pcp
[Top] [All Lists]

Re: [pcp] pcp updates: QA for pmlogger - handle stale primary control an

To: Nathan Scott <nathans@xxxxxxxxxx>
Subject: Re: [pcp] pcp updates: QA for pmlogger - handle stale primary control and socket files
From: Mark Goodwin <mgoodwin@xxxxxxxxxx>
Date: Wed, 29 Jul 2015 21:19:33 +1000
Cc: pcp <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <172342487.185727.1438154956909.JavaMail.zimbra@xxxxxxxxxx>
References: <55B84483.3080604@xxxxxxxxxx> <172342487.185727.1438154956909.JavaMail.zimbra@xxxxxxxxxx>
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:31.0) Gecko/20100101 Thunderbird/31.3.0
On 07/29/2015 05:29 PM, Nathan Scott wrote:
Hi Mark,

----- Original Message -----

new qa/808 will fail unless you have the earlier commit, see below.

Changes committed to git://pcp.io/markgw/pcp/pcp.git master



I pushed some fixes to 280 and 430 that are related today - and those

thanks

made me wonder about the upgrade scenario.  Could you test the case
where we have a primary logger running from an older PCP (hardlink'd)
and upgrade?  Its not clear if that process will go smoothly now.

yes good point - I'll test it. I suspect the solution will be a %postun
scriptlet in the RPM spec to remove all pmlogger control files and sockets
and links thereto, whether hard links or soft. But that assumes we'll always
want to restart pmlogger services after an upgrade; is that reasonable?
(and ditto similar for Deb upgrades).

Cheers

<Prev in Thread] Current Thread [Next in Thread>