8 minute Upstart delay during LTSP client boot

[James Hunt]

On 13/11/12 14:53, Matthew Wyneken wrote:
> Thanks for responding, Steve!
> 
> Here's what's in mountall.log:
> 
> Connected to Plymouth
> / is local
> /proc is virtual
> /sys is virtual
> /sys/fs/fuse/connections is virtual
> /sys/kernel/debug is virtual
> /sys/kernel/security is virtual
> /dev is virtual
> /dev/pts is virtual
> /tmp is local
> /run is virtual
> /run/lock is virtual
> /run/shm is virtual
> /rofs is local
> /cow is virtual
> mount / [510] exited normally
> mount /proc [511] exited normally
> mount /sys [512] exited normally
> mount /dev [513] exited normally
> mount /dev/pts [514] exited normally
> mount /run [515] exited normally
> mount /rofs [516] exited normally
> mount /cow [517] exited normally
> swap finished
> local 1/3 remote 0/0 virtual 0/11 swap 0/0
> local 1/3 remote 0/0 virtual 1/11 swap 0/0
> local 1/3 remote 0/0 virtual 2/11 swap 0/0
> local 1/3 remote 0/0 virtual 3/11 swap 0/0
> local 1/3 remote 0/0 virtual 4/11 swap 0/0
> local 1/3 remote 0/0 virtual 5/11 swap 0/0
> local 2/3 remote 0/0 virtual 5/11 swap 0/0
> local 2/3 remote 0/0 virtual 6/11 swap 0/0
> mounting /sys/fs/fuse/connections
> mount /sys/fs/fuse/connections [567] exited normally
> local 2/3 remote 0/0 virtual 7/11 swap 0/0
> mounting /sys/kernel/debug
> mount /sys/kernel/debug [568] exited normally
> local 2/3 remote 0/0 virtual 8/11 swap 0/0
> mounting /sys/kernel/security
> mount /sys/kernel/security [569] exited normally
> local 2/3 remote 0/0 virtual 9/11 swap 0/0
> local 3/3 remote 0/0 virtual 9/11 swap 0/0
> mounting /run/lock
> mount /run/lock [570] exited normally
> local 3/3 remote 0/0 virtual 10/11 swap 0/0
> mounting /run/shm
> mount /run/shm [571] exited normally
> virtual finished
> local finished
> remote finished
> All filesystems mounted
> local 3/3 remote 0/0 virtual 11/11 swap 0/0
> 
> --
> Matthew
> 
Can't see anything amiss there, but a delay of the size you're seeing suggests
networking problems to me.

Try booting adding "debug" on the kernel command-line, then taking a look in
/run/initramfs/initramfs.debug.

I haven't used LTSP so the following are mostly generic ideas:

- I presume the problem you're seeing is on the client side? If so, do you see
the same problem for all clients?

- What happens if you point a natty client at the precise server or a precise
client at the lucid server?

- Are there any useful logs on the server side? Can you enable debug mode on the
server?

- Have you tried package sniffing to see if that helps identify where the
problem is?

I took a quick look at nbd-proxy.c as used by the client and it appears to call
sleep() if it detects a connection delay. There is a debug flag for nbd-proxy
which might be useful.

You could also try booting with "init=/bin/sh" then:

mount -oremount,rw /

Followed by mounting the entries in /etc/fstab manually and observing results.

-- 
Kind regards,

James.
--
James Hunt
____________________________________
http://upstart.ubuntu.com/cookbook
http://upstart.ubuntu.com/cookbook/upstart_cookbook.pdf