[00:41] <DarkStar1> hello has anyone here managed to install courier mailserver on 12.04? I am having issues with the webadmin interface. not loading
[03:22] <Sprocks> my grub menu stays on screen, can someone help? I changed the timeout to 0 already
[05:40] <nsudo> can i share my samba shares over internet ?? how do i configure my router??
[05:42] <andol> nsudo: The question isn't so much if you can share samba over the Internet as if you really want to share Samba over the Internet...
[05:44] <nsudo> well i need my brother to see some home videos he is miles away! they are around 500 gb andol
[05:46] <andol> nsudo: Making them availible across SFTP seems like a much better option, and if he wants to mount the share he can use something like sshfs for that.
[05:47] <nsudo> and andol what about settings in router?
[05:49] <andol> nsudo: I have really no idea what the current configuration is on your router, nor how to change whatever might be needed on it.
[05:51] <nsudo> and andol can it be done just by port forwarding?
[05:53] <andol> If you are running a standard NAT setup, yes, then "port forwarding" is most likely what you want to use.
[06:09] <nsudo> and andol thx
[06:36] <greenowl> hello
[06:38] <greenowl> is anyone here?
[06:39] <greenowl> names
[06:41] <greenowl> hello?
[06:42] <greenowl> quit
[08:18] <ejv> hello, this evening I upgraded from 10.04 to 12.04 and (among other things) it has run amuck on my software raid (mdadm). It fails to assemble, complaining about "invalid raid superblock magic on /dev/sdXX". This array however was *just working* 2 hours ago under 10.04. How do I salvage this mess without losing data? Thanks.
[08:39] <pentagon> What key ID do you use to verify the md5sums?
[09:16] <pentagon> What key ID do you use to verify the md5sums of ubuntu 12?
[09:45] <pentagon> What key ID do you use to verify the md5sums of ubuntu 12?
[09:45] <pentagon> What key ID do you use to verify the md5sums of ubuntu 12?
[09:49] <DarkStar1> Hi anyone here worked with courier mail servers? I'm trying to migrate a server from one to another
[09:59] <pentagon> What key ID do you use to verify the md5sums of ubuntu 12?
[10:20] <pentagon> What key ID do you use to verify the md5sums of ubuntu 12?
[10:24] <pentagon> What key ID do you use to verify the md5sums of ubuntu 12?
[10:32] <TJ-> pentagon: "gpg --verify MD5SUMS.gpg MD5SUMS" will tell you the Key ID
[10:37] <pentagon> TJ-: what key id do you use?
[10:38] <TJ-> pentagon: The command will tell you!
[10:39] <pentagon> TJ-: its about making sure
[10:39] <pentagon> What do you use.
[10:39] <TJ-> pentagon: That's how you make sure
[10:39] <pentagon> Make sure of what?
[10:39] <pentagon> I don't think anyone understands how things work.
[10:40] <pentagon> Or you are just harassing me more.
[10:40] <pentagon> What are you trying to prove?
[10:41] <pentagon> Intelligence bieng frowned upon causes discomfort?
[10:41] <TJ-> You get the key-ID from the command. You look-up the key on the keyserver. You examine the keys that have signed the key to be sure it's one you're prepared to trust.
[10:41] <pentagon> Ok, you have made your point.
[10:41] <pentagon> I am uncomfortable.
[10:41] <TJ-> https://help.ubuntu.com/community/VerifyIsoHowto
[10:42] <pentagon> Are you a link bot?
[10:42] <pentagon> I asked a focused question.
[10:42] <pentagon> Is #defocus this whole network ?
[10:42] <pentagon> What key ID do you use to verify the md5sums of ubuntu 12?
[10:42] <TJ-> I've just told you
[10:43] <pentagon> is it a hard question ?
[10:43] <pentagon> look in your .gpg folder
[10:43] <pentagon> tell me the KEY ID
[10:43] <pentagon> What key ID do you use to verify the md5sums of ubuntu 12?
[10:43] <pentagon> do you sopeak english?
[10:43] <pentagon> do you understand these words?
[10:43] <pentagon> KEY ID
[10:44] <pentagon> its a short string
[10:44] <pentagon> KEY ID
[10:46] <pentagon> do you know how to check TJ ?
