samba - Nov 2022 - accidentally upgraded DC to 4.17.3 ... didn't work

Am 22.11.22 um 12:00 schrieb Stefan G. Weichinger via samba:
> Additional observations on ADC2:
> 
> # tail log.samba
> [2022/11/22 11:52:06.058000,? 1] 
>
../../source4/dsdb/kcc/garbage_collect_tombstones.c:67(garbage_collect_tombstones_part)
>  ? Doing a full scan on 
> CN=Configuration,DC=arbeitsgruppe,DC=ikw-amstetten,DC=at and looking for 
> deleted objects
> [2022/11/22 11:53:57.118027,? 1] 
>
../../source4/dns_server/dns_query.c:1140(dns_server_process_query_got_auth)
>  ? dns_server_process_query_got_auth: Failed to add SOA record: 
> WERR_DNS_ERROR_RCODE_FORMAT_ERROR
> [2022/11/22 11:53:57.959838,? 1] 
>
../../source4/dns_server/dns_query.c:1140(dns_server_process_query_got_auth)
>  ? dns_server_process_query_got_auth: Failed to add SOA record: 
> WERR_DNS_ERROR_RCODE_FORMAT_ERROR
> [2022/11/22 11:54:24.196900,? 1] 
>
../../source4/dns_server/dns_query.c:1140(dns_server_process_query_got_auth)
>  ? dns_server_process_query_got_auth: Failed to add SOA record: 
> WERR_DNS_ERROR_RCODE_FORMAT_ERROR
> [2022/11/22 11:54:25.032117,? 1] 
>
../../source4/dns_server/dns_query.c:1140(dns_server_process_query_got_auth)
>  ? dns_server_process_query_got_auth: Failed to add SOA record: 
> WERR_DNS_ERROR_RCODE_FORMAT_ERROR
I fixed this one by removing wrong A-Records from the DNS:

there was one entry per VLAN for the DCs ...
> # tail log.wb-ARBEITSGRUPPE
> [2022/11/20 00:00:07.109539,? 1] 
> ../../source3/winbindd/winbindd.c:364(winbindd_sig_hup_handler)
>  ? Reloading services after SIGHUP
> [2022/11/22 09:03:04.040791,? 0] 
> ../../source3/winbindd/winbindd_dual.c:1957(winbindd_sig_term_handler)
>  ? Got sig[15] terminate (is_parent=0)
> [2022/11/22 11:22:52.368820,? 0] 
> ../../source3/winbindd/winbindd_dual.c:1957(winbindd_sig_term_handler)
>  ? Got sig[15] terminate (is_parent=0)
> [2022/11/22 11:50:17.402770,? 0] 
> ../../source3/winbindd/winbindd_dual.c:1957(winbindd_sig_term_handler)
>  ? Got sig[15] terminate (is_parent=0)
> [2022/11/22 11:58:34.500594,? 0] 
> ../../source3/winbindd/winbindd_dual.c:1957(winbindd_sig_term_handler)
>  ? Got sig[15] terminate (is_parent=0)
winbindd still just terminates somehow

increased loglevel to 5

->

[2022/11/22 13:19:27.912603,  5] 
../../source3/winbindd/winbindd_dual_srv.c:72(_wbint_InitConnection)
   _wbint_InitConnection: ARBEITSGRUPPE returning without initialization 
online = 1

Forgive my flooding, but I really search for a solution and would like 
to avoid downgrading again.

Compared the ouput of

ps axf | egrep "samba|smbd|winbindd"

on both DCs. Yes, different samba-versions.

adc1, samba 4.16.6:

     938 ?        S      0:00  \_ samba: tfork waiter process(941)
     941 ?        S      0:04  |   \_ samba: task[winbindd] pre-fork master
     946 ?        S      0:00  |       \_ samba: tfork waiter process(950)
     950 ?        Ss     0:39  |           \_ /usr/sbin/winbindd -D 
--option=server role check:inhibit=yes --foreground
    1125 ?        S      0:03  |               \_ winbindd: domain child 
[ARBEITSGRUPPE]
    1128 ?        S      0:03  |               \_ winbindd: idmap child


adc2, samba 4.17.3:

    6203 ?        S      0:00  \_ samba: tfork waiter process(6205)
    6205 ?        S      0:00  |   \_ samba: task[winbindd] pre-fork master
    6209 ?        S      0:00  |       \_ samba: tfork waiter process(6218)
    6218 ?        Ss     0:00  |           \_ /usr/sbin/winbindd -D 
--option=server role check:inhibit=yes --foreground
    6229 ?        S      0:00  |               \_ winbindd: domain child 
[ARBEITSGRUPPE]


here that "idmap child" process is missing.

Is that OK? good or bad?

So far I always checked stuff like "wbinfo -u" after upgrades and 
interpreted a working output as "things went OK".