Matthew Saltzman wrote:
> The ltrace convinced me that the problem was with the /etc/dhcpd.conf
> file, although that was not the last file accessed. VMware manages a
> special DHCP on its own virtual subnet for the virtual machine. They
> recommend that those already managing DHCP include an entry for the
> new subnet with no "range" parameter. Linuxconf, however, insists
> that the range cannot be empty. I had edited the dhcpd.conf file by
> hand, and when linuxconf tried to use the missing parameter (but *not*
> when it loaded it), it performed a strdup() to an uninitialized
> pointer.
It sounds like you had an incomplete range directive, rather than none
at all. I run DHCP, and have several subnets that have no range
parameter. Linuxconf doesn't segfault.
You don't really need anything other than the definition for the subnet.
Try using:
subnet 192.168.0.0 netmask 255.255.255.0 {}
> Commenting out my changes and reinstalling them with linuxconf
> solved the problem, although the range of the new subnet is not
> empty. I'll be reporting this to VMware. It would be nice if
> linuxconf were a bit more robust to bad config files, however.
I agree. I've seen about a half a dozen bad configs crash linuxconf.
Some more error checking would probably help...
> Thanks for the tip! I'm going to have to try this with the GNOME
> CD player, which crashes my entire system big-time with my Ricoh
> CD-R/W, but that's another post for another day...
The entire system? I don't know if ltrace can help you there.... If
the _system_ is crashing, there's probably something wrong with the
hardware, or the kernel driver. Is that IDE or SCSI? ;)
MSG
--
To unsubscribe: mail [EMAIL PROTECTED] with "unsubscribe"
as the Subject.
