[10:25] <stevenshiau> Hi
[10:25] <stevenshiau> I am new to upstart 0.6.3
[10:26] <stevenshiau> I'd like to know if any "service" in /etc/init/ will mount /dev/pts, /dev/shm... (in Ubuntu 9.10)
[10:26] <stevenshiau> I can not find it
[10:26] <stevenshiau> If anybody knows that, please tell me
[10:26] <stevenshiau> thanks
[11:26] <mbiebl> stevenshiau: the mountall binary is responsible for that → /etc/init/mountall.conf
[11:32] <stevenshiau> mbiebl: Thanks
[11:32] <stevenshiau> But I did not see any devpts mode or something similar in /etc/init/mountall.conf
[11:33] <stevenshiau> description     "Mount filesystems on boot"
[11:33] <stevenshiau> start on startup
[11:33] <stevenshiau> expect daemon
[11:33] <stevenshiau> task
[11:33] <stevenshiau> emits virtual-filesystems
[11:33] <stevenshiau> emits local-filesystems
[11:33] <stevenshiau> emits remote-filesystems
[11:33] <stevenshiau> emits all-swaps
[11:33] <stevenshiau> emits all-filesystems
[11:33] <stevenshiau> emits filesystem
[11:33] <stevenshiau> Any one of them is responsible for mounting /dev/pts?
[11:36] <mbiebl> It's in the source code of mountall
[11:36] <mbiebl> apt-get source mountall
[11:36] <stevenshiau> ok
[11:36] <stevenshiau> tkx
[11:36] <stevenshiau> got it
[11:37] <stevenshiau> Actually I am implementing a diskless system based on this
[11:37] <stevenshiau> Trying to figure out how to start the correct "services"
[11:38] <stevenshiau> Haven't figure it out
[11:38] <stevenshiau> Any hint will be appreciated
[11:40] <stevenshiau> The problem now I have is, looks like mountall.conf is run, but there is no /dev/pts, /dev/shm...
[11:44] <stevenshiau> how can I force mountall to mount them automatically?
[11:46] <mbiebl> stevenshiau: you can run the mountall binary by hand
[11:46] <mbiebl> it should mount those file systems automatically
[11:46] <stevenshiau> yes?
[11:47] <stevenshiau> Could you please show me how? Sorry for asking such a question
[11:47] <mbiebl> maybe you get an error message, why it fails for you
[11:48] <mbiebl> Make that "mountall --verbose"
[11:49] <mbiebl> for even more debug output use "mountall --debug"
[11:50] <mbiebl> you can change the mountall.conf job file and add that parameter to the mountall call
[11:50] <stevenshiau> got it
[11:50] <stevenshiau> Let me try here
[11:53] <stevenshiau> mbiebl: looks like mountall will umount the filesystem first, then mount them again?
[11:53] <stevenshiau> Is that possible I can keep the existing ones?
[18:01] <ion> http://lwn.net/Articles/356764/
[18:02] <Keybuk> lol
[18:02] <Keybuk> saw it
[18:25] <ion> Looking at http://lkml.org/lkml/2009/1/17/37 and the others like it (http://stanse.fi.muni.cz/bugs.html), perhaps they should use the GCC __attribute__ cleanup extension to handle the mutex unlocking. :-P
[18:26] <sadmac> "Even if I don't unpull, I don't _ever_ want to see a driver like this 
[18:26] <sadmac> outside the merge window. And dammit, James, you should have realized 
[18:26] <sadmac> that.
[18:27] <sadmac> And dammit, James, I'm a release engineer, not a doctor.
[21:26] <ion> sadmac: Out of curiosity: is RH using or transitioning to Upstart?
[21:26] <ion> RHEL
[21:27] <notting> all existing releases of RHEL use sysvinit; that's obviously not going to change in an update
[21:27] <sadmac> ion: ^^let notting tell you
[21:29] <notting> rhel does get derived from fedora, though.
[22:16] <ion> Nice, http://www.gnu.org/software/gdb/news/reversible.html
[22:28] <sadmac> ion: kind of scary