[10:46] <pentagon> look in your .gpg folder
[10:47] <TJ-> pentagon: I've told you how to get the key ID, and how to verify it via the web-of-trust. That is how a security-conscious admin verifies the key - not by asking anonymous IRC chatters which key ID they use.
[10:48] <pentagon> TJ-: dont make assumptions
[10:48] <pentagon> I asked a question because I want to know the answer to my question.
[10:48] <pentagon> Don't play the I know better than you game of the 90's
[10:48] <pentagon> this is 2012 everyone on irc isnt a sucker anymore
[10:48] <pentagon> fool me once shame on you
[10:49] <TJ-> pentagon: I'll let you find other answers seeing as you don't like mine.
[10:49] <pentagon> What key ID do you use to verify the md5sums?
[10:49] <pentagon> Do you see how you havent answered?
[10:49] <pentagon> Are you a robot??
[10:49] <TJ-> !polite | pentagon
[10:49] <pentagon> ARe you not programmed to answer my question?
[10:50] <pentagon> volunteer or not
[10:50] <pentagon> you dont need to ramble on
[10:50] <pentagon> if you dont want to answer MY question then dont
[10:50] <pentagon> dont give me an answer to another question I DID NOT ASK
[10:51] <pentagon> now please let someone else respond if you will not answer
[10:51] <Doonz> !polite | pentagon
[10:52] <pentagon> Anyone else out there? What key ID do you use to verify the md5sums?
[10:57] <pentagon> TJ-: can I give you some advice?
[11:58] <DarkStar1> Hi anyone here worked with courier mail servers?
[12:10] <DarkStar1> Anyone alive here?
[12:11] <dsavi> me, sort of
[12:12] <TJ-> DarkStar1: I used to, but now I use Dovecot
[12:13] <dsavi> TJ-, I know exactly what you mean
[12:14] <DarkStar1> TJ-: why did you switch?
[12:14] <dsavi> it was not an informed decision
[12:15] <DarkStar1> I need to setup a mail server and migrate an existing server by the end of today :(
[12:16] <TJ-> DarkStar1: not reliable enough. too many changes in the code base. config files all over the place. Not so easy to administer for fire-and-forget deployments
[12:17] <TJ-> DarkStar1: Postfix+Dovecot (or +Courier) will do Maildir-style mailboxes
[12:17] <DarkStar1> Guess I'll give dovecot a try then. I've been struggling with this all day
[12:17] <TJ-> DarkStar1: which part is causing you grief?
[12:18] <DarkStar1> Well I thought I had the mysql authentication sorted
[12:19] <DarkStar1> as I was given aa dump of a mailbox but I can't seem to autheticate with it
[12:19] <DarkStar1> TJ-: even the webadmin doesn't work
[12:20] <TJ-> DarkStar1: I recall having 'issues' with mysql auth on courier years ago now; don't recall how I solved it though
[12:21] <DarkStar1> can dovecot authenticate via mysql?
[12:21] <TJ-> DarkStar1: all I can recommend, is pay close attention to the log-files, turn on maximum debug logging, and use Google copiously to understand things!
[12:22] <DarkStar1> TJ-:  I have been doing thsat for 24hrs+ now and still not moved forward :)
[12:22] <TJ-> DarkStar1: Yes
[12:22] <TJ-> DarkStar1: ouch :( Rings bells for me, that does!
[12:23] <DarkStar1> Awesome!! DO you have a decent doc for setting this up I can follow. I don't have many hows of the day left to do this. and I still have to learn how to use the imapsync tool I bought earlier
[12:23] <DarkStar1> many hours* not hows :)
[12:24] <TJ-> DarkStar1: will this help? http://wiki.dovecot.org/HowTo/DovecotLDAPostfixAdminMySQL
[12:25] <TJ-> DarkStar1: There's also this Linode guide for Ubuntu (based on 10.04 but its almost identical for 12.04) http://library.linode.com/email/postfix/dovecot-mysql-ubuntu-10.04-lucid
[12:26] <DarkStar1> Cheers dude/dudette. I'll get installing right now
[12:37] <DarkStar1> TJ-: what's the quickest way to test if the imap server is running? I haven't allocted a name for the server yet so it's only via ip
[12:38] <TJ-> telnet to its port
[12:41] <DarkStar1> TJ-: thx
[13:00] <samba35> when i am trying to upgrade my system i am gettting lot of error
[13:00] <dsavi> can someone help me setting up smtp with postfix?
