[Gluster-users] Any success on FreeBSD ?

[Csaba Henk]

On 2011-04-29, Leon Meßner <l.messner at physik.tu-berlin.de> wrote:
> On Mon, Apr 25, 2011 at 10:54:24PM +0530, Venky Shankar wrote:
>> On Monday 25 April 2011 02:12 AM, Leon Meßner wrote:
>> > Hi,
>> >
>> > i wanted to know if anyone succeeded in building glusterfs-3.1.4 on
>> > FreeBSD. The problems will probably by manifold. One of them is there is
>> > no TCP_KEEPALIVE. I am using some older versions of glusterfs on
>> > FreeBSD and know there is no support but i'm not ready to give up hope
>> > so quickly.
>> FreeBSD has socket level keepalive calls that can be used with 
>> setsockopt() (SO_KEEPALIVE). The call to setsockopt() with TCP_KEEPALIVE 
>> should be wrapped under #ifndef GF_BSD_HOST_OS
>
> That worked. I now get past this point and the build bails out with:
>  
>  gcc -DHAVE_CONFIG_H -I. -I. -I../../.. -I../../../libglusterfs/src -I../../../contrib/uuid -D_FILE_OFFSET_BITS=64 -D__USE_FILE_OFFSET64 -D_GNU_SOURCE -DGF_BSD_HOST_OS -I../../../libglusterfs/src -I../../../rpc/rpc-lib/src -fPIC -Wall -g -shared -nostartfiles -I../../../argp-standalone -O0 -g -O2 -MT libgfxdr_la-glusterfs3-xdr.lo -MD -MP -MF .deps/libgfxdr_la-glusterfs3-xdr.Tpo -c glusterfs3-xdr.c  -fPIC -DPIC -o .libs/libgfxdr_la-glusterfs3-xdr.o
> glusterfs3-xdr.c: In function 'xdr_gf_statfs':
> glusterfs3-xdr.c:33: warning: implicit declaration of function 'xdr_u_quad_t'
> glusterfs3-xdr.c: In function 'xdr_gfs3_open_rsp':
> glusterfs3-xdr.c:573: warning: implicit declaration of function 'xdr_quad_t'
> glusterfs3-xdr.c: In function 'xdr_gfs3_dirlist':
> glusterfs3-xdr.c:1460: error: 'gfs3_dirlist' has no member named 'd_fileno'
> glusterfs3-xdr.c:1470: warning: dereferencing type-punned pointer will break strict-aliasing rules
> ...
> *** Error code 1
>  
> Stop in /root/sourcen/glusterfs-3.1.4/rpc/xdr/src.
> ...
>  
> Regards,
> Leon

I have an private commit which solves basic compilation problems on FreeBSD.

https://github.com/csabahenk/glusterfs/commits/topic.freebsd

Although it's against an old state of the code, the issues with types / 
basic system parameters are the same, I guess.

cheers,
Csaba