torkel ( joined #fai.
fai-guy ( joined #fai.
07:16 lazyb0yn3kl: you not only need a correct kernel. an amd64 server will have an amd64 nfsroot
07:16 lazyb0yand tzhat will not work with an i3865 kernel
07:16 lazyb0ybut it should be possible to make an i386 nfsroot on an amd64 system
07:17 lazyb0ythere is a line in the make-fai-nfsroot config that let's you set debootstrap options - i have no experience with amd64, but it should be possible to set the arch there
allee ( left irc: Ping timeout: 480 seconds
allee ( joined #fai.
08:53 stockholmi have a few packages that put data and structure into postgresql during postinstall
08:53 stockholmbut postgresql is diverted during fai install.
08:53 stockholmhow do you guys handle such problems?
08:54 stockholmthose packages fail to install for that reason
09:09 stockholmh01ger, lazyb0y, eartoast_ ?
09:10 h01gerspammer
09:12 h01geranyway, why is it diverted`
09:12 h01ger?
09:16 stockholmpostgres? because it is users start-stop-daemon and that one is replaced by fai?
09:16 stockholmh01ger: i was cooking kaffee latte inbetween
09:16 stockholmand cleaned the trashcan.
09:17 stockholmand the breakfast table
09:17 stockholmbtw: who is comeing to the fai workshop in essen?
09:18 h01gerthen use a firstboot script and install the postgres packages and configs there. (=at the first boot after the fai installation)
09:18 stockholmwhere can i find such a script?
09:19 stockholmi would like to avoid all those errors that others did already :-)
09:19 stockholmh01ger: starting postgres manually during installation is a bad idea?
09:21 h01gerdepends on your situation
09:21 h01geri find firstboot scripts easier and mroe straightforward
09:22 stockholmh01ger: just trying to reinstall the failed packages you mean?
09:26 stockholmh01ger: do you have one?
09:27 h01gerno
09:27 h01geralso because today i mostly use d-i and softupdates
09:27 h01gerand there this problem doesnt exist
oz___ ( joined #fai.
oz__ ( left irc: Read error: Connection reset by peer
10:38 lazyb0ystockholm: fai does explicitly not run services at install time
10:39 lazyb0yif I had to do things that need the services running, I'd probably put them "somewhere" (/root/posgresql-seed.sql?) and invoke a script at first startup
10:40 lazyb0yor, like holger suggests, only do those things after the install, on the first softupdate might also be interesting (no matter if the base install is d-i or something else)
Nick change: oz___ -> oz_
10:47 h01geractually, on the first softupdate is better then on firstboot. less extra logic.
11:38 stockholmh01ger, lazyb0y: now i am thinking of a softupdate at /etc/rc2.d/S99fai-softupdate, if the file /softupdate is there
11:39 stockholmbut isnt that a rather ugly solution? how can i keep packages from being installed at fai-installtime and keep them for a later softupdate?
11:44 h01gerdefine a class SOFTUPDATE, use "PACKAGES install $CLASSNAME" there
12:10 stockholmh01ger: ohhh, shiny!
12:12 stockholmh01ger: and $Classname is CLASS_THAT_NEEDS_POSTRESS_TO_RUN_IN_ORDER_TO_WORK ?
12:15 h01gerdepends on your classes
Action: h01ger would use POSTGRESQL or such
12:21 stockholmrather then POSTGRESS? :-)
12:22 stockholm"install" is the same then "aptitude" in this case, right? /me searches in the fai-guide.pfs
12:24 h01geraptitude should work too
Action: juri_ would use SQLPOSTGRE. ;P
12:27 juri_i think in reverse polish notation. they don't make a drug to cure that yet.
12:27 stockholmjuri_: you have our full compassion!
12:28 juri_only actually, on my CDs, it would end up as "SERVERSQLPOSTGRE" :)
12:51 stockholmthe class SOFTUPDATE is defined during softupdates automatically?
12:56 stockholmnix
12:58 stockholmhow can i detect if it is a softupdate that goes on?
12:59 stockholmeartoast_: ?
Action: juri_ shrugs
13:05 h01gerstockholm, search the bts, i filed a bug about this
13:07 lazyb0ystockholm: "how can i detect if it is a softupdate that goes on?" : FAI_ACTION="softupdate"
13:08 lazyb0yhmm, has anybody an idea what I am dealing with when screen output gets black with an USB install, after boot, and after the "warning beeps" at the begin of the install were to be heard?
13:08 stockholmlazyb0y: ah, right.
