On Tue, Jul 22, 2014 at 12:59:04PM -0500, wireless wrote: > > 2014-07-22 9:40 GMT-03:00 Craig Small <csm...@enc.com.au>: > >> On Mon, Jul 21, 2014 at 09:51:03AM -0500, wireless wrote: > >>> On 07/21/14 08:21, Sebastian Martinez wrote: > >>>> this inteface fails randomly but when i run the poller manualy works ok > >> Hmm, it sounds like load or something is not being correctly separated > >> from other pollers. > >> > >> I'm wondering if moving the engine temp directory to a ram drive might > >> help? The other idea is to make fping use a popen if possible. > > > who i do this?. > > In the codes [1]: The ram drive is basically a matter of finding one and then adjusting the configuration to use that for the engine temp drive.
> If have to audit the code(s) and find the opportunities to use popen. > You have to understand that popen in php, and other scripting languages, > is not the same as popen as it is in a "C" program. Yes.. php has some, let's say, interesting ideas when it comes to forks and other processes. > You start playing around with one part of files and soon you'll find > the monotonic nature of the mysql database engine, might be a bottleneck > that is only resolved buy upgrading to postgresql. Lot of folks are I agree, postgresql is much better when the going gets tough. Unfortunately JFFNMS does not really use the power of postgresql fully. The latest version does work with postgresql now, which is a bonus. For big installations get your DBA to do some analysis on the postgresql database, there is room for improvement but its beyond me. > kernel tuned exclusively for linux_jffnms_mysql performance. In fact > that ".config" file would be of keen interest to the wider jffnms > community and maybe should be part of the jffnms website (Craig?). I don't have enough equipment at home to stress JFFNMS. Unfortunately my day job they'd take a dim view of putting some unauthorised device on the network, so no dice there either. I have smashed the mysql database in a previous job, it was groaning under the weight. I found removing the raw logs regularly helped there. RRD is ok, but not ideal, you need to have fast disks for it. Sadly, I've got a lot of that fixed in another project but its still a work in progress. It's asynchronous, multi-threaded and works well with databases. I've had to put delays in just not to overwhelm the SNMP agents. It's also incomplete :/ - Craig -- Craig Small (@smallsees) http://enc.com.au/ csmall at : enc.com.au Debian GNU/Linux http://www.debian.org/ csmall at : debian.org GPG fingerprint: 5D2F B320 B825 D939 04D2 0519 3938 F96B DF50 FEA5 ------------------------------------------------------------------------------ Want fast and easy access to all the code in your enterprise? Index and search up to 200,000 lines of code with a free copy of Black Duck Code Sight - the same software that powers the world's largest code search on Ohloh, the Black Duck Open Hub! Try it now. http://p.sf.net/sfu/bds _______________________________________________ jffnms-users mailing list jffnms-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jffnms-users
