[09:10] <kraut> moin
[12:17] <juliux> hi all
[12:19] <winsubu> halo guys, i have to put up a gateway running ubuntu, i want to know can i have two or more default routes (multi homes gateway) in ubuntu?
[12:21] <kraut> winsubu: you didn't understood the meaning of a default route.
[12:21] <kraut> winsubu: you could have severall routs, but only _one_ default route
[12:21] <kraut> oh, sorry
[12:21] <kraut> yes, that's possible, but it's not perfectly working
[12:22] <kraut> the route must shutdown completly to use the redundant route.
[12:22] <kraut> winsubu: you could do for example a ping check over the active default route and if it fails, you could deactivate it and use the other one.
[12:22] <kraut> but remind that this isn't session-transparent.
[12:24] <winsubu> kraut, ok well i have two adsl routers with ip's so i was thinking putting the one router as the default and add an ip alias to the nic and putting the other as it's default route then having squid run on the box, you could connect to IP:3128 -> going via router ip1 | or connecting to squid IP/ALIAS:3128 -> going via router ip 2
[12:24] <winsubu> so i guess it can be done :) man i am impressed with ubuntu,
[12:25] <winsubu> i am using gentoo currently but my head is so sore from all the source compiles and manual kernel compiles,
[12:25] <kraut> winsubu: yes, but round-robbin won't work
[12:25] <kraut> gentoo sucks hell
[12:25] <winsubu> kraut, thats ok, any way i can get a round robin to work with my setup?
[12:25] <kraut> it has only advantages, if you are a programmer and you need the sources
[12:25] <winsubu> hehehe SHHH dont say that too loud
[12:25] <kraut> i don't think a round-robbin would be possible
[12:26] <kraut> because your external IP will flap then
[12:26] <kraut> and your sessions will fail
[12:26] <winsubu> unless i have two config files, do a cron to do a ping every 5 mins, if the ping fails or time out have cron execute a script to rename the net config and do a networking restart?
[12:28] <winsubu> ok thats a great help thanks kraut 
[12:31] <ICU> hmm there is no round-robin in this scenario, is it?
[12:33] <ICU> you could probably do the trick by using a combination of destination/source nat and routing but this would require to have at least a proper routing. i'm not really sure how to handle that
[12:34] <ICU> perhaps you could solve this by using ip only, but i'm not that familiar with the ip interface :P
[12:36] <winsubu> well you could get something like monit to ping the routers every 5 mins and if no response comes back then rename the netowork settings file and restart the network, then you could be using router 1 or router 2 but thats quite manual :P
[12:41] <dthacker> winsubu: are interested in load balancing, failover, or both?
[12:41] <winsubu> yes,
[12:42] <dthacker> :)
[12:50] <winsubu> dthacker, you got a howto?
[12:51] <ICU> hmm oh no need to do that
[12:52] <dthacker> winsubu:  I was trying to think of a way to apply what I know about linux-ha, ldirectord and crossroads to your situation, but I have not found a way to apply these to what you are doing.
[12:52] <ICU> if ping fails you could just delete the old default route and add an new one
[12:53] <dthacker> linux-ha will move a virtual ip back and forth between two linux machines, but it must live on the machines.  Can't run it on routers ;)
[12:55] <dthacker> using the KISS principle (Keep It Simple Silly) then I would follow ICU's suggestion
[01:29] <stiV> hi everyone ... not sure this is the right channel, but maybe someone can help or redirect me somewhere i can ask my question: i have my own debian mirror, (self signed etc...) and some own packages. the packages work fine, except one package has an initscript that starts/stops a daemon. (perl) the problem is when i upgrade the package and the daemon is already running when the script tries to start it or already stopped whe
[01:29] <stiV> the "Hangup" does not come from the initscript - when i do this manually (stopping multiple times, starting multiple times) everything works ok --- if running, it tells me and if already stopped in tells me that too.
[01:30] <stiV> exit status is always 0 (checked by echoing $? right after the commands)
[01:31] <stiV> anyone have a suggestion how i can get my postinst script to ignore this and just go one (everything works if i remove that starting part, but i want it to do that...!)
[01:31] <stiV> "go one" = "go on"
[01:32] <stiV> google is not really helping there
[01:32] <stiV> ah and i tried to remove the "set -e" --> doesn't help
[01:51] <ScottK> stiV: #ubuntu-motu is generally a better channel for packaging related questions (and I think that's what that will amount to).
