystemRequirements
[18:55] <mathiaz> kirkland: I'm writing up the hardware requirement for UEC testing
[18:55] <smoser> root volumen on ebs is very interesting.... i wonder jjohansen if you could hibernate
[18:55] <smoser> and resume from hibernation on ebs volume
[18:56] <mathiaz> kirkland: ty
[18:56] <kirkland> mathiaz: no problem; feel free to update that page if you have any other suggestions
[18:56] <jjohansen> hrmm, that would be interesting.  I haven't looked into ebs at all but as long as it could be mounted early it should be possible
[18:56] <kirkland> mathiaz: this was intended as a rough guide for our users
[18:58] <erichammond> smoser: I don't think hibernation is an option for the new "stop" state on EC2.  It is possible to do things like change the kernel associated with the instance while it is stopped.
[18:59] <smoser> well, yes, but if you didn't change the kernel, you could presumably resume from hibernate
[18:59] <erichammond> smoser: "stop" is a shutdown
[18:59] <smoser> change kernel or do other things to the volume that would result in inconsistent state for resume
[19:00] <smoser> hm.. yeah, but if started from user space
[19:01] <erichammond> smoser: I recommend the new buckets be named the same as the existing ones, replacing "us" with "us-west-1" and I recommend grabbing them quickly.
[19:01] <smoser> i think i might own them already
[19:03] <smoser> erichammond, http://paste.ubuntu.com/334039/
[19:03] <neonfreon>  /wg 1
[19:03] <smoser> http://paste.ubuntu.com/334040/ is the complete list of canonical owned buckets
[19:04] <smoser> my land grab paid off
[19:04] <erichammond> smoser: I wouldn't change the existing "us" ones to "us-east-1".  It gets confusing to have multiple buckets for a single region.
[19:06] <MTecknology> I've been reading up on remote rsyslog but I'm seeing that the information is passed in clear text and there's no method to actually prevent these messages from being altered
[19:06] <smoser> erichammond, yeah, i think you'reright.  i think we'll stay "consistent" in that the naming convention is <basename>-<S3_LOCATION>
[19:06] <smoser> for a region
[19:10] <erichammond> smoser: I'm going to be mostly unavailable for the next 6-10 hours, but if you have any quick questions about migration, feel free to give me a call on my cell.
[19:10] <erichammond> bug 492037
[19:11] <smoser> nice. thanks, erichammond (the bug).
[19:12] <erichammond> smoser: I'd recommend dropping whatever else you were doing today and performing the migration.  EC2 customers are now waiting on *you* to use the new region that Amazon has released.  Welcome to EC2 AMI support and maintenance :)
[19:12] <smoser> :)
[19:12] <smoser> i dont know.. its kind of nice to have people wait on me. makes me feel powerful
[19:12] <zul> crazy chicano
[19:19] <mathiaz> kirkland: could you have a quick look at https://wiki.ubuntu.com/UECTesting - in the Test/Demo Plan section
[19:19] <kirkland> mathiaz: sure
[19:19] <mathiaz> kirkland: does the hardware requirements and network topology requirements look sane/comprehensible?
[19:23] <kirkland> mathiaz: yeah, totally
[19:24] <kirkland> mathiaz: you should be able to get up to 4 small guests on each of the NCs
[19:24] <kirkland> mathiaz: and host a decent number of images
[19:24] <mathiaz> kirkland: hm - 4 guests per NCs - that's 8 guests max
[19:24] <mathiaz> kirkland: I'd rather have 8 GB minimum then
[19:25] <mathiaz> kirkland: to max out at 16 guests
[19:26] <kirkland> mathiaz: well you might be able to get away with 256MB guests
[19:26] <kirkland> mathiaz: you'll also need to tweak the eucalyptus.conf to allow for more than one vm per cpu core
[19:27] <mathiaz> kirkland: well - considering that we wanna test the default configuration, I'd rather increase the RAM
[19:27] <kirkland> mathiaz: fair enough
[19:28] <mathiaz> kirkland: more than one vm per cpu core -> does this mean that there should be at least 8 cores on the NCs?
[19:28] <mathiaz> kirkland: to be able to run up to 8 VMs?
[19:28] <kirkland> mathiaz: well, the default configuration is only 1 VM per core
[19:28] <kirkland> mathiaz: so 8 cores would give you up to 8 vm's
[19:29] <mathiaz> kirkland: hm - so the limiting factor here is the number of core, not the amount of RAM?
[19:29] <mathiaz> kirkland: a small guest is 1 core + 512 M of RAM?
