| 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 |
Have a customer running UNIX 4.0b and ASE 1.4 on a 4100 system. He has
several nfs services under ASE and finds that the 'quota' command is
seeing 40 second delays each time it tries to touch an ASE nfs service
file.
Thats to say, if in the /etc/fstab file he has a single entry for a
nfs mounted filesystem that is managed by ASE 1.4, he sees a 40 second
delay in the response of the 'quota' command. If there are 2 of these
ASE managed filesystems in the /etc/fstab file, he sees a 80 second
delay in the compleation of the quota command. The filesystems have
the ADVFS type of structure to them.
If the customer exports a ADVFS filesystem in the normal nfs maner (not
controled by ASE) he sees no delays in the quota commnad compleation.
I have been trying to duplicate this problem but the systems running
ASE here at the center have all been blown out of the water by
previous trouble shooting test and I am spending more time trying to
fix these systems that trouble shooting MY problem.
Any ideas?????
Jim Kiley
Atlanta CSC UNIX support
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 1902.1 | ipmt of ASE lockup of rquotad | RHETT::KILEY | Wed Feb 26 1997 19:41 | 28 | |
We have further proven that the "rquotad" process locks up and becomes
inaccessible on this customers site but only on systems with ASE.
If you do a ps looking for rquotad you get the following.
core2% ps alxw | grep rquotad
0 14716 1539 0 42 0 1.66M 176K socket S ??
0:00.51 rpc.rquotad
0 14778 1539 0 44 0 1.66M 176K socket S ??
0:00.43 rpc.rquotad
350 16270 15614 0 44 0 1.57M 136K 125f9094 S + ttyp4
0:00.00 grep rquotad
notice that there are 2 rquotad's running and neither of the will
respond to an rpcinfo commnad. SEE FOLLOWING
core2% rpcinfo -u core2 100011
rpcinfo: RPC: Timed out
program 100011 version 0 is not available
Why and how is ASE locking up the rquotad?
Jim Kiley
Atlanta CSC UNIX support
PS. I'll ipmt this one!!
| |||||
| 1902.2 | reproduced at engineering | RHETT::KILEY | Tue Mar 11 1997 12:27 | 11 | |
This problem has been reproduced by engineering.
It was speculated that the problem may be caused by a problem with rpc
not correctly handeling long path names, however, that has yet to be
proven.
no patch is available yet.
Jim Kiley
Atlanta CSC
| |||||
| 1902.3 | quick update - patch is coming | dust.zk3.dec.com::Marshall | Rob Marshall USEG | Fri Apr 04 1997 11:32 | 14 |
Hi,
I know this is a bit late, but I don't usually have time to read
this conference.
The problem that is referred to here has been fixed with a patched
rpc.rquotad. The official patch is not out, yet, but there is
one coming. The problem had to do with a system call to get the
host's IP address for use with the connect() system call. I
changed it to use either gethostbyname("localhost") or, an IP
address of all zeroes if that failed.
Rob Marshall
USEG
| |||||