[13:00] <samba35> E: Sub-process /usr/bin/dpkg returned an error code (1)
[13:00] <samba35> this is last line of error
[13:00] <samba35> how do i clear apt cache and rebuild apt database
[13:00] <RoyK> try apt-get install -f or dpkg --configure -a
[13:01] <RoyK> apt-get clean && apt-get update
[13:02] <samba35> ok
[13:04] <samba35> hmm same error ,is it becase of i am behind a firewal ? but i am able to update other ubuntu update but some package are still giveing same error -inface some days back i did update for 100+ mb and it work
[13:05] <dsavi> can someone help me with this? telnet mydomain.com smtp works from my local internet connection, but it refuses connection requests from other places
[13:06] <dsavi> the port is open according to nmap
[13:08] <TJ-> dsavi: what do the mail logs show? (/var/log/mail*)
[13:15] <dsavi> TJ-, successful connection: postfix/smtpd[23208]: connect from <redacted>[<ip address>]
[13:15] <dsavi> TJ-, but the unsuccessful connections don't show up
[13:17] <TJ-> dsavi: So they're being rebuffed by the firewall maybe?
[13:17] <dsavi> TJ-, it's possible, I don't know how to check that
[13:18] <dsavi> TJ-, I made an iptables rule to accept traffic on *:25, although it was a copied/pasted command so I don't know how well it worked
[13:19] <TJ-> dsavi: Are you, by any chance, trying to connect to port 25 of a consumer ISP? Many ISPs block port 25 incoming (and some, outgoing) to prevent open spam relays
[13:20] <dsavi> TJ-, I'm telneting from one of my vpses to an linode instance where postfix is running
[13:20] <dsavi> and that's unsuccessful
[13:23] <TJ-> Linode? Well they don't block. To be clear, you've confirmed that postfix will accept external connections?
[13:24] <TJ-> dsavi: What I'm getting at is, where are the 'other places' you talked about that fail to connect? Could they be blocked for outgoing port 25 by ISPs?
[13:25] <dsavi> TJ-, I tested from two VPSes, one at dreamhost and the other from godaddy or something
[13:26] <dsavi> TJ-, and then from my mobile broadband connection which I'm currently using, also the only one that made a successful connection
[13:26] <TJ-> dsavi: The way I'd test it is to set up a simple tcp server on a different port than 25, and see if you can telnet onto it instead
[13:27] <TJ-> dsavi: This post shows how to use netcat for that: http://ubuntuforums.org/showthread.php?t=828870
[13:28] <dsavi> TJ-, would dovecot count? telnet mydomain.com pop3 works from everywhere
[13:28] <TJ-> dsavi: Yes
[13:29] <TJ-> dsavi: Sounds like port 25 being blocked somewhere then. Try accessing another well-known SMTP server from Dreamhost/Godaddy to see if they are blocked
[13:29] <dsavi> TJ-, could it have anything to do with what mynetworks is set to in postfix/main.cf?
[13:29] <dsavi> I don't know what that config option does at all
[13:31] <TJ-> dsavi: Yes. But if that was responsible you'd see log entries in mail.error or similar log-file
[13:47] <DarkStar1> TJ-: did you setup dovecot with mysql?
[13:47] <DarkStar1> sorry to interrupt. I am a bit pushed for time. :(
[13:49] <TJ-> DarkStar1: Only for testing - I use Unix accounts
[13:50] <DarkStar1> TJ-: were your tests successful?
[13:50] <TJ-> DarkStar1: yes
[13:51] <DarkStar1> Cool. Can you remember or give hints as the which were the critical files to modify to get it working?
[13:51] <TJ-> I can't sorry. I know I didn't have any issues doing it - it was all straightforward
[13:51] <DarkStar1> I was looking at the sql-conf file and I'm a little confused with the differences between some query sections
[13:53] <DarkStar1> Aah crap can someone help me pls. it seems I have a runaway telnet process
[13:54] <DarkStar1> I was just trying to telnet into the imap port and couldn't so when I grep'ed the process I realise that I couldn't kill it.
