HI all, I have install 2 domains both on linux servers running debian samba 3.0.20b-2+b1. (Latest) I have both domains trusting each other. Domain A have 300 users and the other domain B have 3000 users. I have "winbind" on the nsswitch.conf for both PDCs. I have not errors runnning wbinfo -u, or wbinfo -g except when I run it on Domain A PDC. Domain users group which all 3000 users are at failed to show up at the output. The rest of the domain groups are displayed. Looking in the winbindd log:- (Domain B PDC = BAUGLIR; Domain B=UWCSTU) -------------------------------------------------------------------- [2005/12/14 18:36:42, 10] nsswitch/winbindd_rpc.c:lookup_groupmem(539) rpc: lookup_groupmem UWCSTU sid=S-1-5-21-2723404422-2550591724-2764062575-513 [2005/12/14 18:36:52, 0] rpc_client/cli_pipe.c:rpc_api_pipe(438) cli_pipe: return critical error. Error was Call timed out: server did not respond after 10000 milliseconds [2005/12/14 18:36:52, 0] rpc_client/cli_pipe.c:rpc_api_pipe(438) cli_pipe: return critical error. Error was Call timed out: server did not respond after 10000 milliseconds [2005/12/14 18:36:52, 10] nsswitch/winbindd_rpc.c:lookup_groupmem(539) rpc: lookup_groupmem UWCSTU sid=S-1-5-21-2723404422-2550591724-2764062575-513 [2005/12/14 18:36:52, 3] nsswitch/winbindd_cm.c:connection_ok(819) Connection to BAUGLIR for domain UWCSTU has died or was never started (fd == -1) [2005/12/14 18:36:52, 0] rpc_client/cli_pipe.c:cli_rpc_close(1767) cli_rpc_open failed on pipe \samr to machine BAUGLIR. Error was Call timed out: server did not respond after 10000 milliseconds ------------------------------------ Can anyone know who to cache winbind well or increase the pagesize? I guess the timeout is because of the 3000 entries. Regards, adrian
Hi, are you telling me to install nscd and it will solve my problem? Also i read somewhere in the samba website that you should not run nscd with winbind. Is that true? If it is, what are some ways of improving the performance of winbind and how can I make it scale? Thanks for your replies. adrian Vijay Avarachen wrote:> I am not sure if this will help but I was getting strange errors and > often dead winbinds due to the large amount of users and groups. I have > had great success with setting up OpenLDAP for idmap backend. Now all > my Linux machines are authenticating users and I also use nscd to speed > things up and ease the load on OpenLDAP. > > > On 12/14/05, *Adrian Chow* <achow@uwcsea.edu.sg > <mailto:achow@uwcsea.edu.sg>> wrote: > > HI all, > > I have install 2 domains both on linux servers running debian samba > 3.0.20b-2+b1. (Latest) > > I have both domains trusting each other. > > Domain A have 300 users and the other domain B have 3000 users. I have > "winbind" on the nsswitch.conf for both PDCs. > > I have not errors runnning wbinfo -u, or wbinfo -g except when I run it > on Domain A PDC. Domain users group which all 3000 users are at failed > to show up at the output. The rest of the domain groups are displayed. > > Looking in the winbindd log:- (Domain B PDC = BAUGLIR; Domain B=UWCSTU) > -------------------------------------------------------------------- > [2005/12/14 18:36:42, 10] nsswitch/winbindd_rpc.c:lookup_groupmem(539) > rpc: lookup_groupmem UWCSTU > sid=S-1-5-21-2723404422-2550591724-2764062575-513 > [2005/12/14 18:36:52, 0] rpc_client/cli_pipe.c:rpc_api_pipe(438) > cli_pipe: return critical error. Error was Call timed out: server > did > not respond after 10000 milliseconds > [2005/12/14 18:36:52, 0] rpc_client/cli_pipe.c:rpc_api_pipe(438) > cli_pipe: return critical error. Error was Call timed out: server did > not respond after 10000 milliseconds > [2005/12/14 18:36:52, 10] nsswitch/winbindd_rpc.c:lookup_groupmem(539) > rpc: lookup_groupmem UWCSTU > sid=S-1-5-21-2723404422-2550591724-2764062575-513 > [2005/12/14 18:36:52, 3] nsswitch/winbindd_cm.c:connection_ok(819) > Connection to BAUGLIR for domain UWCSTU has died or was never started > (fd == -1) > [2005/12/14 18:36:52, 