[OpenIndiana-discuss] SMF service that should not shutdown with the OS

Jim Klimov jimklimov at cos.ru
Mon Jun 16 06:37:53 UTC 2014


On 2014-06-15 14:42, Oscar del Rio wrote:
> On 14/06/2014 6:09 PM, Jim Klimov wrote:
>> an interesting puzzle came up: how can one define an SMF service
>> so that it would ignore the system-wide shutdown and would only
>> (perhaps ungracefully) die off when the kernel/zone goes down?

> Perhaps a service without a "stop" method?
>
> http://docs.oracle.com/cd/E26502_01/html/E28980/firstboot-2.html
>
> <!--
>      The exec attribute below can be changed to a command that SMF
>      should execute to stop the service.  See smf_method(5) for more
>      details.
> -->
>      <exec_method timeout_seconds="60" type="method" name="stop"
>         exec=":true"/>

Thanks, that's a nice hint. I'll tell the admins on location to
try this in quiet time :) I wonder however if it would apply to
an inetd service instance (something I did not realize before) -
i.e. can we override the stop method for an inetd instance?..
Does that make any systematic sense at all? ;)

Or should they rather spawn a separate complete inetd clone
service (without an actual stop method) with this telnet server
instance re-attached to it, if two inetd's are at all possible
in one system? I don't think it would be fatal to override the
one and only system inetd service if need be, though... It does
not have many uses here anyway (there are also some netcat port
redirectors made as "quick workarounds" that stuck forever, but
it would not hurt to keep them up as long as possible either).

Heh, that's all something time and experiment would tell - but
thanks anyhow for giving a direction to try out :)

I'll let the list know if they come back and this experiment
leads to something fruitful (or not).

Thanks,
//Jim Klimov




More information about the openindiana-discuss mailing list