[13:54] <DarkStar1> every time i grep it (ps -ef | grep telnet) I find it has a different pid
[13:55] <DarkStar1> how can I kill it?
[13:55] <TJ-> DarkStar1: £killall telnet"
[13:55] <TJ-> DarkStar1: "killall telnet"
[13:57] <DarkStar1> doesn't work
[13:57] <DarkStar1> I get a no process found
[13:58] <TJ-> DarkStar1: Is a shell process re-spawning it? If so, kill the parent process (the PPID is the column after the PID in "ps -efly" output)
[13:59] <TJ-> DarkStar1: Are you trying to kill the telnetd service by any chance?
[13:59] <TJ-> that's not the 'telnet' client of course
[14:00] <DarkStar1> TJ-: was trying to kill the client
[14:01] <DarkStar1> I can't nor it seems can I kill the parent
[14:01] <DarkStar1> How can I stop the service?
[14:01] <TJ-> DarkStar1: Well you must have something set to respawn it if you're seeing lots of different PIDs for the telnet client
[14:01] <TJ-> DarkStar1: "sudo service <service-name> stop"
[14:03] <DarkStar1> didn't work. doesn't recognise telnet
[14:03] <TJ-> DarkStar1: Well no, because the service is a daemon, so it'd be "telnetd" if anything. But that won't stop your telnet client running
[14:04] <DarkStar1> but it should stop a respawn right?
[14:05] <DarkStar1> TJ-: not telnetd either
[14:05] <TJ-> DarkStar1: nope. Services are set to respawn if they stop, for obvious reasons. That doesn't affect any clients for the protocols those services answer to locally
[14:08] <TJ-> DarkStar1: going back to the beginning - when you used the 'telnet' client on the DreamHost/GoDaddy VPSes to test the SMTP connection to Linode, what was the *exact* command-line you used to start telnet?
[14:11] <DarkStar1> TJ-:  I did a "telnet localhost imap2"
[14:12] <dsavi> TJ-, I think you're thinking of me not DarkStar1
[14:13] <DarkStar1> then in the interactive mode (whilst trying to test the courier imap server) typed a 'AB LOGIN "user" "secret"
[14:13] <DarkStar1> then killed the process by pressing a ctrl+c
[14:13] <DarkStar1> because I had to do something but couldn't exit telnet for some reason
[14:14] <TJ-> dsavi: I admit things are confusing! I got your SMPT issue confused with DarkStar1's telnet issue!
[14:14] <DarkStar1> :)
[14:14] <dsavi> TJ-, Haha, well, that happens :)
[14:14] <TJ-> Can't you both just get together and merge the working parts of your mail servers? :P
[14:15] <dsavi> Haha :D
[14:15] <DarkStar1> sure
[14:15] <dsavi> The ironic thing is that I'm setting this up so I can receive *ONE* email from Amazon SES so I can migrate our email delivery to them
[14:16] <DarkStar1> I only need telnet to test that the imap server is up and running anyway. The bigger issue is setting up dovecot to which I need help really badly
[14:16] <DarkStar1> for which*
[14:17] <TJ-> DarkStar1: the mysql parts? The only thing I recall is, I had to ensure that the query was selecting the correct columns from the correct table in the correct database - all pretty obvious
[14:18] <TJ-> dsavi: On your problem of SMTP not connecting - I was wondering if you were trying to use 'telnet' from Dreamhost/GoDaddy to Linode, and if so, if you were doing "telnet <mailserver-ip> 25" ?
[14:19] <dsavi> TJ-, I did telnet mydomain.com smtp
[14:19] <dsavi> I'll try 25
[14:20] <dsavi> TJ-, 25 and smtp have the same results
[14:20] <TJ-> dsavi: That was my point - making sure you were specifying the destination port :)
[14:21] <dsavi> TJ-, but really, this telnet thing is irrelevant, my real issue is that my domain doesn't receive email
[14:21] <dsavi> TJ-, I try to send email to one address, and I get this error from GMail: Google tried to deliver your message, but it was rejected by the recipient domain. The error that the other server returned was: 553 553 Unknown recipient. (state 13).
[14:22] <TJ-> dsavi: Have you checked the entire path, starting with the MX records in DNS?
