[OpenAFS] build errors.. was: RE: vos create failure on attached drive array ?

Jason Mccabe Jason.McCabe@mci.com
Thu, 11 Aug 2005 12:28:39 -0400

Looks like configure had trouble finding my relatively new build tools.
So I made some changes to Makefile.sun4x_58(.in) to use GCC and LEX.
Also MT_CFLAGS had to be changed from -mt (for SunCC) to -pthreads
(for GCC).

But now the linking of TVICED is failing because it's missing
this symbol:

  pthread_mutex_t rx_event_mutex;

from rx_event.c.

It appeared that LIBS for tviced should include librx.a,
but it doesn't in the MAKE from the source package, so I
added it.  My compiled version of librx.a doesn't define
that symbol anyway.. probably because of this logic in

 * This mutex protects the following global variables:
 * rxevent_initialized


pthread_mutex_t rx_event_mutex;


#endif /* AFS_PTHREAD_ENV */

Any ideas on how to fix that for the whole MAKE?

By the way, I configure:

   ./configure --with-afs-sysname=sun4x_58 --enable-transarc-paths


-----Original Message-----
From: openafs-info-admin@openafs.org [mailto:openafs-info-admin@openafs.org]
On Behalf Of Derrick J Brashear
Sent: Wednesday, August 10, 2005 12:09 PM
To: openafs-info@openafs.org
Subject: Re: [OpenAFS] RE: vos create failure on attached drive array ?

On Wed, 10 Aug 2005, Jason Mccabe wrote:

> This is running Veritas.  I've seen references to needing
> to run a NAMEI version of the AFS server (whatever that is).
> I suppose I could try to build AFS from source unless someone
> already has suitable Solaris 8 64- or 32-bit binaries.

configure --enable-namei-fileserver
and then make as normal.

However, 104 is
#define VVOLEXISTS      104     /* Volume already exists */

I'd guess you have leftover crap in /vicepa, unless the volserver is 
getting strange errors from the OS.

OpenAFS-info mailing list