[prev in list] [next in list] [prev in thread] [next in thread] 

List:       amd-dev
Subject:    Re: AIX bos.net.nfs.client 5.1.0.15 breaks amd (5.1.0.10 works fine)
From:       Don Davis <davis () roguewave ! com>
Date:       2002-01-29 23:24:10
[Download RAW message or body]

Erez Zadok wrote:

> In message <3C55E2BB.E00C8C69@roguewave.com>, Don Davis writes:
> > Howdy,
> >
> > We recently applied some updates needed by db2.  This worked great
> > except that amd stopped working. One of the sets installed was
> > bos.net.nfs.client 5.1.0.15 when we backed this patch set out it works.
> >
> > This problem has occured on multiple versions of amd including 6.1a4.
> > (which has some other issues that I am trying to work on)
> >
> >
> > ------------------------------------------------------ from the log.
> >
> > Jan 28 08:31:12 tempest amd[426088]/info:  file server localhost type
> > local starts up
> > Jan 28 08:31:12 tempest amd[434282]/error: /amd: mount: Invalid argument
> >
> > Jan 28 08:31:13 tempest amd[434282]/error: /amd: mount: Invalid argument
> >
> > Jan 28 08:31:13 tempest amd[434282]/fatal: mount_toplvl: Invalid
> > argument
> > Jan 28 08:31:13 tempest amd[426088]/error: /amd: mount (afs_cont):
> > Invalid argument
> >
> > ---------------------------------------------------------------
> >
> > I realize that AIX5.x is not even on the map however we need it for our
> > developers and we rely heavily on amd.
> > Any help you can offer would be greatly appreciated.
> >
> >
> > Donald Davis,
> > Unix System Adminstrator.
> > Roguewave Software, Inc.
>
> We need some details: what has changed in AIX b/t these two patches?

AMD stopped working. As far as I can tell the files which changed are.

/usr/lib/drivers/autofs.ext
/usr/sbin/rpc.mountd
/usr/sbin/nfso
/usr/lib/drivers/nfs.ext
/usr/sbin/mknfs
/usr/lib/ras/autoload/nfsdb.cmd
/usr/sbin/nfso
/usr/sbin/automountd
/usr/sbin/rpc.mountd
/usr/sbin/autofsmount
/sbin/helpers/aufsmnthelp
/sbin/helpers/nfsmnthelp

>
>
> What of am-utils are you using?
>

We were using a pre-ldap version (6.0.0?) which another admin built. I
unfortunately big fingered the binary attempting to build
am-utils-6.1a4 which builds and seems to run however I am still having some
problems with the mapping. Currently we are running with a binary from one of
the 4.3.3 systems (amd 5.2.2.2).

>
> How did you managed to compile am-utils to aix5, given that it hasn't been
> ported to this OS version yet?

3 things.

I configured it like this.

./configure  --prefix=/usr/rwlocal --without-hesiod --without-ldap
--disable-fast-install

I tweeked on config.guess (I could probably just have changed the case
lable).

diff -c config.guess /usr/rwlocal/src/am-utils-6.1a4/config.guess
*** config.guess        Wed Nov 22 03:10:05 2000
--- /usr/rwlocal/src/am-utils-6.1a4/config.guess        Fri Jan 25 15:48:32
2002
***************
*** 484,489 ****
--- 484,503 ----
        fi
        echo ${IBM_ARCH}-ibm-aix${IBM_REV}
        exit 0 ;;
+     *:AIX:*:5)
+       IBM_CPU_ID=`/usr/sbin/lsdev -C -c processor -S available | head -1 |
awk '{ print $1 }'`
+       if /usr/sbin/lsattr -EHl ${IBM_CPU_ID} | grep POWER >/dev/null 2>&1;
then
+               IBM_ARCH=rs6000
+       else
+               IBM_ARCH=powerpc
+       fi
+       if [ -x /usr/bin/oslevel ] ; then
+               IBM_REV=`/usr/bin/oslevel`
+       else
+               IBM_REV=5.${UNAME_RELEASE}
+       fi
+       echo ${IBM_ARCH}-ibm-aix${IBM_REV}
+       exit 0 ;;
      *:AIX:*:*)
        echo rs6000-ibm-aix
       exit 0 ;;

I fixed a problem with an undefined subroutine. (if DEBUG=0);

#diff -c libamu/xutil.c /usr/rwlocal/src/am-utils-6.1a4/libamu/xutil.c
*** libamu/xutil.c      Tue Jan  9 19:35:23 2001
--- /usr/rwlocal/src/am-utils-6.1a4/libamu/xutil.c      Fri Jan 25 11:07:28
2002
***************
*** 415,420 ****
--- 415,421 ----
  {
    return cmdoption(opt, dbg_opt, &debug_flags);
  }
+ #endif /* DEBUG */


  void
***************
*** 429,435 ****
    real_plog(XLOG_DEBUG, fmt, ap);
    va_end(ap);
  }
! #endif /* DEBUG */


  void
--- 430,436 ----
    real_plog(XLOG_DEBUG, fmt, ap);
    va_end(ap);
  }
!


  void
#

It seems to work with the new ibm patches backed out, though I am having some
problems getting it to map the system specific directories.
I suspect that this has to do with my not understanding the mapping well
enough though..It would be nice to have the patches installed.

Let me know if I have provided enough info here.

Don.




_______________________________________________
amd-dev mailing list: amd-dev@cs.columbia.edu
Am-utils: http://www.am-utils.org
[prev in list] [next in list] [prev in thread] [next in thread]