Hi,

We had not one but two power outages last night, during which
my Ultra 10 went down. Now, when I try to power it up, the boot
process hangs at the stage after it says "Setting default IPv4
interface for multicast." I've tried powering off again,
disconnecting the network and power cables, and then letting
the box sit for several minutes before reconnecting the cables
and attempting another boot.

My box is an Ultra 10 running SunOS 5.8. Everything seems to
initialize correctly, until it gets to the point where it says
"WARNING: asppd has bee superseded; see  pppd(1m). After that,
I get two "NIS domainname is <domainname>" messages, both of
which state the same domain name. Then the rpc services seem to
start fine: "starting rpc services: rpcbind keyserv done." Next
the netmask for hme0 is set correctly. Then the "Setting default
IPv4 interface for multicast: add net 224.0/4: gateway <hostname>"
message comes, after which the system hangs.

Any clues about how to fix this?

Thanks in advance,

Scott

At 12/5/03 10:45, Scott Fordin bespake thusly:
> Hi,
> 
> We had not one but two power outages last night, during which
> my Ultra 10 went down. Now, when I try to power it up, the boot
> process hangs at the stage after it says "Setting default IPv4
> interface for multicast." I've tried powering off again,
> disconnecting the network and power cables, and then letting
> the box sit for several minutes before reconnecting the cables
> and attempting another boot.
> 
> My box is an Ultra 10 running SunOS 5.8. Everything seems to
> initialize correctly, until it gets to the point where it says
> "WARNING: asppd has bee superseded; see  pppd(1m). After that,
> I get two "NIS domainname is <domainname>" messages, both of
> which state the same domain name. Then the rpc services seem to
> start fine: "starting rpc services: rpcbind keyserv done." Next
> the netmask for hme0 is set correctly. Then the "Setting default
> IPv4 interface for multicast: add net 224.0/4: gateway <hostname>"
> message comes, after which the system hangs.

  That message is issued only after the function is completed,
  meaning that this is the last thing it did correctly.

  Suspect your machine can't get to the NIS server.  Made even
  worse if /etc/nsswitch.conf has nis in front of files.

  Search the rc2.d directory for the interface multicast message,
  think it is in S71rpc.  From there, you can add debug echo
  statements to S71 and S72inetsvc to see where it is hanging.

  First check your NIS server to make sure it is alive, particularly
  if this machine is statically bound rather than broadcast.  And
  make sure the NIS server can reach this box.


-- 
Still a Raiders fan, but no longer sure why.

In article <3fd0d23a$1_2@newspeer2.tds.net>, Scott Fordin <sfordin@ohelp.com> writes:
> Hi,
> 
> We had not one but two power outages last night, during which
> my Ultra 10 went down. Now, when I try to power it up, the boot
> process hangs at the stage after it says "Setting default IPv4
> interface for multicast." I've tried powering off again,
> disconnecting the network and power cables, and then letting
> the box sit for several minutes before reconnecting the cables
> and attempting another boot.
> 
> My box is an Ultra 10 running SunOS 5.8. Everything seems to
> initialize correctly, until it gets to the point where it says
> "WARNING: asppd has bee superseded; see  pppd(1m). After that,
> I get two "NIS domainname is <domainname>" messages, both of
> which state the same domain name. Then the rpc services seem to
> start fine: "starting rpc services: rpcbind keyserv done." Next
> the netmask for hme0 is set correctly. Then the "Setting default
> IPv4 interface for multicast: add net 224.0/4: gateway <hostname>"
> message comes, after which the system hangs.
> 
> Any clues about how to fix this?
> 
> Thanks in advance,
> 
> Scott
> 

Likely there is no contact with the NIS server:
no network connection, or wrong network (NIS server unreachable).
Try to ping the machine from another machine (preferable from the
NIS server).

-- 
Michael Tosch
IT Specialist
HP Managed Services Germany
Phone +49 2407 575 313

Scott Fordin wrote:
  > We had not one but two power outages last night, during which
 > my Ultra 10 went down. Now, when I try to power it up, the boot
 > process hangs at the stage after it says "Setting default IPv4
 > interface for multicast." I've tried powering off again,
 > disconnecting the network and power cables, and then letting
 > the box sit for several minutes before reconnecting the cables
 > and attempting another boot.

Thanks, Lon and Michael, for your replies. I've finally gotten
it to boot, and it looks like the problem was that the system
was hanging when trying to trying to remount an NFS share from
a Linux machine that had also gone down during the power outage.
Once I brought that Linux box back up, the NFS mount succeeded
and everything continued as normal.

I suppose this leaves me with two questions though:

    1. Why wouldn't the NFS mount attempt simply time out after
       a reasonable amount of time? I mean, I let the thing sit
       there overnight, but nothing. I would expect it to give
       up trying after awhile.

    2. I know this may seem like an obvious question, but short
       of booting from a CD or diskette, how can I interrupt the
       boot process so I can get to an ok prompt so I can then
       modify nsswitch.conf or vfstab? I know I know this, but
       my brain is drawing a blank right now. Too many deadlines!

