[Nut-upsuser] Nut-upsuser Digest, Vol 72, Issue 21

[Raphaël]

Le 23/06/2011 14:00, nut-upsuser-request at lists.alioth.debian.org a écrit :
> Send Nut-upsuser mailing list submissions to
> 	nut-upsuser at lists.alioth.debian.org
>
> To subscribe or unsubscribe via the World Wide Web, visit
> 	http://lists.alioth.debian.org/cgi-bin/mailman/listinfo/nut-upsuser
> or, via email, send a message with subject or body 'help' to
> 	nut-upsuser-request at lists.alioth.debian.org
>
> You can reach the person managing the list at
> 	nut-upsuser-owner at lists.alioth.debian.org
>
> When replying, please edit your Subject line so it is more specific
> than "Re: Contents of Nut-upsuser digest..."
>
>
> Today's Topics:
>
>     1. Re: Restarting NUT twice in the same time creates 2 instances
>        of usbhid-ups driver (Arjen de Korte)
>
>
> ----------------------------------------------------------------------
>
> Message: 1
> Date: Thu, 23 Jun 2011 11:53:46 +0200
> From: Arjen de Korte<nut+users at de-korte.org>
> To: nut-upsuser at lists.alioth.debian.org
> Subject: Re: [Nut-upsuser] Restarting NUT twice in the same time
> 	creates 2 instances of usbhid-ups driver
> Message-ID:
> 	<20110623115346.Horde.IpRfXUisJlFOAw0q00qnMAA at mail.de-korte.org>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed; DelSp=Yes
>
> Citeren Rapha?l<raphael.broutin at ve-hotech.com>:
>
>> When reading /etc/init.d/nut, i see that when restarting nut, script
>> stop nut and start it again.
>> When starting again the driver is one more time started without any
>> verification that its already launched.
> Stopping a driver is an asynchronous process. So you either have to
> check if the PID file is removed before assuming the driver has
> actually stopped or you need to wait a couple of seconds before
> (re)starting the drivers.
>
>> I think there is a problem in code there.
> This is an inherent effect of the method that is used to signal
> drivers to stop. On systems with multiple driver instances, you don't
> want to hang around waiting for a driver to actually stop before going
> to the next one.
>
>> have you a solution to bypass this problem without rewriting /etc/init.d/nut
> The easiest fix is to add a delay between
>
>      stop nut
>      sleep 2
>      start nut
>
> Best regards, Arjen
Testing PID before launching a new instance of driver is a good solution 
but need to be implemented in /etc/init.d/nut

The sleep solution is incorrect because if i do in parallel twice 
(stop,sleep,start). The two starts will be done at the same time ( i 
mean stop will be already done) and there will be 2 instance of driver.
Or i dont understand.

Testing PID before and launching new instance of driver if pid does not 
exist) sould be the right solution but only if  2 UPS needs 1 driver 
else this solution fails

Regards.


-- 
Raphaël