Re: nfs client. Error 111 on mount. ACK!
Peter T. Waltenberg (p.waltenberg@irl.cri.nz)
Tue, 14 Oct 1997 08:50:12 +1300 (NZDT)
On 13-Oct-97 Chris Powell wrote:
> Wellll...
>
> I thought all was well. True, moving the script that mounts remote
> filesystems _after_ the portmapper started was good; it did away with
> those pesky startup errors.
>
> However, it also introduces a new error: system load average heads to
> 1.00 and stays there. It seems to be that a process called "lockd" (0
> resident pages, 0% CPU) is the cause of the high load. It only appears
> on ps if I reorder the startups as described.
>
> I've moved back to the old way. A pause-and-error-sequence at startup is
> far more palatable than a constantly high load average...
>
Yes, I found the same. And did the same.
I applied Bil Hawes sunrpc patches, they don't fix the problem either.
Reading rc.inet2 it seems that the NFS mount would ideally proceed cleanly
even without portmap running. The intent seems to be to allow mounting of
most of the filesystem (including /usr where portmap lives) via NFS before
the daemons load.
Peter
----------------------------------
E-Mail: P.Waltenberg@irl.cri.nz
Date: 14-Oct-97
Time: 08:43:14
----------------------------------