Thanks again,

Scott

> My box is an Ultra 10 running SunOS 5.8. Everything seems to
> initialize correctly, until it gets to the point where it says
> "WARNING: asppd has bee superseded; see  pppd(1m). After that,
> I get two "NIS domainname is <domainname>" messages, both of
> which state the same domain name. Then the rpc services seem to
> start fine: "starting rpc services: rpcbind keyserv done." Next
> the netmask for hme0 is set correctly. Then the "Setting default
> IPv4 interface for multicast: add net 224.0/4: gateway <hostname>"
> message comes, after which the system hangs.
> 
> Any clues about how to fix this?
> 
> Thanks in advance,
> 
> Scott
> 

At 12/6/03 06:48, Scott Fordin's monkeys randomly typed:

> Scott Fordin wrote:
>   > We had not one but two power outages last night, during which
>  > my Ultra 10 went down. Now, when I try to power it up, the boot
>  > process hangs at the stage after it says "Setting default IPv4
>  > interface for multicast." I've tried powering off again,
>  > disconnecting the network and power cables, and then letting
>  > the box sit for several minutes before reconnecting the cables
>  > and attempting another boot.
> 
> Thanks, Lon and Michael, for your replies. I've finally gotten
> it to boot, and it looks like the problem was that the system
> was hanging when trying to trying to remount an NFS share from
> a Linux machine that had also gone down during the power outage.
> Once I brought that Linux box back up, the NFS mount succeeded
> and everything continued as normal.
> 
> I suppose this leaves me with two questions though:
> 
>     1. Why wouldn't the NFS mount attempt simply time out after
>        a reasonable amount of time? I mean, I let the thing sit
>        there overnight, but nothing. I would expect it to give
>        up trying after awhile.

  You are extremely likely to be mounting something that
  doesn't have a shell that could send an interrupt and are
  probably also mounting with a hard mount.

  I believe these type of NFS mounts may time out in 2038 when
  the unix clock expires, but certainly not before.

  It is a very very bad idea to be doing this in runlevel 2
  at such an early point in the process.  The only worse thing
  I can recall is a customer that decided to place their
  NIS maps on an NFS server that was a client of the NIS server
  that owned the maps.

  However, your symptom and error message are exactly what
  you get when the system cannot contact the NIS server.


> 
>     2. I know this may seem like an obvious question, but short
>        of booting from a CD or diskette, how can I interrupt the
>        boot process so I can get to an ok prompt so I can then
>        modify nsswitch.conf or vfstab? I know I know this, but
>        my brain is drawing a blank right now. Too many deadlines!

  Well, on the console you can try CTL-C, but for some hangs it
  is less useful than sacrificing a user^H^H^H^H chicken.

-- 
Still a Raiders fan, but no longer sure why.

"Scott Fordin" <sfordin@ohelp.com> wrote in message
news:3fd1ec2a$1_3@newspeer2.tds.net...
> Scott Fordin wrote:
>   > We had not one but two power outages last night, during which
>  > my Ultra 10 went down. Now, when I try to power it up, the boot
>  > process hangs at the stage after it says "Setting default IPv4
>  > interface for multicast." I've tried powering off again,
>  > disconnecting the network and power cables, and then letting
>  > the box sit for several minutes before reconnecting the cables
>  > and attempting another boot.
>
> Thanks, Lon and Michael, for your replies. I've finally gotten
> it to boot, and it looks like the problem was that the system
> was hanging when trying to trying to remount an NFS share from
> a Linux machine that had also gone down during the power outage.
> Once I brought that Linux box back up, the NFS mount succeeded
> and everything continued as normal.
>
> I suppose this leaves me with two questions though:
>
>     1. Why wouldn't the NFS mount attempt simply time out after
>        a reasonable amount of time? I mean, I let the thing sit
>        there overnight, but nothing. I would expect it to give
>        up trying after awhile.

The default mount option is to attempt the mount in the foreground and hard.
So if the mount is not available you are not  going anywhere. One thing you
can try to prevent this in the future is to set the mount options in the
last column in the vfstab file to bg. This will attempt the mount in the
background. It may be slow coming up as it will try for some minutes to
mount, but then will eventualy stop trying and  continue the boot process. I
am not sure but I think once the system comes up, it will keep trying the
mount and if and when the mount is available it will be mounted.
>
>     2. I know this may seem like an obvious question, but short
>        of booting from a CD or diskette, how can I interrupt the
>        boot process so I can get to an ok prompt so I can then
>        modify nsswitch.conf or vfstab? I know I know this, but
>        my brain is drawing a blank right now. Too many deadlines!

A Stop A should bring you to the ok prompt from where you can boot - s and
then either modify the nsswitch file or comment out the mount point in the
vfstab file.


See ya

Paul

Scott Fordin wrote:
> 
>    1. Why wouldn't the NFS mount attempt simply time out after
>       a reasonable amount of time? I mean, I let the thing sit
>       there overnight, but nothing. I would expect it to give
>       up trying after awhile.

NFS mounts have options of "soft" and "retry=<n>" which will fail the 
mount after a number of retries (<n> above, default 10000) with soft 
mounting, or "forever" with non-soft mounting (which is the default).

Beardy wrote:
> Scott Fordin wrote:
> 
>>
>>    1. Why wouldn't the NFS mount attempt simply time out after
>>       a reasonable amount of time? I mean, I let the thing sit
>>       there overnight, but nothing. I would expect it to give
>>       up trying after awhile.
> 
> 
> NFS mounts have options of "soft" and "retry=<n>" which will fail the 
> mount after a number of retries (<n> above, default 10000) with soft 
> mounting, or "forever" with non-soft mounting (which is the default).
> 

Excellent tips from all. Thanks again to everyone for your help.

Scott

Beardy wrote:
> Scott Fordin wrote:
> 
>>
>>    1. Why wouldn't the NFS mount attempt simply time out after
>>       a reasonable amount of time? I mean, I let the thing sit
>>       there overnight, but nothing. I would expect it to give
>>       up trying after awhile.
> 
> 
> NFS mounts have options of "soft" and "retry=<n>" which will fail the 
> mount after a number of retries (<n> above, default 10000) with soft 
> mounting, or "forever" with non-soft mounting (which is the default).
> 



ITYM "bg" rather than "soft".

-- 
Tony