problems with services starting up after a system crash

[Messiah]

Dear all,

If a system crashes, some services won't start. Like sshd, which is not a good thing. I have had this a couple of times, and I am lucky that I have webmin running on that system (which did startup) so I still can remotely do some things on it.

After some looking I saw that this happens because of the file /var/run/sshd.pid that still exists, because the system crashed. After removing that file, and issuing /etc/init.d/sshd start the ssh daemon starts fine.

A couple of questions:
- Is this supposed to be like this? I know it's not good to crash your system, but these things happen.
- How can I avoid this properly? Should I change some init scripts that after mounting the local discs the directory /var/run/ should be emptied? Or is there another option to solve this?
- Is it likely that those changed init scripts could made it to the baselayout package? I think this issue is important.

Thanks in advance,

Yavuz Aydin

[rac]

I think /etc/init.d/bootmisc is supposed to clean /var/run.
_________________
For every higher wall, there is a taller ladder

[Messiah]

eeermm...all-righty then
After spitting bootmisc out I see what you say.

However, this is a server running in a very productional environment
So we have grsec compiled in, and I have read some problems about some services not starting at boot time. I remember that 3 services always fail at boot time on this server, and if I remember correctly, bootmisc is one of them. I thought that it could not harm, but it does.

So an off-topic question...when are these problems with grsec and starting services being solved?

Hmm..this is a dilemma. Should I disable grsec or not?

Does anyone know under which circumstances (e.g. with which options of grsec) these problems occur?