[02:03] <stiV> thanks ;)
[02:13] <[diablo] > hi guys
[02:18] <[diablo] > http://tinyurl.com/327rgv <--- my swap from kde to gnome... I actually like Gnome for the first time in 7 years
[02:19] <[diablo] > none what so ever
[02:20] <[diablo] > mere conversation
[03:33] <Drazha> hm, for some strange reason when I try to install smth with aptitutde install its asking me for the install CD
[03:33] <Drazha> why????
[03:35] <ScottK> Probably because you have the CD active in your sources.list
[03:36] <dendrobates> Drazha: sudo vim /etc/apt/sources.list
[03:44] <spiekey> hi
[03:44] <spiekey> where can i change the umask value golbally in ubuntu?
[03:44] <spiekey> i changed it in /etc/login.defs, but when i now log in its still 022(i changed it to 007)
[03:49] <dendrobates> spiekey: it is probably being overwritten by /etc/profile.
[03:49] <nealmcb> spiekey: hmmm - googling leads me to a nautilus problem, but this may be old:  http://ubuntuforums.org/archive/index.php/t-130937.html
[03:49] <nealmcb> what desktop do you use?
[03:51] <nealmcb> debian bug 314796
[03:51] <ubotu> Debian bug 314796 in libgnomevfs2-0 "Nautilus doesn't respect ACL when creating new files" [Normal,Fixed]  http://bugs.debian.org/314796
[03:53] <dendrobates> spiekey: Just changing it in /etc/profile should fix your problem.
[03:53] <Drazha> hm
[03:53] <ScottK> Is there anyone here who is familiar with using milters with Sendmail that would have a moment to discuss it?  I'm trying to make sure the Ubuntu package for dkim-filter has a default that's reasonable for both Postfix and Sendmail.
[03:53] <Drazha> I installed this same and identical copy yesterday and did the same thing and never got this message
[03:54] <nealmcb> yeah - hmm - that bug is for a different problem
[03:55] <jdstrand> ScottK: what is your question.  I have used milter, but haven't used dkim-filter.
[03:56] <ScottK> The dkim-filter can listen on a Unix socket or on TCP, e.g. /var/run/dkim-filter/dkim-filter.sock or inet:8891@localhost or inet:12345@192.0.2.1
[03:56] <ScottK> The Debian package defaults to the socket
[03:56] <ScottK> This is problematic for a chrooted Postfix.
[03:57] <ScottK> So I changed it in Ubuntu to use the loopback.
[03:57] <ScottK> No chroot issues that way.
[03:57] <ScottK> My question is, is that a problem for Sendmail?
[03:58] <jdstrand> I have always used the unix socket.  let me pull out my bat book
[03:58] <ScottK> Thanks.
[03:59] <jdstrand> while I am looking-- have you considered a hard link (haven't looked at the postfix packaging...)
[04:00] <ScottK> That would work I think, but would be more complexity.
[04:00] <jdstrand> yeah-- no sweat.  something along the lines of:
[04:01] <jdstrand> INPUT_MAIL_FILTER(`prog', `S=inet:port@localhost')
[04:02] <jdstrand> obviously there is more you can do with that, but it is fully supported and easy enough to adapt the current milter settings for sendmail (if there is one)
[04:03] <jdstrand> basically replace the current `S=/var/run/dkim-filter/dkim-filter.sock' with `S=inet:8891@localhost'
[04:04] <ScottK> Right.
[04:05] <ScottK> OK.  The first upload I did "works", but isn't what I think we should release with.
[04:05] <ScottK> Let me see what I can do with that.  Thanks.
[04:05] <jdstrand> np
[04:06] <jdstrand> just for thoroughness, sendmail can support 'local' 'inet' and 'inet6' ('unix' is a synonym for 'local')
[04:07] <ScottK> Thanks.
[04:09] <jdstrand> I also had a typo in the text I said to replace, but I bet you spotted it
[04:09] <jdstrand> s#`S=/var/run/dkim-filter/dkim-filter.sock'#`S=local:/var/run/dkim-filter/dkim-filter.sock'#
[04:11] <ScottK> I hadn't, but I haven't actually sat down to update the package yet.  I hope I would have then.
[04:12] <ScottK> It's next in the queue after the current bit I'm doing.
[04:12] <spiekey> dendrobates thanks!
[04:53] <Drazha> does anyone know if plesk 8.2 will work on ubuntu 7.04?
[04:59] <dantalizing> Drazha, its certified for 6.06, I would imagine if there are issues, they'd be minor.