[14:22] <dsavi> TJ-, here's where the problem begins, I don't currently have access to change the MX records
[14:23] <TJ-> dsavi: It could be that the Google mail-servers are using a different mail-exchanger than the one you're expecting. DNS zone changes can take time to propagate
[14:23] <TJ-> dsavi: Does the current MX for the domain match the server you're working on?
[14:25] <dsavi> TJ-, it doesn't, but since the linode guide said nothing about changing the mx records, I thought that the server specified in the mx records might be pre configured correctly, and that my server was the one with bad config
[14:27] <TJ-> dsavi: The MX record is the authority on where mail for the domain is delivered. All MTAs look up that when determining where to deliver email to. You can check it using "dig -t MX <domain.tld>"
[14:29] <TJ-> dsavi: (do the 'dig' from another system, not the server in question nor the server hosting the DNS zone!)
[14:29] <DarkStar1> Yeah I thought I had done that correctlyI think my system is too unstable. what's the comand for restart. shutdown -r now doesn't seem to work properly
[14:30] <DarkStar1> doesn't seem means it's taking a long time to restart
[14:30] <dsavi> TJ-, yep, did it and the server doesn't match, but I suspect the domain is the one used by linode to redirect mail to their instances
[14:30] <TJ-> DarkStar1: Sometimes it will if processes are 'stuck'
[14:30] <DarkStar1> + I get errors when I sudo to root
[14:31] <DarkStar1> TJ-: it's been like that for a long time now. On two separeat installations
[14:32] <TJ-> DarkStar1: Some hardware doesn't properly handle the shutdown signal;
[14:33] <TJ-> dsavi: The only way I can think for you to receive the email, is to create an account on the mail server currently configured in the MX record for the domain, so the server will accept the username@
[14:34] <dsavi> TJ-, Okay
[14:35] <dsavi> I'll have to get access to the linode web backend
[14:35] <dsavi> :/
[14:35] <TJ-> huh?
[14:35] <TJ-> dsavi: You mean the email is to an @linode.com username?
[14:36] <dsavi> TJ-, no, it's account@mydomain.com
[14:36] <dsavi> But I don't have access to the linode control panel where I could change the mx records for that domain
[14:37] <dsavi> I'm the server admin and web dev but I haven't needed access to that until now
[14:37] <TJ-> dsavi: Ahhh, .... drat?
[14:37] <dsavi> TJ-, Yeah, I'll just have to get my boss to get me access or change the mx records or whatever
[14:40] <TJ-> dsavi: You could set up a local redirect if all you're doing is testing it. Assuming the mail server is known as "mail.mydomain.tld" you'd add to "/etc/hosts" the line "<mail-server-ip> mail.mydomain.tld" - that way, that system would use '<mail-server-ip>' instead of using a DNS MX lookup
[14:41] <dsavi> TJ-, All I need is to receive this one email from Amazon web services, so I don't think a local redirect would work
[14:42] <TJ-> dsavi: Ahhh... can you change the email address the AWS sends to?
[14:43] <dsavi> TJ-, Here's what I'm trying to do: I want to migrate our email delivery to Amazon Simple Email Service, but to do it, you need to confirm that you control an email address by having a confirmation email sent to it
[14:44] <TJ-> dsavi: Ahhhh!!!! OK, so there's no way around it then
[14:45] <dsavi> TJ-, Yeah, email delivery was terrible on our old sendmail setup, got rejected by other servers seemingly at random, so I decided to migrate to Amazon SES or Sendgrid (We chose SES)
[14:45] <dsavi> So you see my dilemma :P
[14:46] <TJ-> dsavi: It's chicken-and-egg. Most delivery problems are due to not having correct SPF type records on your DNS that match your mail servers
[14:47] <dsavi> Yeah
[14:47] <dsavi> TJ-, Anyway, I need to go now, I think I know what to do next (Change mx records), so thanks a lot for your help :)
[14:47] <TJ-> Good luck!
[14:48] <dsavi> :)
[16:41] <DarkStar1> Which package installs the auth-sql.conf.ext? because it's missing in the conf.d directory
[16:41] <DarkStar1> for dovecot
[17:06] <edgy> # dpkg -l linux-headers-`uname -r`
[17:06] <edgy> No packages found matching linux-headers-3.2.0-31-generic.
