Oracle FAQ Your Portal to the Oracle Knowledge Grid

Home -> Community -> Usenet -> c.d.o.server -> 10gR2 tru64 permissions

10gR2 tru64 permissions

From: <>
Date: 27 Mar 2007 07:36:27 -0700
Message-ID: <>


I have some problems on a TRU64 box running a 10gR2 database. They may all have the same cause but I can't find it.

First while trying dbua to upgrade from 8i, I get at 10% an ORA-01031, found some notes from oracle saying the owner group of Oracle_home should be dba. In this box it is oinsatll. I have recursiveley chgreped my oracle_home, restarted my dbua but I still got the error message.
Another strange thing about it is after the error message dbua seems to progress but very very very slowly ( 10% in ten minutes, then I have stopped it at 34% three hours after the db is only 400 Mo ...)

Then I tried to work on an other instance on the same computer. While trying to connect by sqlplus using SID I got the following error ORA-27140: attach to post/wait facility failed. Have tried to change permissions to 6751 on oracle binary according to an oracle note but with no results .

And the last and the least this other instance was running for some days when suddenly ( i can swear I wasn't making anything at this time) the connection became impossible whith error ORA-12537 TNS connection closed. Listner log shows :

TNS-12518: TNS:listener could not hand off client connection
TNS-12547: TNS:lost contact
TNS-12560: TNS:protocol adapter error
TNS-00517: Lost contact

Compaq Tru64 UNIX Error: 32: Broken pipe

listener trace shows :
nsglbgetRSPidx: returning ecode=0
sntpcall: only 0 bytes read
sntpcall: Can't read from pipe; err[1] = 32 nserror: nsres: id=7, op=72, ns=12547, ns2=12560; nt[0]=517, nt[1]=32, nt[2]=0; ora[0]=0, ora[1]=0, ora[2]=0
nsglhe: Direct handoff errors: ns=12547, ns2=12560

I have checked log size rmed it, without any effects, And then I have'nt had any more ideas or clues from Inet...

At this time I have a database which is running but I can't find anyway to connect ...

Looking for some possible causes to these problems, I found an undotbs at more than 19 Go.

Seems there are a lot of permissions problems on this box. Could they be the cause of all the errors ?

So, Am i right if I summarize the permissions in that way :

oracle user belongs to dba group then oinstall. all directories belong to group dba.
binaries receive 755 peremissions
only oracle binary receives 6751

If I try to fix my permissions in that way, Will i have the posibility to connect without restarting the computer ( It is part of a TRU64 cluster if I restart the computer an another node will restart the missing services ) ...

Any help in any of the problems would be greatly appreciated.

Best regards Received on Tue Mar 27 2007 - 09:36:27 CDT

Original text of this message