[19:33] <kirkland> mathiaz: it's both
[19:34] <kirkland> mathiaz: you know how ec2 has various different machine types?  -t m1.small or -t c1.medium, etc?
[19:34]  * mathiaz nods
[19:34] <kirkland> mathiaz: in the ec2 world, the machine types that start with "m" are "memory" rich machines
[19:34] <kirkland> mathiaz: and the "c" ones are "cpu" rich machines
[19:34] <kirkland> (kind of)
[19:34] <kirkland> mathiaz: but Amazon gets to decide what those machine type/sizes are
[19:35] <kirkland> mathiaz: one of the advantages of running your own cloud is that YOU get to decide how big the machine types are
[19:35] <kirkland> mathiaz: with UEC, you can make any of your machine types any size you want
[19:35] <kirkland> mathiaz: it's in the web interface, on the configuration tab
[19:35] <mathiaz> kirkland: ok - so what are the default machine types?
[19:35] <kirkland> mathiaz: there are defaults, of course, but you can trivially change those
[19:35] <mathiaz> kirkland: ok - so what are the default machine types in UEC?
[19:35] <kirkland> mathiaz: let me check ...
[19:37] <kirkland> mathiaz: http://rookery.canonical.com/~kirkland/screenshot.png
[19:40] <kirkland> mathiaz: those are the defaults; you see how easy it is for an admin to change though ...
[19:41] <mathiaz> kirkland: great - thanks
[19:41] <kirkland> mathiaz: i'm trying to find where the config is to change the allowed vm's per core
[19:48] <RoAkSoAx> are there any lucid server iso's available yet?
[19:49] <kirkland> RoAkSoAx: there was for a while, but they're broken now
[19:49] <kirkland> RoAkSoAx: the last good one I have was from 27 Nov
[19:51] <RoAkSoAx> kirkland, is there a place where I can get it?
[19:51] <RoAkSoAx> since I cannot seem to find any in
[19:51] <RoAkSoAx> the website
[19:52] <majuk> Hey guys. I'm having an issue with networking. My server has 2 1g/s NICs and is connected to a 1g/s switch and a 100mb/s router that routes the rest of the hosts. The switch also connects the main T1 gateway. The problem is, it seems the server is using the router as a gateway as my entire network caps out at 100kb/s, including the server.
[19:54] <kirkland> RoAkSoAx: you can download from me, if you like
[19:54] <majuk> Anyone have any ideas on how to diagnose that this is actually the problem? tracepaths show the server hitting the T1 gateway first, so it looks like it's functioning correctly, but I should be able to get more than 100k/s or at LEAST 100k/s on the server AND through the router to the other hosts.
[19:54] <kirkland> RoAkSoAx: in the mean time, you should bug cjwatson or someone in ubuntu-devel about getting the daily server iso's fixed
[19:55] <RoAkSoAx> kirkland, I'll bug them then
[19:55] <kirkland> RoAkSoAx: cjwatson may be gone for the day, poke slangasek
[19:56] <RoAkSoAx> ok will do :)
[19:57]  * majuk cries
[20:02] <mathiaz> kirkland: seems that 8 cores + 4 RAM is enough max out both RAM and nb of cores
[20:02] <mathiaz> kirkland: about disk space - the capacity defined in the type is how much space is allocated to the guest?
[20:02] <mathiaz> kirkland: does this take into account the cache image?
[20:03] <kirkland> mathiaz: no, it does not
[20:03] <kirkland> mathiaz: eucalyptus can be a pig about disk space, i learned
[20:03] <kirkland> mathiaz: back in the jaunty cycle, i was trying to use a 32GB SSD in one of my NCs
[20:04] <kirkland> mathiaz: i never could get an instance to launch, failed for strange, non-exception-handled reasons
[20:04] <mathiaz> kirkland: right - I remember
[20:04] <mathiaz> kirkland: now how much space is needed by the image cache?
[20:04] <mathiaz> kirkland: I guess it depends on the image
[20:04] <kirkland> right, exactly
[20:04] <kirkland> mathiaz: and how many different images you're talking about
[20:05] <kirkland> mathiaz: each image will be cached once
[20:05] <kirkland> mathiaz: if you only have 1 image, then the requirement isn't very high
[20:05] <kirkland> mathiaz: and you can clear that cache at any time
[20:05] <mathiaz> kirkland: yeah - according to my calculs, I can run up to 8 guests on one NC
[20:05] <kirkland> mathiaz: will just take longer each time you start a non-cached image
[20:05] <mathiaz> kirkland: which means up to 8 different images
[20:06] <mathiaz> kirkland: would 20 Gb / image cache be a good approximation?