[17:06] <edgy> why there is no headers in the server but there is in the desktop?
[17:12] <qman__> edgy, that just means it's never been installed on your system
[17:12] <qman__> apt-cache search linux-headers-`uname -r`
[17:22] <edgy> qman__: but why it's not installed by default in the server but installed by default in the desktop? why is it needed?
[17:23] <qman__> edgy, those are part of build tools, which are not included in either by default
[17:23] <qman__> they may have been added if you installed a driver or software that requires them
[17:23] <qman__> out of the box, compilers and associated things to compile software (headers and dev packages) are not installed
[17:24] <edgy> $ apt-cache show linux-headers-`uname -r` |grep Task
[17:24] <edgy> Task: ubuntu-desktop, ubuntu-usb, edubuntu-desktop, edubuntu-usb, xubuntu-desktop, mythbuntu-frontend, mythbuntu-desktop, mythbuntu-backend-slave, mythbuntu-backend-master, lubuntu-desktop
[17:24] <edgy> qman__: doesn't this mean it's installed by default with ubuntu-desktop?
[17:24] <qman__> possibly, it never used to be, but jockey might drag it in
[17:25] <qman__> or some other software that requires compiling at install
[17:27] <edgy> qman__: apt-cache show ubuntu-desktop would tell you that it's recommended so it's not that jockey drag it, it's really installed by default
[17:27] <qman__> oh, that's why
[17:27] <qman__> recommend
[17:27] <qman__> recommends only became installed by default a couple years ago
[17:28] <qman__> but, that is why it gets installed, so that things like DKMS and jockey can compile stuff when you ask them to
[17:28] <qman__> server doesn't have those things out of the box and as such doesn't need headers either
[17:29] <edgy> reasonable
[17:31] <edgy> though now I found it's strange that the headers are recommended for ubuntu-desktop but not kubuntu-desktop
[17:33] <qman__> does seem strange, but who knows
[17:34] <edgy> qman__: now the important issue for me is this
[17:34] <edgy> I am getting "error: physical volume pv0 not found." with grub updates
[17:34] <edgy> what could be wrong please?
[17:36] <edgy> there is a bug at https://bugs.launchpad.net/ubuntu/+source/grub2/+bug/1014159 but no response!
[17:38] <qman__> edgy, unfortunately I don't know enough about it to know why that is happening, but it appears that you don't have a volume called pv0, and it's looking for one anyway
[17:38] <qman__> I'd check out the grub configs and see if anything refers to that
[17:40] <edgy> qman__: good hint, let me check
[17:55] <edgy> don't knoq qman__ but when I do grep -r pv0 on /etc/lvm/archive/ there are entries for it
[18:06] <AaronMickDee> I installed ubuntu server on an old IDE drive. I had an extra SATA drive, so I 'dd if=/dev/sda of=/dev/sdb'. How do I update grub2 to know to boot from /dev/sdb instead of the ide drive?
[18:10] <edgy> AaronMickDee: aren't you goint to remove the IDE disk?
[18:13] <edgy> AaronMickDee: I think you just remove the IDE and put the SATA and it would work, if not, run update-grub and it would generate the correct grub.cfg file
[18:14] <edgy> AaronMickDee: if you want to keep both may be you just need to change hd0 to hd1 in the config file and try
[18:28] <qman__> AaronMickDee, sudo grub-install /dev/sdb
[18:28] <qman__> then remove the old disk and you're set
[18:45] <AaronMickDee> qman: Thanks!
[18:49] <AlanBell> hi all
[18:50] <AlanBell> I am contemplating a server upgrade from 10.04 to 12.04 and the server is running KVM with a bunch of assorted VMs on it
[18:50] <AlanBell> if I pause all the VMs, do a do-release-upgrade and unpause the VMs will they just cheerfully carry on? Anyone tested that?
[18:51] <qman__> sounds dangerous
[18:53] <AlanBell> does doesn't it :)
[18:54] <AlanBell> can you run kvm in an ec2 t1.micro?
[18:54] <TJ-> AlanBell: I can't think of anything that would knock them out. Do you have any special external devices with virtio ?
