Re: starting service order

[Marco Ermini]

<quota chi="Jan-Henrik Haukeland">
[...]
> I think as a start we should add (1) and (2) to our TODO list. What do the
> other commiters think?
[...]
</quota>

What about having a state "startup in progress" and/or "shutdown in
progress" for processes long to startup? and the ability to "fork" a
startup/shutdown in a way that doesn't "block" monit?

This is what I usually see when I startup a process in my application
server (I have a dependency between the apache instance and the "managed"
instance, it is a BEA WebLogic configured in a manager-managed setup; Sun
15k, Solaris 8): a message stating "monit: cannot read status from the
monit daemon" which blocks monit output for about 17 minutes (!) on a very
big machine :-/ with dozens of CPUs:

---8x---
address@hidden # monit status
The monit daemon uptime: 7h 2m

Device 'rootfs'                     accessible
Device 'weblogicfs'                 accessible
Host '3pi_CSU_physic_1'             online with all services
Host '3pi_CSU_physic_2'             online with all services
Host '3pi_CSU_load_balancer'        online with all services
Process 'nodemanager'               running
Process 'manager'                   running
Process 'managed_02_7001'           running
Process 'managed_02_7002'           running
Process 'managed_02_7003'           running
Process 'managed_02_7004'           not monitored
Process 'managed_02_7005'           not monitored
Process 'managed_03_7001'           not monitored
Process 'managed_03_7002'           not monitored
Process 'managed_03_7003'           not monitored
Process 'managed_03_7004'           not monitored
Process 'managed_03_7005'           not monitored
Host 'oracle_givs146'               online with all services
Process 'apache_02_8081'            running
Process 'apache_02_8082'            running
Process 'apache_02_8083'            running
Process 'apache_02_8084'            not monitored
Process 'apache_02_8085'            not monitored
Process 'apache_03_8081'            not monitored
Process 'apache_03_8082'            not monitored
Process 'apache_03_8083'            not monitored
Process 'apache_03_8084'            not monitored
Process 'apache_03_8085'            not monitored
address@hidden #
address@hidden #
address@hidden # monit start apache_02_8084
address@hidden # monit status
monit: cannot read status from the monit daemon
address@hidden # monit status
monit: cannot read status from the monit daemon
address@hidden # monit status
monit: cannot read status from the monit daemon
address@hidden # monit status
monit: cannot read status from the monit daemon
address@hidden # monit status
The monit daemon uptime: 7h 19m

Device 'rootfs'                     accessible
Device 'weblogicfs'                 accessible
Host '3pi_CSU_physic_1'             online with all services
Host '3pi_CSU_physic_2'             online with all services
Host '3pi_CSU_load_balancer'        online with all services
Process 'nodemanager'               running
Process 'manager'                   running
Process 'managed_02_7001'           running
Process 'managed_02_7002'           running
Process 'managed_02_7003'           running
Process 'managed_02_7004'           running
Process 'managed_02_7005'           not monitored
Process 'managed_03_7001'           not monitored
Process 'managed_03_7002'           not monitored
Process 'managed_03_7003'           not monitored
Process 'managed_03_7004'           not monitored
Process 'managed_03_7005'           not monitored
Host 'oracle_givs146'               online with all services
Process 'apache_02_8081'            running
Process 'apache_02_8082'            running
Process 'apache_02_8083'            running
Process 'apache_02_8084'            running
Process 'apache_02_8085'            not monitored
Process 'apache_03_8081'            not monitored
Process 'apache_03_8082'            not monitored
Process 'apache_03_8083'            not monitored
Process 'apache_03_8084'            not monitored
Process 'apache_03_8085'            not monitored
address@hidden #
---8x---

This is just because the bash script which is starting the processes (the
BEA WebLogic + the call to apachectl) is starting.

I think this could also improve system startup if using Monit as a sysV
replaacemente, since you can start services which are not dependent each
others at the same time, which usually doesn't happen in a SysV setup.


Regards.
-- 
Marco Ermini
http://www.markoer.org
Dubium sapientiae initium. (Descartes)
<< This message is for the designated recipient only and may contain
privileged or confidential information. If you have received it in error,
please notify the sender immediately and delete the original. Any other
use of the email by you is prohibited. >>