Re: [appliances-discuss-owner@opensolaris.org: Re:

Mike,

Thanks for your help on this thus far.  Been very helpful.

I ended up putting ssh on the machine not before I broke the
installation and had to pretty much reinstall solaris from
scratch.  I think that I'd done enough tweaking with it to
damage it beyond repair and it was probably due for a
re-install anyway.

I've only just gotten back to this recently and it looks as
if for some reason, SMF is still running startd for the
console-login service as shown by svcs -p.

online          9:53:47 svc:/system/sac:default
                9:53:47      179 sac
                9:53:47      180 ttymon
online          9:53:48
svc:/system/boot-archive-update:default
online          9:53:48 svc:/system/cron:default
                9:53:47      178 cron
online          9:53:48 svc:/system/system-log:default
                9:53:48      193 syslogd
online          9:53:48 svc:/system/utmp:default
                9:53:48      195 utmpd
online          9:53:48 svc:/network/ssh:default
                9:53:48      199 sshd
online          9:53:48 svc:/system/console-login:default
                9:53:48      203 svc.startd
online          9:53:48 svc:/system/dumpadm:default
online          9:53:48 svc:/milestone/multi-user:default
online          9:53:48
svc:/milestone/multi-user-server:default
online          9:53:48 svc:/network/service:default

# pstack 203
203:    /lib/svc/bin/svc.startd
 d1e24a67 nanosleep (ccb6dd58, 0)
 08071795 utmpx_mark_init (cb, 83d7000) + 2a  08069c8b
exec_method (834f180, 0, 82f8ba8, 80e4008, 0) + 3a
 0806a395 method_run (ccb6dfd0, 0, ccb6dfd4) + 351  0806ab6b
method_thread (82fb3c0) + 169
 d1e24476 _thr_setup (cd873800) + 4e
 d1e24760 _lwp_start (cd873800, 0, 0, ccb6dff8, d1e24760,
cd873800)

I looked this up in the code:

60 int
     61 utmpx_mark_init(pid_t pid, char *prefix)
     62 {
     63 	struct utmpx ut, *oldu;
     64 	int tmplen;
     65 	int ret;
     66 
     67 	while (st->st_initial &&
!utmpx_truncated)
     68 		(void) usleep(200 * USEC_PER_MSEC);
     69 
     70 	/*
     71 	 * Clean out any preexisting records for this PID,
as they
must be
     72 	 * inaccurate.
     73 	 */
     74 	utmpx_mark_dead(pid, 0, B_TRUE);
     75

Seems likely that it is blocking on the while statement.  Is
there anything stale perhaps that is being persisted across
both that may still be in the read-only state?

/var - mounted RW on /tmp (copied from /memRoot/root/var in
start up script) /etc - mounted RW on /tmp (copied from
/memRoot/root/etc in start up script)

Only thing I could think of is something under /lib? 
Console-login seems to start OK if I mount / as RW on
startup.

Brad
 
 
This message posted from opensolaris.org
_______________________________________________
appliances-discuss mailing list
appliances-discussopensolaris.org
http://mail.opensolaris.org/mailman/listinfo/appl
iances-discuss