[18:56] <AlanBell> TJ-: no, it is pretty standard local drives on the host
[18:56] <AlanBell> it kind of sounds dangerous, but the more I think about it the more I think it should work
[18:56] <TJ-> AlanBell: Will you be changing the kernel version?
[18:56] <AlanBell> sure, 10.04 to 12.04
[18:57] <AlanBell> running the standard server kernel
[18:57] <TJ-> Which AMIs from and to?
[18:57] <AlanBell> not doing it on Amazon, doing it on real hardware
[18:58] <AlanBell> the ec2 question was because I was thinking of doing a dry run on ec2
[18:58] <TJ-> AlanBell: oh OK - your question made me think it was EC2
[18:58] <AlanBell> I don't have the vt capable hardware kicking about to trial this on
[18:59] <TJ-> AlanBell: Do you have/can you create enough spare space to install a 12.04 server in parallel, and test it? If you're using LVM for example, that would be trivial to do.
[19:00] <TJ-> AlanBell: You could even install the 12.04 into a VM on a bare metal/LVM partition to reduce down-time to a bare minimum
[19:02] <AlanBell> plenty of space, not quite sure I understand that scenario though
[19:02] <AlanBell> you mean dual boot 12.04?
[19:02] <TJ-> AlanBell: I was thinking of minimising the time to install a dual-boot for the purposes of testing 12.04 on the bare metal, if keeping the VMs operational is critical
[19:03] <TJ-> AlanBell: If you can afford for the VMs to be offline then you can do it the other way - from CD/USB live to the new partition/volume
[19:03] <AlanBell> oh, right, yeah, keeping them operational is kind of critical, but a bit of downtime is OK
[19:04] <AlanBell> so you are thinking of doing a fresh install of the host and moving the VMs across
[19:04] <AlanBell> rather than a do-release-upgrade
[19:05] <qman__> that's what I would do, set up a new server, migrate the VMs
[19:05] <qman__> and that way if it doesn't work you still have the old one
[19:05] <TJ-> AlanBell: That's how I usually do it. Create an LV, use it as the device for a VM, install into it, then "update-grub" on the host to add it to the boot menu, then finally reboot the host to test it. Any problems - boot back into the regular OS and read the VMs logs to figure out what went wrong
[19:06] <TJ-> AlanBell: On my systems everything is LVM - if you have the same option, it makes the move relatively easy to back out of instantly
[19:08] <TJ-> AlanBell: The VMs don't require much from the host OS so you can do that part of an upgrade test relatively fast. 2nd stage would be a dist-upgrade of the host (assuming you have complex configs that need to be migrated) or else use apt-mark on the host to generate and copy over the list of installed packages to the new 12.04 install.
[19:10] <AlanBell> yeah, networking config is a bit complex
[19:11] <TJ-> AlanBell: Then dist-upgrade is probably the final solution, but this allows you to dip your toes in the water so-to-speak, without ending up falling in the river!
[19:11] <AlanBell> don't have easy access to the grub menu (remote server, can restart remotely console stuff is a pita)
[19:11] <TJ-> AlanBell: Well that's not insurmountable - I've had that on occasion too. You simply make sure you pre-edit /boot/grub/grub.cfg and set which OS entry should be the default boot entry.
[19:12] <AlanBell> yeah, was thinking through that
[19:12] <AlanBell> little bit scary, but should work
[19:12] <RoyK> not too scary ;)
[19:13] <TJ-> AlanBell: The other thing is - you might want to check this - but I seem to recall that there was some work underway recently in KVM to support guest-in-guest. If it has landed for kernel 3.2 and the kvm/qemu in 12.04, you could run your VMs from inside a VM running the new 12.04 :s
[19:13] <AlanBell> yeah, someone told me that earlier, I think I have a chicken and egg problem with that though
[19:13] <TJ-> AlanBell: So you'd have 10.04(host) > (12.04(guest/host) > VMs
[19:14] <AlanBell> I think the outer one has to be 12.04 or it doesn't pass the VT flag inwards
[19:14] <AlanBell> however I might be able to do that on my laptop
[19:15] <TJ-> AlanBell: Well, you can install the 3.2 kernel on its own to the current 10.04 install, and boot from that. I've done that a few times although I didn't test extensively that *every* possible service still works.
