pcp
[Top] [All Lists]

Re: [pcp] pmda persistent indom cache access issues, part 2

To: Ken McDonell <kenj@xxxxxxxxxxxxxxxx>, Marko Myllynen <myllynen@xxxxxxxxxx>
Subject: Re: [pcp] pmda persistent indom cache access issues, part 2
From: Nathan Scott <nathans@xxxxxxxxxx>
Date: Thu, 25 Feb 2016 16:35:58 -0500 (EST)
Cc: pcp developers <pcp@xxxxxxxxxxx>
Delivered-to: pcp@xxxxxxxxxxx
In-reply-to: <56CF5866.20407@xxxxxxxxxxxxxxxx>
References: <56CED581.8020305@xxxxxxxxxx> <56CF5866.20407@xxxxxxxxxxxxxxxx>
Reply-to: Nathan Scott <nathans@xxxxxxxxxx>
Thread-index: rAQaWzmHx58PLSCfOWzU142cRQPOtA==
Thread-topic: pmda persistent indom cache access issues, part 2

----- Original Message -----
> [...]
> If this is not possible, then the only options I can think of (assuming mode
> 777 and mode 1777 directories in /var are unpalatable) would be a
> libpcp_pmda extension to use (a new) $PCP_PMDA_CACHE_DIR variable from the
> environment or possibly a new method in libpcp_pmda to change the indom
> cache directory (for just the caller PMDA), e.g. pmdaCacheInit(char *dir)
> 
> Either would be trivial if the use case is compelling ... the environment
> variable would not require any library ABI or wrapper code changes.

FWIW, from oracle experience - its a slight pain to go the Install script route
and a bit of a maintenance headache (apart from being obscure / non-obvious) -
have to remember to update the ./Install script every time an indom is added to
the PMDA.

An option to explicitly set the directory for an individual PMDA sounds good.

cheers.

--
Nathan

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