[05:02] <Drazha> dantalizing, hmm, so I guess its trial and error then :)
[05:02] <dantalizing> its more fun that way
[05:03] <Drazha> :) I have just about enough of "fun" in my life :)
[05:33] <dendrobates>  bug 118977
[05:33] <ubotu> Launchpad bug 118977 in samba "winbindd will not start do to invalid cache path" [High,New]  https://launchpad.net/bugs/118977
[06:06] <jdstrand> dendrobates: you know, it just occurred to me that you said auth-client-config is in main now-- I was thinking you said 0.4 made it into universe before the freeze
[06:07] <jdstrand> dendrobates: I would like to revise my statement.  I am very, *very* pleased.  :) thanks!
[06:34] <dendrobates> jdstrand: it was approved for main.  It might take some time for it to show up.
[07:40] <bdmurray> dendrobates: bug 130324 looks interesting but I am unsure how to classify / flag it for the server team.
[07:40] <ubotu> Launchpad bug 130324 in Ubuntu "LAMP fails to create necessary files in /var when LVM used" [Undecided,Incomplete]  https://launchpad.net/bugs/130324
[07:44] <dendrobates> bdmurray: make it an apache bug
[07:45] <bdmurray> dendrobates: okay, do you think it is apache or maybe LVM?
[07:51] <dendrobates> bdmurray: I'm not sure but if you want us to look at it apache is better.
[10:56] <joshritger> can anyone tell me if i can use ubuntu server to share two drives that are formated in ntfs and have files on them already?
[10:57] <lcdd> yes, you can read off them
[10:57] <joshritger> lcdd: can I write to them from a networked computer?
[11:00] <mralphabet> joshritger: reading / writing is controlled by the OS that the drives are connected to.  NTFS writing in linux is not not something I would describe as stable.  It makes no difference where the write comes from, it is still the local machine handing the data to the drive to be written.
[11:00] <gamble6x> As far as I can see ntfs-3g is not in aptitude for 6.06 LTS
[11:01] <ScottK> gamble6x: Do you have Universe enabled?
[11:01] <gamble6x> but arguably you could install that and it should work fine as long as you haven't done any NTFS encryption.
[11:01] <joshritger> I am trying to setup a fileserver that will be able to share files with my ubuntu box and my windows box, I know about samba
[11:02] <gamble6x> yes
[11:02] <joshritger> can I share the two drives that are ntfs from my ubuntu server and read and write to them from windows or ubuntu?
[11:04] <gamble6x> joshritger: as I said.  ntfs-3g seems to be "pretty good" at reading/writing NTFS drives.  So if you mount the drives on an ubuntu server with ntfs-3g it should work well enough shared through Samba or whatever you like.
[11:05] <joshritger> ok
[11:05] <gamble6x> however, if you are able, it would be best to backup the data elsewhere and convert the drives to something Linux native.
[11:05] <joshritger> ok, which filesystem do you recommend?
[11:06] <gamble6x> but because ntfs-3g is not in the repos (as best I can tell) you'll have to install it yourself.
[11:06] <joshritger> k
[11:06] <joshritger> which linux native filesystem do you recommend?
[11:07] <ScottK> It was first packaged in Edgy, so no, it's not in Dapper.
[11:07] <joshritger> ok
[11:07] <qman> ext3 is the tried and true basic system...reiserFS is good with lots of small files, and XFS is good with big files
[11:07] <joshritger> ok
[11:07] <ScottK> ReiserFS is also essentially unmaintained.
[11:07] <joshritger> i think i will  go  with ext3
[11:07] <joshritger> I just have to figure out where to back up my stuff too first
[11:08] <joshritger> LOL
[11:08] <gamble6x> ext3 is still my top pick.  Others are slightly faster at moving smaller or larger files depending, but overall ext3 is a nice average speed journalized filesystem that is rock solid.
[11:08] <joshritger> ok, i think i will try that as soon as I get my server up and running
[11:09] <qman> yeah, the performance difference is in my experience not that great, unless you really tune it to a specific application
[11:09] <gamble6x> but I'll freely admit that as an opinion.  Others might tell you riser or something else.
[11:09] <joshritger> I am most familiar with ext3 so I think that is what i will go with
[11:09] <qman> I'd sooner buy faster disks
[11:10] <joshritger> most of my files that are used off of the disks are music files, but some occasional storage also
[11:10] <qman> well, the "small" cutoff for reiserFS being better is around one megabyte
[11:11] <joshritger> ok, so ext3 sounds best for now
[11:11] <qman> so music files aren't going to gain from it, and XFS is going to be about a wash, since they're not really big either
[11:12] <joshritger> thanks for the help, maybe once I actually get it running I can ask some more questions on more tedius matters
[11:12] <qman> if you're having trouble finding extra space to back up to, you might look into resizing partitions, but that's also considered somewhat unstable
[11:13] <joshritger> ok
[11:13] <joshritger> I have enough space I just have to figure out what to move where