13:09 juri_lazyboy: you've tried hitting shift to wake the screen up, right? :)
13:10 lazyb0yjuri_: ;) - but no - :(
14:12 lazyb0yhmm, seems that happens when some modules are being loaded...
14:13 lazyb0yanybody experiences with faiing xseries?
siggg ( joined #fai.
siggg ( left irc:
fai-guy ( left irc: Quit: leaving
16:06 eartoast_stockholm: FAI_ACTION
16:14 stockholmlazyb0y: xseries? sun?
16:15 stockholmlazyb0y: they use a raid adapter that pretends to be a floppy and fails to be installed on
16:15 lazyb0ystockholm: ibm
16:15 h01gerITYM lenovo?
16:15 stockholmah THAT x series
16:16 lazyb0yno, they only sold laptops and notebooks...
16:16 stockholmoh, not THAT x series anyway
16:16 stockholmthere seem to me many of those
16:16 lazyb0yerm, anything wrong with THAT?
16:16 stockholmlazyb0y: no, not at all.
16:17 lazyb0yshould I better run if I just got the order to install these?
16:17 stockholmthere are just many x's in computing
16:17 h01gerlazyboy, anyway, i did, whats the probem?
16:17 stockholmyou could run if you get those sun servers with x in
16:17 h01ger+l
16:18 stockholmor ask me :-)
16:18 stockholmit works here now
16:18 lazyb0y:)
16:18 lazyb0yanyway, grml seems to boot nice, but a fai usb / cd, just after loding modules, just a second before the "I'm gonna install now" warning beep, the screen goes blank
16:19 lazyb0yI guess I'm better off going to get the net install setup up and running...
16:19 lazyb0yI just thought, I could "quickly" test if generall install with example config works nice while getting the final config right.
16:20 lazyb0yh01ger: any horrible stuff to expect here?
16:21 lazyb0yit's x3650
16:21 h01ger3650?
Action: h01ger only knows some xseries modells, but not that one
16:24 lazyb0y:)
16:24 lazyb0ybrand new stuff
16:26 lazyb0yprobably hard to know all those models...
17:02 stockholm
17:04 lazyb0ystockholm: ?
17:33 stockholmlazyb0y: ich denke über essen nach.
17:33 stockholmlazyb0y: wer kommt denn da?
17:33 stockholmoder wird das eine lizenzfragenschlammschlacht?
17:34 h01gerstockholm, fährst du nach .ar?
17:36 stockholmh01ger: nee, co2, geld, zeit
17:36 h01gerah. wo du co2 sagst, erinnere ich mich
17:37 stockholmdoof das.
17:37 h01gernextes jahr kannst du bahn fahn
17:37 stockholmwas war das leben einfach 2004
17:37 stockholmda habe ich darüber garnicht nachgedacht.
17:37 stockholmja. :-)
Action: h01ger schon und es trotzdem gemacht ;)
17:38 stockholmund dieses jahr wieder. :-)
17:39 stockholmich glaube wenn ich das angebot mit debian day und marketing team früher gekriegt hätte hätte ich noch anders drüber gedacht
17:39 h01gerja. für jeden windows rechner, der durch einen linuxrechner ersetzt wird, darf ich 400km fliegen und bleib co2 neutral ;)
17:39 h01gerist es denn schon zu spät?
17:39 stockholmja
17:39 h01gerah
17:40 stockholminzwischen ist die sommerplanung mit allem drum und drann gelaufen.
Action: h01ger nickt
17:41 stockholmsigrids job macht es auch nicht leichter
17:41 stockholmweil die halt immer noch sehr sehr viel arbeitet
19:25 stockholmwhat is the opposite of "softupdate"? :-)
19:28 stockholmi need to define a opposite class for SOFTUPDATE and need a good name
leehartley (~lyric@ joined #fai.
20:35 leehartleyanyone here?
20:35 leehartleyi get Mounting aufs on /root failed: no such device error
20:35 leehartleyi emailed the list, but thought i would see if anyone here knew
leehartley (~lyric@ left irc: Quit: leehartley
istaz ( joined #fai.
istaz_ ( left irc: Ping timeout: 480 seconds
sepski ( joined #fai.
21:56 h01gerstockholm, INSTALL?
21:56 h01gerhi sepski :)
21:56 sepskih01ger,  :)
22:03 stockholmFAI_ACTION="install"; if [ "X$FAI_ACTION" != "Xsoftupdate" ] ; then  echo $FAI_ACTION; echo "softupdate" ; fi
22:03 stockholmFAI_ACTION="install"; if [ "X$FAI_ACTION" != "Xsoftupdate" ] ; then  echo $FAI_ACTION; echo "softupdate" install
22:03 stockholmsoftupdate
22:04 stockholmhu?
