Suggested change to LFS bootscripts

Jack Brown jbrown at kmts.ca
Sat Oct 19 18:57:19 PDT 2002


On Sat, 19 Oct 2002 20:01:41 +0000, DJ Lucas wrote:

> Gerard Beekmans  wrote in message:
> 
>> Another one:
>>
>> if you're using NFS you're killing the network before you had a chance
> to run
>> the mountfs script which umount the NFS shares. By disabling the
> network
>> before unmounting NFS, it'll go haywire and you'd get long delays
>> while NFS needs to timeout before mountfs continues.
>>
>> --
>> Gerard Beekmans
> 
> Yes, I stumbled onto that shortly after I posted here..on the blfs
> thread (wiser ifdown-eth0 script is needed).  I see a logic problem here
> as mountfs will depend on the network being up, unless a script is
> written specifically for network mounts.  But in some cases network
> depends on having not run sendsignals, while mountfs must be run after
> sendsignals.  BLFS is aiming not to change what LFS has set for the
> defaults with the bootscripts on this issue.  Also, at the moment, it is
> only one program that causes a problem, the future is where a problem
> could really lay (where network dependent daemons are lost as
> sendsignals runs before the network should be down).  Again this breaks
> mountfs in only the case of dhcpcd.  I'd say to drop that client, but I
> can't see having only the ISC client as you are forced to build the
> server as well.  At anyrate, this is major OT for this list.  I think
> furthur discussion is needed in blfs-dev with your above notations.  Any
> suggestions would be much appreciated. :-)
> 
> Thanks for the assistance,
> 
> DJ Lucas

Actually you aren't forced to build the server. This is exactly the set up
I use right now.

(Hint: make -C dhcpcd install)

Jack Brown
-- 
Unsubscribe: send email to listar at linuxfromscratch.org
and put 'unsubscribe lfs-dev' in the subject header of the message



More information about the lfs-dev mailing list