[20:06] <kirkland> yeah, that should be safe
[20:06] <mathiaz> kirkland: that would mean at least 160 GB for image caching on each NC
[20:06] <kirkland> mathiaz: our images are pretty small, now
[20:06] <kirkland> mathiaz: like under 1GB compressed
[20:07] <mathiaz> kirkland: what is used on the NC - raw or qcow2?
[20:07] <kirkland> mathiaz: oh, no, you're way over shooting
[20:07] <azteech> majuk, if memory each connection will only be as fast as the slowest connection on your network - because you are using a 100mb/s router - the network most likely will never go faster. you need to upgrade to a router that is capable of running the gigabit speed.
[20:07] <kirkland> mathiaz: the cache is just the bare image itself
[20:07] <mathiaz> kirkland: oh ok - so 5 GB /image cache?
[20:07] <kirkland> mathiaz: let me check my NC
[20:08] <mathiaz> kirkland: isn't the NC cpying the image at some point?
[20:08] <azteech> majuk, meant to say if memory servers me correctly
[20:08] <kirkland> mathiaz: yes, to run the image
[20:08] <kirkland> mathiaz: the cache, though, as I understand it, is just the local, master copy of the image
[20:08] <majuk> Hmmmm.... so you mean since the 100mb/s router is connected through the gig switch, the gig switch will only run at 100mb/s speeds?
[20:08] <mathiaz> kirkland: and that gets copy for each instance
[20:09] <majuk> azteech! ^^
[20:09] <mathiaz> kirkland: and there is also another file allocated for each type of instance (capacity)
[20:09] <kirkland> mathiaz: http://pastebin.ubuntu.com/334084/
[20:10] <kirkland> mathiaz: this node has run 2 different images (emi's)
[20:10] <kirkland> mathiaz: each of which is ~564M cached
[20:10] <azteech> majuk, that is what I am saying.
[20:10] <mathiaz> kirkland: what's in the emi-* directories?
[20:10] <kirkland> mathiaz: after the *instance* is terminated, eucalyptus cleans up the backing disk image
[20:11] <kirkland> mathiaz: -rw-r--r-- 1 eucalyptus eucalyptus 2.1G 2009-11-24 21:30 disk
[20:11] <kirkland> -rw-r--r-- 1 eucalyptus eucalyptus 5.9K 2009-11-24 21:31 disk-digest
[20:11] <kirkland> mathiaz: where those are sparse files
[20:11] <kirkland> mathiaz: do you want access to my NC to poke around?
[20:12] <mathiaz> kirkland: right - and these are m1.small?
[20:12] <mathiaz> kirkland: oh yeah - if possible
[20:12] <mathiaz> kirkland: that would be easier I guess
[20:15] <majuk> azteech! Alright, thanks man
[20:16] <azteech> majuk, yw
[20:28] <kirkland> smoser: yo
[20:28] <kirkland> smoser: mathiaz has a question for you about uec vs. ec2
[20:28] <mathiaz> smoser: yeah - seems that on UEC, there isn't any local storage in instances
[20:29] <mathiaz> smoser: IIRC on EC2 you get some scratch space on /mnt (like 100s of GB)
[20:29] <mathiaz> smoser: it seems that UEC doesn't provide that
[20:29] <smoser> mathiaz, not true
[20:29] <smoser> :)
[20:30] <smoser> hold on
[20:30]  * mathiaz stops breathing
[20:32] <aubre> mathiaz: on my extra large instance I have around 17gb in /mnt
[20:32] <mathiaz> aubre: on EC2?
[20:32] <aubre> doh
[20:33] <aubre> nope
[20:33] <aubre> talking UEC
[20:33] <mathiaz> aubre: is /mnt a separate partition?
[20:33] <mathiaz> aubre: does it use another disk?
[20:33] <mathiaz> aubre: what's the kvm command on the NC?
[20:33] <aubre> mathiaz: it is /dev/sda2
[20:34] <aubre> mathiaz: and I didn't have to do anything to make it happen, it was automatically there when I created the instance
[20:34] <aubre> mathiaz: I am using the 64-bit img from the store btw
[20:34] <mathiaz> aubre: right - extra large gives you 20 Gb by default
[20:35] <aubre> mathiaz: on ec2, doesn't what you put in /mnt go away when you close the image?
[20:36] <mathiaz> aubre: yes - that's correct - it's just scratch space
[20:36] <mathiaz> aubre: I don't seem to have that on my UEC instances though
[20:36] <aubre> mathiaz: what size are you using?
