Re: pass-through feature seems to cache data

Roman Medina-Heigl Hernandez wrote: Please, Thomas, could you name any of the CPAN modules which is useful with SQL backend. I'd like to have a look at it/them. Till now all CPAN modules I've reviewed aren't useful with a SQL backend. Thanks a lot. I'm talking about generic database access modu

Re: pass-through feature seems to cache data

Thomas Anders wrote: > Roman Medina-Heigl Hernandez wrote: > >> I knew it. But Thomas specifically referred to this CPAN's module. I'm >> trying to find info about Net-SNMP Perl module, examples, etc. It's not >> easy, it seems to be a great lack of info about it :-( > > > I referred to CPAN's N

Re: pass-through feature seems to cache data

Roman Medina-Heigl Hernandez wrote: I knew it. But Thomas specifically referred to this CPAN's module. I'm trying to find info about Net-SNMP Perl module, examples, etc. It's not easy, it seems to be a great lack of info about it :-( I referred to CPAN's Net::SNMP, because it offers a few usefu

Re: pass-through feature seems to cache data

Dave Shield wrote: > On Thu, 2005-10-13 at 00:59 +0200, Roman Medina-Heigl Hernandez wrote: > >>I've digged into CPAN's Net::SNMP module and I haven't found any >>function/method useful for "serving" SNMP tables based on a SQL table, >>which is what I'm looking for. > > Please note that (despite

Re: pass-through feature seems to cache data

On Thu, 2005-10-13 at 00:59 +0200, Roman Medina-Heigl Hernandez wrote: > I've digged into CPAN's Net::SNMP module and I haven't found any > function/method useful for "serving" SNMP tables based on a SQL table, > which is what I'm looking for. Please note that (despite the similar name), the Net::

Re: pass-through feature seems to cache data

On Tue, 2005-10-11 at 16:49 +0200, Roman Medina-Heigl Hernandez wrote: > Another question, then... which is the preferred method to extend > GET/SET functionality? Writing code :-) Actually, that's a serious answer. The pass mechanism is a useful "quick-n-dirty" approach for adding a new extensi

Re: pass-through feature seems to cache data

Thomas Anders wrote: > Roman Medina-Heigl Hernandez wrote: > >> Actually I'm going to have a look to stuff kindly posted by Turbo >> Frediksson: >> http://www.bayour.com/bacula-snmp/ >> Btw, it doesn't seem trivial at all. >> >> I can't believe there isn't any perl library for this. > > > There

Re: pass-through feature seems to cache data

Roman Medina-Heigl Hernandez wrote: Actually I'm going to have a look to stuff kindly posted by Turbo Frediksson: http://www.bayour.com/bacula-snmp/ Btw, it doesn't seem trivial at all. I can't believe there isn't any perl library for this. There are several. First, I'd highly recommend using

Re: pass-through feature seems to cache data

Dave Shield wrote: > On Tue, 2005-10-11 at 16:49 +0200, Roman Medina-Heigl Hernandez wrote: > >> Another question, then... which is the preferred method to extend >>GET/SET functionality? > > > Writing code :-) > > Actually, that's a serious answer. > The pass mechanism is a useful "quick-n-dir

Re: pass-through feature seems to cache data

Dave Shield wrote: The proper fix is to recode the 'pass.c' implementation to use the v5 APIs, and one of the cache-related helpers - either the cache helper directly (as per 'extend.c'), or via the 'stash-cache' helper. How hard is this expected to be? May this also be the preferred way to fi

Re: pass-through feature seems to cache data

Thanks a lot ! I've recompiled net-snmp by modifying: include/net-snmp/net-snmp-config.h and it works in "real-time" mode. bye! --Alexandre. Roman Medina-Heigl Hernandez wrote: Dave Shield wrote: The workaround is to edit the file include/net-snmp/net-snmp-config.h remove the d

Re: pass-through feature seems to cache data

Dave Shield wrote: > The workaround is to edit the file > include/net-snmp/net-snmp-config.h > remove the definition of EXCACHETIME, and re-compile the agent. > > The proper fix is to recode the 'pass.c' implementation to use > the v5 APIs, and one of the cache-related helpers - either the >

Re: pass-through feature seems to cache data

AP> the pass-through feature of net-snmp ... AP> seems to cache data for 30 seconds RMHH> ... newer versions doesn't perform caching. At least RMMH>according to the Faq: Note that with the 4.2.x and 5.0.x lines, return values are cached within the agent for 30 seconds, rat

Re: pass-through feature seems to cache data

You're right. I've also tested the same version on Linux and caching persists :-( If somebody knows a workaround/fix for this I'd be also interested. Cheers, -Roman Alexandre Pashai wrote: > ...I've tested this feature with the latest version of net-snmp running > on Solaris 9: > > bash-2.05# /u

Re: pass-through feature seems to cache data

...I've tested this feature with the latest version of net-snmp running on Solaris 9: bash-2.05# /usr/local/sbin/snmpd -version NET-SNMP version: 5.2.1.2 Web: http://www.net-snmp.org/ Email: net-snmp-coders@lists.sourceforge.net ...and I confirm that 'pass' directiv

Re: pass-through feature seems to cache data

I have the same problem too. Snmpd seems to refresh cache when it receives a SET request, which seems logical to me. The problem happens when you have other external methods of updating values which should be returned by Snmpd (for instance, if you have data store on a database, and want Snmpd to r

pass-through feature seems to cache data

hi all, thep ass-through feature of net-snmp works quite well. But, it seems to cache data for 30 seconds while i expect it to have a "real-time" behavior... Is it possible for net-snmp to exec the script on each snmpget request ? Thanks for answers. Alexandre -