22:04 stockholmhow did that paste work out that way?
22:04 stockholmlet me try again:
22:04 stockholmFAI_ACTION="install"; if [ "X$FAI_ACTION" != "Xsoftupdate" ] ; then  echo $FAI_ACTION; echo "softupdate" ; fi
22:04 stockholmand that prints
22:04 stockholminstall
22:05 stockholmsoftupdate
22:05 stockholmand i get exactly the same for == or =
22:05 stockholmthe syntax looks right to me
22:08 stockholmand why does this work?
22:08 stockholmFAI_ACTION="install"; if [ $FAI_ACTION = softupdate ] ; then  echo $FAI_ACTION; echo "softupdate" ; fi
22:09 h01gerstockholm, did you search the bts for my bug about this?
22:09 stockholmi did, and did not find anything
22:09 stockholmi sarched both for softupdate and holger
22:10 stockholmwhat is the bug?
22:10 h01ger
Action: h01ger sighs. 2005
22:12 stockholmhas anyone considered a fork?
22:12 h01gerhai
22:12 h01gerholger automated infrastructure
22:12 h01ger+s
22:12 h01ger:)
22:12 stockholmso?
22:12 h01gerbut thomas is also really good at maintaining fai
22:12 stockholmi guess you would have plenty of fans
22:12 stockholmhe lacks people skills
22:13 h01gerhe is very conservative in changing fai (which is also good), and he puts a _lot_ of time in this
22:14 stockholmwell, if he put less time in it himself and let others contribute more there would be a breal even at some point.
22:14 stockholmbreak even, even
Action: h01ger is too busy and to little bothered
22:15 h01geri can fix this in my configspace i dont even have to run the fork
22:16 h01gers/run the/create a/
22:16 h01gerbut i'd liked fai softupdates capabilities much much more "marketed" ("being done marketing for") = more developed, more advertized
22:17 h01germore emphasized, thats the word
22:17 stockholmsure.
22:17 stockholmnot just that
22:19 h01gerstockholm, svn/fai/branches/people/stockholm/ - please do
22:19 h01gerif you have something interesting and proof its working, thomas is happy to take it
22:20 stockholmsure.
22:21 stockholmsince this is marketing releated it would fall under my jurisdiction :-)
22:21 stockholmhaha
22:21 h01geri should do the same with a fai guide written in a wiki like, which "becomes" - but i dont have the time. to show thomas this is better. which i know it is. but he doesnt believe me until i show it with the fai guide. he doesnt believe it from the debian-edu-doc. i dont have the time to provide this "demo" (or fork) (and also not the need. fai works for me.) but its sti
22:21 h01gerll a pity
22:22 h01gerdebian-edu-doc is now translated to five languages.
22:22 stockholmhe is shoveling fais grave
22:22 h01geri now even have "hope" (i think its possible) to have a more or less update version for lenny _very soon_. becomes distribuuted doc-development in a wiki is fast.
22:22 stockholmits quite obvious
22:22 h01gerno
22:23 h01gerits big in clusters.
22:23 h01gerit is
22:23 h01gerand elsewhere. as said, i can workaround fai's limitations in my configspace easily
22:23 stockholmhaveing a project prosper is about building a community around it.
22:23 stockholmyou drive people away by frustrating them
22:23 stockholmyou yourself used plenty of frustrated words in the last ten minutes
sepski ( left irc: Quit: Leaving
22:25 stockholmso do many others i talk to
22:26 stockholmoh well
22:30 h01geri do think thomas is doing a very good job not always hunting the latest trends. but.
22:30 h01geranway
22:30 h01gerstockholm, have you looked at puppet? many people are using it now, where i always think "i'd do that with fai". but thats because i know fai. i wonder if puppet is better or if its that the development is more open?
22:30 h01gerand probably not only "looked" :)
23:30 stockholmh01ger: i actually did NOT look at puppet because i wanted to use the debian way
23:31 stockholmh01ger: still awake?
23:32 h01gersort of
23:32 stockholmyour [ $ROOTCOMMAND ] patch does not seem to work properly.
23:32 stockholmor i broke something
23:33 stockholmi guess i did, but why do i have a "FAI_" class?
23:52 h01gerdid you replace $ROOTCOMMAND with $ROOTCMD ?
--- Thu Jul 10 2008

Generated by 2.1_h01ger1 by Jeff Waugh - find it at! Customized by Holger 'h01ger' Levsen