[20:36] <kirkland> mathiaz: you're running c1.medium
[20:36] <kirkland> mathiaz: which is only a 5G instance
[20:36] <mathiaz> kirkland: right - so I see a /dev/sda2 in the instance
[20:36] <mathiaz> kirkland: but it's not mounted by default in the image
[20:37] <aubre> mathiaz: hmm
[20:37] <smoser> ok. i'm back.
[20:37] <aubre> mathiaz: so it may be a function of which image you are using?
[20:37] <mathiaz> kirkland: on the NC side, it's just one big 5Gb file
[20:37] <aubre> mathiaz: so you could just fdisk it and mount it on your own
[20:37] <smoser> euca-describe-availability-zones verbose
[20:37] <mathiaz> aubre: right
[20:37] <smoser> that tells you how much "disk" you have
[20:38] <smoser> if your root filesystem uses all of that "disk" you dont get any more. whatever you dont use goes in /dev/sda2
[20:38] <aubre> mathiaz: try the images from the store, mine automagically mounted /dev/sda2 to /mnt
[20:38] <smoser> thats basically the case.
[20:38] <smoser> it "shoudl work".
[20:39] <mathiaz> smoser: ok - seems like this is what I get indeed
[20:39] <mathiaz> smoser: (except that sda2 is not mounted on /mnt by default)
[20:39] <mathiaz> kirkland: which emi- were you using?
[20:39] <smoser> mathiaz, what is the image ? the released karmic should do that.
[20:39] <smoser> the lucid probably has issues (mountall)
[20:39] <smoser> or could have issues
[20:40] <aubre> smoser: mathiaz my released karmic automounts
[20:40] <mathiaz> smoser: that being said - I thought that on EC2 the scratch space on /mnt was much bigger
[20:40] <smoser> on ec2 it is.
[20:40] <smoser> 160G
[20:40] <mathiaz> smoser: and euca-describe-instances verbose on EC2 would show a disk space of 160?
[20:40] <aubre> smoser: on all instance sizes?
[20:40] <smoser> http://aws.amazon.com/ec2/instance-types/
[20:41] <smoser> euca-describe-instances verbose is a euca specific tool
[20:41] <smoser> since you can configure your eucalyptus sizes
[20:41] <smoser> you cannot do so for ec2 . you get what they say.
[20:41] <mathiaz> smoser: right - gotcha
[20:41] <mathiaz> smoser: so everything looks good to me.
[20:42] <mathiaz> smoser: except that the emi I'm using doesn't mount sda2 - I don't which image is actually used though
[20:43] <smoser> http://docs.amazonwebservices.com/AWSEC2/latest/DeveloperGuide/concepts-amis-and-instances.html#instance-types has real info on what you get on ec2
[20:44] <smoser> mathiaz, the rc karmic images i think had issues with that. i dont really recall, but released karmic shoudl work
[20:44] <smoser> and obviously we need to make lucid work
[20:50] <marks256> when Newegg says that piece of hardware (RAID controller) is supported under "Linux (RedHat, SuSE, Debian, Mandrake, TurboLinux, CentOS, etc.)" does that "etc" include Ubuntu? I believe ubuntu is based on Debian...
[20:59] <ScottK> Ubuntu is based on Debian.
[20:59] <ScottK> The kernels are not the same however, so it's not guaranteed to be close enough.
[20:59] <ScottK> Etc probably would include Ubuntu, but no way to know for sure.
[21:02] <marks256> ScottK, i asked on #ubuntu and everyone who replied said it should work. Thanks anyway :)
[21:16] <MTecknology> I did this 'ufw allow proto tcp from 138.247.0.0/16 to any port 22' and now trying to make an ssh connection to my server isn't working. I did default deny as well.  My IP starts with 138.247
[21:19] <MTecknology> Maybe it's not ufw; I'm getting this error "ssh_exchange_identification: Connection closed by remote host"
[21:22] <MTecknology> nevermind... hosts.deny
[21:46] <unit3> I'm having some weird issues with LVM on one of my servers, and some of the /dev/vg/lv links are missing.
[21:46] <unit3> they exist under /dev/mapper/whatever, but all the references in config files and stuff point to the other paths.
[21:46] <unit3> is there an easy way to get udev to recreate those paths?
[21:49] <Aison> damn, after some uptime, I get an endless lop of these messages: Dec  3 22:47:32 mediaserv kernel: [  387.818430] saa7146: interrupt_hw(): warning: interrupt enabled, but not handled properly.(0xe7fcfbb7)
[21:49] <Aison> both cores are used 100%
[21:49] <Aison> the machine is almost not reacting
[21:49] <unit3> that's a video capture card, right?
