[Pvfs2-users] problem with pvfs2 over diskless cluster

[Murali Vilayannur]

Hi Rob,
> Thanks for the detailed report. How did you end up fixing the "cannot 
> find db.h" compile errors?
I suspect Wenbin did not run configure with --with-db=</path/to/db>
> This "DB_THREAD mandates memory allocation flag on DBT data" seems to 
> indicate that the semantics have changed in Berkeley DB, and they're 
> requiring a certain way of returning data that wasn't required before 
> (probably for thread safety)? That's irritating.
Well.. this was after the patch and I am 100% certain that the patch is 
not the right way to fix the error if at all it fixes. Since the value 
of val.flags was not being set, db was complaining loudly.
i.e. DB_THREAD flag semantics have not changed since bdb 4.3 (atleast)
Snipper from Google's code search (I love this feature!)

if (check_thread <javascript:searchRef('check_thread')> && DB_IS_THREADED <javascript:searchRef('DB_IS_THREADED')>(dbp <javascript:searchRef('dbp')>) &&
	    !F_ISSET <javascript:searchRef('F_ISSET')>(dbt <javascript:searchRef('dbt')>, DB_DBT_MALLOC <javascript:searchRef('DB_DBT_MALLOC')> | DB_DBT_REALLOC <javascript:searchRef('DB_DBT_REALLOC')> | DB_DBT_USERMEM <javascript:searchRef('DB_DBT_USERMEM')>)) {
		CDB___db_err <javascript:searchRef('CDB___db_err')>(dbenv <javascript:searchRef('dbenv')>,
		    "DB_THREAD mandates memory allocation flag on DBT %s",
		    name <javascript:searchRef('name')>);
		return (EINVAL <javascript:searchRef('EINVAL')>);
	}




i.e. it requires that the flag value be set to one of the valid DB_DBT_* 
options.
Thanks,
Murali