| Title: | ase |
| Moderator: | SMURF::GROSSO |
| Created: | Thu Jul 29 1993 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 2114 |
| Total number of notes: | 7347 |
2 node TruCluster V1.4, UNIX V4.0b with all current patches.
There are 2 large drd services, then an nfs service. The nfs service is
also "mounted" to each server, as this is the Oracle Home area.
Every time we boot up the systems, the first system to come up starts
the ASE as it should, and then the system mount all nfs directories.
Everytime we get the following:
clnttcp_create: RPC: Port mapper failure - RPC Timed out
rpc.statd: cannot talk to statd at oracle_nfs.
oracle_nfs is the nfs service name. Now, the directory finally does get
mounted, but it mounts it as NFS2, not NFS3. There is a later nfs
service that also gets mounted, and it gets mounted okay with no
errors.
Why are we getting this error?
Also, during some intensive nfs reads, we get the following error:
NFS3 RFS_GETATTR failed for server ora_stage: RPC: Timed out
NFS3 RFS_READ failed for server ora_stage: RPC: Timed out
ora_stage is the second nfs service. When the above happens, the
jobs running die. (we also get this sometimes on oracle_nfs).
Reasons/suggestions???
Thanks,
/jim
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 1947.1 | Any ideas?? | DYOSW5::WILDER | Does virtual reality get swapped? | Mon Mar 17 1997 22:02 | 8 |
HELP! Any ideas on the base note?? I logged a CSC call but since it was
only "urgent", I still haven't gotten a call back.
Ideas/suggestions??
Thanks,
/jim
| |||||
| 1947.2 | Possible answer... | JULIET::LOLA_JI | Tue Mar 18 1997 16:56 | 31 | |
Jim,
I'm running up against the same problem with one of my clients. Their
configuration is 2 A/S 2100A's, DU 4.0B, TruCluster Production Server
1.4, with all the patches. All their services are NFS Services
including their Mail.
After a bit of investigation, it looks like the automount binary may be
the problem. Atleast I'm going to test a fix that I had created. The
-p option of automount does not appear to be recognized when I manually
run /sbin/init.d/nfsmount start.
What I have done is to recompile the automount source code after
verifying that the -p option is supported. I'll be testing this out at
my customer's site on Friday to see if it works. I'll let you know by
the following Monday.
If this does fix the problem, I'll QAR this. In the meantime, you may
want to QAR this anyway.
If you wish to contact me directly, send e-mail to :
jim@dwarf.wro.dec.com
I'll update you with my findings.
Jim Lola
UNIX Systems Consultant
Digital NSIS
| |||||
| 1947.3 | I'm not running automount | DYOSW5::WILDER | Does virtual reality get swapped? | Tue Mar 18 1997 22:05 | 9 |
Well, I'm not running automount at this site, so your situation may not
help me, but I'll look at anything.
Anything else? I logged a call with Colorado yesterday at 2:20 p.m. and
still haven't heard from them. They tried once, got a busy signal, and
didn't try again. Guess I'll have to escalate this.
/jim
| |||||