0] rpc_client/cli_pipe.c:cli_rpc_close(1767) > cli_rpc_open failed on pipe \samr to machine BAUGLIR. Error was > Call > timed out: server did not respond after 10000 milliseconds > > ------------------------------------ > > Can anyone know who to cache winbind well or increase the pagesize? I > guess the timeout is because of the 3000 entries. > > Regards, > > adrian > > -- > To unsubscribe from this list go to the following URL and read the > instructions: https://lists.samba.org/mailman/listinfo/samba > <https://lists.samba.org/mailman/listinfo/samba> > > > > > -- > "Knowledge is the only wealth that grows as you spend it, and diminishes > as you save it." > -- ancient Sanskrit saying
it has always been mentioned, that idmap_rid is the better backend in large organizations greez Adrian Chow wrote:> Hi, > > are you telling me to install nscd and it will solve my problem? Also i > read somewhere in the samba website that you should not run nscd with > winbind. Is that true? If it is, what are some ways of improving the > performance of winbind and how can I make it scale? > > Thanks for your replies. > > adrian > > Vijay Avarachen wrote: > >> I am not sure if this will help but I was getting strange errors and >> often dead winbinds due to the large amount of users and groups. I >> have had great success with setting up OpenLDAP for idmap backend. >> Now all my Linux machines are authenticating users and I also use nscd >> to speed things up and ease the load on OpenLDAP. >> >> On 12/14/05, *Adrian Chow* <achow@uwcsea.edu.sg >> <mailto:achow@uwcsea.edu.sg>> wrote: >> >> HI all, >> >> I have install 2 domains both on linux servers running debian samba >> 3.0.20b-2+b1. (Latest) >> >> I have both domains trusting each other. >> >> Domain A have 300 users and the other domain B have 3000 users. I >> have >> "winbind" on the nsswitch.conf for both PDCs. >> >> I have not errors runnning wbinfo -u, or wbinfo -g except when I >> run it >> on Domain A PDC. Domain users group which all 3000 users are at >> failed >> to show up at the output. The rest of the domain groups are >> displayed. >> >> Looking in the winbindd log:- (Domain B PDC = BAUGLIR; Domain >> B=UWCSTU) >> -------------------------------------------------------------------- >> [2005/12/14 18:36:42, 10] >> nsswitch/winbindd_rpc.c:lookup_groupmem(539) >> rpc: lookup_groupmem UWCSTU >> sid=S-1-5-21-2723404422-2550591724-2764062575-513 >> [2005/12/14 18:36:52, 0] rpc_client/cli_pipe.c:rpc_api_pipe(438) >> cli_pipe: return critical error. Error was Call timed out: server >> did >> not respond after 10000 milliseconds >> [2005/12/14 18:36:52, 0] rpc_client/cli_pipe.c:rpc_api_pipe(438) >> cli_pipe: return critical error. Error was Call timed out: >> server did >> not respond after 10000 milliseconds >> [2005/12/14 18:36:52, 10] >> nsswitch/winbindd_rpc.c:lookup_groupmem(539) >> rpc: lookup_groupmem UWCSTU >> sid=S-1-5-21-2723404422-2550591724-2764062575-513 >> [2005/12/14 18:36:52, 3] nsswitch/winbindd_cm.c:connection_ok(819) >> Connection to BAUGLIR for domain UWCSTU has died or was never >> started >> (fd == -1) >> [2005/12/14 18:36:52, 0] rpc_client/cli_pipe.c:cli_rpc_close(1767) >> cli_rpc_open failed on pipe \samr to machine BAUGLIR. Error was >> Call >> timed out: server did not respond after 10000 milliseconds >> >> ------------------------------------ >> >> Can anyone know who to cache winbind well or increase the >> pagesize? I >> guess the timeout is because of the 3000 entries. >> >> Regards, >> >> adrian >> >> -- >> To unsubscribe from this list go to the following URL and read the >> instructions: https://lists.samba.org/mailman/listinfo/samba >> <https://lists.samba.org/mailman/listinfo/samba> >> >> >> >> >> -- >> "Knowledge is the only wealth that grows as you spend it, and >> diminishes as you save it." >> -- ancient Sanskrit saying > >-- Michael Gasch Max Planck Institute for Evolutionary Anthropology Department of Human Evolution (IT) Deutscher Platz 6 D-04103 Leipzig Germany Phone: 49 (0)341 - 3550 137