[Linux-cluster] Latest cvs
Ion Alberdi
ialberdi at histor.fr
Mon Jun 20 09:43:51 UTC 2005
Ion Alberdi wrote:
> I added this logs in find_root_by_ref:
>
> resource_t
> *
>
> find_root_by_ref(resource_t **reslist, char
> *ref)
>
> {
>
> resource_t
> *curr;
>
> int
> x;
>
>
>
> //MODIF
> ION
>
> clulog(LOG_NOTICE,"We are in find_root_by_ref we enter the
> loop\n");
>
> //MODIF
> ION
>
>
>
> list_do(reslist, curr)
> {
>
> //MODIF
> ION
>
>
> clulog(LOG_NOTICE,"%d\n",curr);
>
> //MODIF
> ION
>
>
>
> if (curr->r_rule->rr_root ==
> 0){
>
> //MODIF
> ION
>
> clulog(LOG_NOTICE,"curr->r_rule->rr_root ==
> 0\n");
>
> //MODIF
> ION
>
>
> continue;
>
>
> }
>
>
> /*
>
> This should be one operation - the primary
> attr
>
> is generally at the head of the
> array.
>
>
> */
>
> for (x = 0; curr->r_attrs &&
> curr->r_attrs[x].ra_name;
>
> x++)
> {
>
> if (!(curr->r_attrs[x].ra_flags &
> RA_PRIMARY)){
>
> //MODIF
> ION
>
> clulog(LOG_NOTICE,"!(curr->r_attrs[x].ra_flags &
> RA_PRIMARY)\n");
>
> //MODIF
> ION
>
>
> continue;
>
>
> }
>
> if (strcmp(ref,
> curr->r_attrs[x].ra_value)){
>
> //MODIF
> ION
>
> clulog(LOG_NOTICE,"strcmp(ref,
> curr->r_attrs[x].ra_value)=true\n");
>
> //MODIF
> ION
>
>
> continue;
>
>
> }
>
> return
> curr;
>
>
> }
>
> } while (!list_done(reslist,
> curr));
>
>
>
> return
> NULL;
>
> }
> what happens in gump:
>
> Jun 20 11:30:11 gump clurgmgrd[7231]: <notice> find_root_by_ref
> parameters:resources 134613668,groupname datamover
> Jun 20 11:30:11 gump clurgmgrd[7231]: <notice> We are in
> find_root_by_ref we enter the loop
> Jun 20 11:30:11 gump clurgmgrd[7231]: <notice> 134642544
> Jun 20 11:30:11 gump clurgmgrd[7231]: <notice> curr->r_rule->rr_root == 0
> Jun 20 11:30:11 gump clurgmgrd[7231]: <notice> 134617456
> Jun 20 11:30:11 gump clurgmgrd[7231]: <notice> find_root_by_ref in
> groups.c returned res 134617456
> Jun 20 11:30:11 gump clurgmgrd[7231]: <notice> res_exec in restree.c
> returned rv 0
> Jun 20 11:30:11 gump clurgmgrd[7231]: <notice> res_exec in restree.c
> returned rv 0
> Jun 20 11:30:11 gump clurgmgrd[7231]: <notice> _res_op_by_level in
> restree.c returned rv 0
> Jun 20 11:30:11 gump clurgmgrd[7231]: <notice> res_start in groups.c
> returned ret 0
> Jun 20 11:30:11 gump clurgmgrd[7231]: <notice> group_op in rg_state.c
> returned ret 0
> Jun 20 11:30:11 gump clurgmgrd[7231]: <notice> Service datamover started
>
> what happens in buba:
>
> Jun 20 11:31:29 buba clurgmgrd[10706]: <notice> find_root_by_ref
> parameters:resources 134613668,groupname datamover
> Jun 20 11:31:29 buba clurgmgrd[10706]: <notice> We are in
> find_root_by_ref we enter the loop
> Jun 20 11:31:29 buba clurgmgrd[10706]: <notice> 134616568
> Jun 20 11:31:29 buba clurgmgrd[10706]: <notice> curr->r_rule->rr_root
> == 0
> Jun 20 11:31:29 buba clurgmgrd[10706]: <notice> find_root_by_ref in
> groups.c returned res 0
> Jun 20 11:31:29 buba clurgmgrd[10706]: <notice> group_op in rg_state.c
> returned ret -1
> Jun 20 11:31:29 buba clurgmgrd[10706]: <warning> #68: Failed to start
> datamover; return value: 1
> Jun 20 11:31:29 buba clurgmgrd[10706]: <notice> Stopping service
> datamover
> Jun 20 11:31:29 buba clurgmgrd[10706]: <notice> find_root_by_ref
> parameters:resources 134613668,groupname datamover
> Jun 20 11:31:29 buba clurgmgrd[10706]: <notice> We are in
> find_root_by_ref we enter the loop
> Jun 20 11:31:29 buba clurgmgrd[10706]: <notice> 134616568
> Jun 20 11:31:29 buba clurgmgrd[10706]: <notice> curr->r_rule->rr_root
> == 0
> Jun 20 11:31:29 buba clurgmgrd[10706]: <notice> find_root_by_ref in
> groups.c returned res 0
> Jun 20 11:31:29 buba clurgmgrd[10706]: <crit> #12: RG datamover failed
> to stop; intervention required
> Jun 20 11:31:29 buba clurgmgrd[10706]: <notice> Service datamover is
> failed
> Jun 20 11:31:29 buba clurgmgrd[10706]: <crit> #13: Service datamover
> failed to stop cleanly
>
> The reslist seems not to be the same in the two nodes,
> I'll check my cluster.conf in my both nodes............
>
Sorry I'd better have waited before sending the previous mail:
the cluster.conf are identical.
So my question-> what does this lists represents, and what are the
files I have to check to see if there are identical
on my two nodes?
More information about the Linux-cluster
mailing list