A common scenario for smaller installations is that all servers reside on the same UPS. This poses a problem when power comes back and all machines boot at the same time. The machines depending on the NFS server will try a mount, get a "connection refused" back, and give up. What I'd like is for the machines to patiently wait for the NFS server to come up, then proceed with the startup. This could of course be extended to other network file systems, but NFS is my particular itch.
Created attachment 144586 [details] nfswait
Ooops... if [ $[ $now - $start ] -gt 120 ]; then should of course be if [ $[ $now - $start ] -gt $RETRANS ]; then
Hm, might be simpler to wire it into netfs rather than a separate script.
Perhaps. The advantage of a separate service is that you can easily turn the function on and off. But a variable somewhere in sysconfig could also do that job.
Any comments on this or something more I'd need to do? Would be nice to have this in standard FC.
I'd rather have a variable in sysconfig somewhere and do the wait in netfs itself.
Created attachment 154340 [details] nfswait.patch Sorry for taking so long. How about this patch?
Aside from the logic seeming backwards (wait iff WAITFORNFS = "no"?), seems ok. Will merge something along those lines post-F7.
The backwards logic is actually intentional. The idea is that the function is on by default, and you turn it off via sysconfig.
Since we're now post-F8, I was hoping to see this feature. Any progress?
re-ping
Putting on tracker. Sorry about the delay.
Adding FutureFeature keyword to RFE's.
Initscripts no longer handle this -> close.