Hello,
We are running the latest Samba from the AIX Toolbox (please see below) and we combine this with Centrify DC + Centrify ADbindproxy to allow for Windows domain user login and home directory access via Samba. Yesterday, winbindd crashed abruptly with a series of errors show below (and annotated). This cut off access to the Samba share for a few hours until the depth of the failure was understood. I was able to simply restart winbindd and things appeared to go back to normal but I'm hoping to understand why this happened and find ways to avoid it. Please advise if anyone has some insight. Thanks!
[root@crcoacs1 tmp]# oslevel -s
7100-05-07-2038
[root@crcoacs1 tmp]# rpm -qa | grep -i samba
samba-winbind-4.12.10-2.ppc
samba-client-4.12.10-2.ppc
samba-devel-4.12.10-2.ppc
samba-winbind-clients-4.12.10-2.ppc
samba-common-4.12.10-2.ppc
samba-libs-4.12.10-2.ppc
samba-4.12.10-2.ppc
...from syslog...
1. This starts the winbindd errors and repeats hundreds of times.
../../auth/kerberos/kerberos_pac.c:327(kerberos_decode_pac)
can't repack the SRV signature: NT_STATUS_NO_MEMORY
2. Then this error takes over and repeast hundreds of times.
../../auth/kerberos/kerberos_pac.c:339(kerberos_decode_pac)
can't repack the RAW PAC: NT_STATUS_NO_MEMORY
3. Talloc Error - Samba Memory Handler
../../source3/lib/popt_common.c:68(popt_s3_talloc_log_fn)
Bad talloc magic value - unknown value
4. Winbindd Panics due to talloc error and dies.
../../source3/lib/util.c:830(smb_panic_s3)
PANIC (pid 1246006): Bad talloc magic value - unknown value
5. Trys to trace
../../lib/util/fault.c:277(log_stack_trace)
unable to produce a stack trace on this platform
Tries to core dump and create the dump file and fails.
../../source3/lib/dumpcore.c:283(dump_core)
Exiting on internal error (core file administratively disabled)
6. Sig Term.
../../source3/winbindd/winbindd.c:245(winbindd_sig_term_handler)
Got sig[15] terminate (is_parent=0)
# end #
------------------------------
Michael Metts
------------------------------