I think I'm seeing a uid wrap/overflow problem in the linux nfs client
or userid libc stuff -- but it might be in the VMS NFS server...
SERVER: VMS multinet NFS server
CLIENT: 2.2.18 linux works ok with the 'nobody' account.
nobody@ogre:/home/falcon > ls -la |grep phurst
dr-x------ 2 nobody nogroup 199680 Jun 26 11:07 phurst
I can list this dir and access files in it.
CLIENT: 2.4.18 linux doesn't work with the 'nobody' account.
nobody@teton:/home/falcon> id
uid=65534(nobody) gid=65534(nogroup) groups=65534(nogroup),65533(nobody)
nobody@teton:/home/falcon> ls -la |grep phurst
dr-x------ 2 4294967294 4294967294 199680 Jun 26 11:07 phurst
In this case the nobody account appears to go across as a -2 and the use
of long int instead of short int somewhere is bouncing the uid/gid
values up to the 4294967294 uid/gid number.. which causes some problems.
Anyone else seeing this? Is this a linux NFS fix that is showing a
multinet bug?
eric
Here is a tcpdump showing the -2/-2 stuff.
teton:~ # tcpdump host pleiad
tcpdump: listening on eth0
11:12:07.657523 teton.2306391190 > tyco.nfs: 108 getattr [|nfs] (DF)
11:12:07.657987 tyco.nfs > teton.2306391190: reply ok 96 getattr DIR
40755 ids -2/-2 sz 1536
11:12:07.658160 teton.2323168406 > tyco.nfs: 108 getattr [|nfs] (DF)
11:12:07.658649 tyco.nfs > teton.2323168406: reply ok 96 getattr DIR
40755 ids -2/-2 sz 1536
-------------------------------------------------------
This sf.net email is sponsored by: Jabber Inc.
Don't miss the IM event of the season | Special offer for OSDN members!
JabberConf 2002, Aug. 20-22, Keystone, CO http://www.jabberconf.com/osdn
_______________________________________________
NFS maillist - [email protected]
https://lists.sourceforge.net/lists/listinfo/nfs