[19:17] <TJ-> AlanBell: On my systems I have a separate /boot/ that shares kernels and grub across multiple OS versions, so its pretty easy for me to do that.
[19:18] <AlanBell> yeah, with these I went for a completely default install
[19:19] <TJ-> 12.04 installer has the option to do complete LVM now
[19:19] <TJ-> Might be worth migrating to an LVM arrangement at the same time, for future-proofing
[19:24]  * RoyK is also running lucid on a kvm host and is considering an upgrade
[19:24] <RoyK> or a bloody reinstall to keep things simple
[19:26] <AlanBell> RoyK: I think I am going to try the brave upgrade, and file bugs if it doesn't work
[19:28]  * TJ- hands out the grey hairs :p
[19:32]  * DarkStar1 grabs a few
[19:34]  * AlanBell tries try.cloud.ubuntu.com
[19:35] <AlanBell> if that will run KVM then I can make a 10.04 amazon ami, install kvm and put a vm on it, then upgrade the host and see what happens
[19:36] <AlanBell> http://paste.ubuntu.com/1252509/ promising
[19:43] <AlanBell> well I have an hour to play with this thing, it starts running 12.04, lets install kvm, kick off a quantal #desktop install in a vm, then upgrade the host to quantal
[19:59] <RoyK> AlanBell: an upgrade will most probably work, I've done that a dozen or so times lucid->precise
[20:00] <AlanBell> yeah, so have I on desktops, just not with kvm running
[20:00] <RoyK> haven't tried to upgrade kvm hosts yet
[20:01] <RoyK> that is, I tried once 12.04 was released, and that failed, had to reinstall
[20:01] <RoyK> but not kvm-related
[20:01] <AlanBell> there is a real nasty in the release notes, looks like all the KVM vms vanish
[20:01] <RoyK> where?
[20:02] <AlanBell> bug 931350
[20:03] <AlanBell> kind of a huge issue really, in terms of the likely laundry bills it could cause
[20:05] <RoyK> heh
[20:05] <RoyK> flagged "medium"
[20:05] <RoyK> that's rather hilarious
[20:05] <AlanBell> and "fixed" in release notes
[20:06] <TJ-> Amazing it hasn't bee fixed. It's not hard to set-up a Lucid bare metal and test VMM to find out what causes the duplicate entry (to fix that), or to create a simply .postinst script the runs the config through 'sed'
[20:06] <RoyK> "it's a bug, but it's fixed once we tell people it won't work"?
[20:06] <TJ-> One of the downsides to a fixed 6-month release cycle - too many bugs and not enough bodies
[20:08] <RoyK> TJ-: these are LTS releases, not your average 6-month releases
[20:08] <AlanBell> indeed, this one is rather a minor manual fix to resolve the problem
[20:08] <TJ-> Makes no difference... the devs are still in the 6-month release cycle
[20:09] <AlanBell> just rather a distressing situation to discover
[20:09] <TJ-> AlanBell: At least there was a release-note! You didn't have to suffer the effect and have to find out why
[20:09] <AlanBell> TJ-: very much so, and a good reminder to always read the release notes it is too!
[20:10] <TJ-> AlanBell: Yes. It's an underrated pastime
[20:11] <tomson> Hi guys. How to create user and folder for him, /home/username. 2. how to limit him just to this particular folder? Any help would be appreciated.
[20:28] <AlanBell> ok, my testing on my hour of Amazon reveals that virt-manager gets disconnected during upgrades and kvm pausing is not the same as suspending, don't think you can pause a vm and reboot and restart it
[20:29] <AlanBell> best thing to do would be to do a clean shutdown of the guests, do the upgrade and fire them up again
[20:29] <AlanBell> and fix the xml if they vanish
[20:29] <AlanBell> nice job for next weekend I think
[23:25] <Wyleyrabbit> Hi everyone
[23:25] <Wyleyrabbit> need some help. I just upgraded my Ubuntu 10.04 LTS server to 12.04 LTS, and everything is great...except I can't get php to talk to a database. Also, I'm getting errors in /var/log/apach2/error.log - mod_fcgid segfaults.
[23:27] <Wyleyrabbit> is this the right channel to be in for this?
[23:30] <Wyleyrabbit> hello?
[23:31] <clunkychicken> hi