[21:49] <Aison> yes
[21:50] <unit3> Sounds like the driver for that is buggy. can you rmmod saa7146?
[21:51] <unit3> and if so, do things even out?
[21:51] <Aison> trying, each keystroke takes 10 seconds ;)
[21:51] <unit3> heheh oh man, that's really messed.
[21:52] <Aison> it's not the first time, it happens all the time ;)
[21:52] <Aison> before with gentoo, now with ubuntu server
[21:52] <unit3> oh man, that's harsh. definitely a bug with that kernel module then.
[21:52] <unit3> File it on launchpad, they'll forward it upstream.
[21:53] <unit3> also, you may wish to load the module with the saa7146_debug parameter set to something > 0. That way, it should log more info about why it's dying to syslog, give you more to report.
[21:53] <Aison> sec... I'm also ask #linuxtv, I know some guys there
[21:53] <unit3> Sure. Really sounds like a driver bug to me though, so I'm not sure there's much they'll be able to do, unless there's a different driver for that card.
[23:05] <Hypnoz> anyone know how to make apt-get install nis silent (prompts for domain name currently), for use during like an unattended pxe install?
[23:09] <billybigrigger> -qq
[23:10] <billybigrigger> Hypnoz, simple apt-get help shows that
[23:10] <billybigrigger> -qq No output except for errors
[23:10] <billybigrigger> -y  Assume Yes to all queries and do not prompt
[23:16] <Hypnoz> -y I believe is only to accept the install, not for the queries the packages prompt during their install
[23:23] <Hypnoz> apt-get -y install nis  didn't work, it still prompted for a domain name
[23:23] <Hypnoz> will try -qq
[23:23] <Hypnoz> still prompts for a domain name
[23:30] <dinger1986> hello does anyone have experience of hylafax?
[23:39] <xperia> hello to all. i have just successful installed red5 on my server and have now a qustion about the protocol rmtp
[23:39] <xperia> if i do in the browser "http://mywebserver.com" it works everything like expected
[23:40] <xperia> but if i do "rmtp://mywebserver.com" i am getting rmtp protocol not registered
[23:40] <xperia> my question is what could be the problem ?
[23:41] <xperia> do i need to put some new lines on my bind name server for this special protocol. i suppose no as i dont has anything to do with domain names or i am wrong ?
[23:44] <unit3> rmtp is a protocol for multimedia apps. Generally web browsers don't speak rmtp.
[23:44] <unit3> what are you trying to do, exactly?
[23:45] <xperia> need to test if my new installed red5 flash media server works good. tested some demos but all freeze my webbrowser or dont do anything till yet
[23:46] <unit3> so you've got some web pages with flash apps that are configured to point at your red5 install, then?
[23:46] <xperia> the port 1935 for rmtp is openend on my router
[23:47] <xperia> till yet i just try only the demos unit3 in the next day i will try to adopt my flash apps to use my new red5 server
[23:47] <unit3> ok, so you've got some demo flash apps in a web page configured to point at your red5 server?
[23:47] <xperia> unit3: yes here is the link http://wificom.ch:5080/
[23:48] <xperia> sorry:  http://wificom.ch:5080/demos
[23:48] <xperia> the bandwidth test application freeze my browser
[23:49] <xperia> the chat application dont do anything. dont know why
[23:49] <xperia> and the port testing application fails !
[23:49] <unit3> right, I see that.
[23:49] <sysErroR> I have openssl installed on my vps, and I compiled an IRCd with ssl support, yet, it can't find openSSL, can anyone help me?
[23:50] <unit3> xperia: this looks like a red5 configuration issue, but I don't know a ton about red5. you'd probably have better luck asking on their support forums or irc channel, since this doesn't seem to be a problem with your actual ubuntu server install.
[23:50] <unit3> I mean, your install is there, it's reachable, and it's complaining about stuff. That seems like a configuration issue with red5.
[23:52] <xperia> unit3: thank you for your helpfull tips. have asked here to veryfie if something is wrong with my server. will join red5 irc chanell but this people arent very helpfull
[23:52] <unit3> yeah, I think your server's fine, this looks like a red5 configuration issue.
[23:53] <unit3> port 1935 is open, so the port tester should work, but it isn't. to me that says there's extra access controls on the red5 side blocking the connection.
[23:53] <unit3> but again, never setup red5 myself, so I'm not really certain.
[23:58] <mathiaz> kane_: hi - is https://wiki.ubuntu.com/UECTesting - Topologies section enough to keep the ball rolling for getting the hardware?