From: Ermanno Polli (Ermanno.Polli@lnf.infn.it)
Date: Tue Mar 18 2003 - 09:16:51 EST
Dear all,
here is the story I haven't told this ML.
First, my configuration.
1) ES40 server, 2 667MHz CPU's, tru64 5.0A
2) PC with two 1.2 GHz Pentium CPU's, SuSE Linux 7.3
In setting up the Linux PC, I exported some filesystems.
When I mounted them on the ES40, I had "invalid credentials". And I was
root!
Well, to make a long story short, working with tcpdump and some
RFC's, I found out that I passed too many GID's in the mount call.
On tru64:
/usr/include/rpc/auth_unix.h:#define NGRPS 32
On SuSE:
/usr/include/rpc/auth_unix.h:#define NGRPS 16
In the RFC1057:
struct auth_unix {
unsigned int stamp;
string machinename<255>;
unsigned int uid;
unsigned int gid;
unsigned int gids<16>;
};
I stripped off some groups from root and voila', the mount succeded.
In this case tru64 didn't adhere to the RFC standard.
After having moved all my users and their data, I checked the move with
my normal user account then I tried to read my mail with pine.
Note that I have the INBOX on tru64 and all the pine folders on SuSE
Linux.
Trying to access some folder, pine frozed. No way of getting out of it.
I had to become root on another terminal window and I "kill -9"-ed pine.
I'm still trying to understand the tcpdump I had during the freeze, but,
as I said in the previuos summary, I suspect some dropping of packets.
An addendum to the previous request:
I tried to copy files from a RedHat linux, and I had the same kind of
file corruption I had with SuSE Linux, so it's not dependent on the
linux flavor.
So, the question is:
has anyone succeded in accessing nfs mounted disks on some linux from
tru64 v5.0A, patch kit #3 and all the CSP's up to the sendmail bug?
Thank you again,
Ermanno Polli
ermanno.polli@lnf.infn.it
This archive was generated by hypermail 2.1.7 : Sat Apr 12 2008 - 10:49:12 EDT