Re: Problems with rrdcached and RRDs::graph
kevin brintnall <[email protected]>
| Newsgroups | gmane.comp.db.rrdtool.devel |
|---|---|
| Message-ID | <[email protected]> |
On Fri, Jan 29, 2010 at 10:16:58AM -0800, SaschaTen wrote: > > you are right there are text-file in the file-system. In this files there > are entries consists of UPDATE, a RRD-file/path and a VALUE. My opinion was > that this is the RRDcache feature, so it cached the incoming value into a > text-file before it flushed this value into the RRD. After I send the flush > command the daemon will check all files for entries for this RRD-file and > collect this data to update the RRD-file. It's like a journal for databases > with the data that was sent to the RRD-file and is it processed or will it > done later. For rollback or if an commit-comand comes to flush the data. I > hope you could floow my mind and I'm not so wrong. The updates are held in RAM, and written to the RRD files on demand whenever a "FLUSH" is received. In order to prevent data loss, the updates may be written to the journal file also. This is the text file you're seeing. This journal is used on start-up to ensure that the daemon's update cache is current after a restart. > Is that right or wrong? And if it's wrong how could I hold the entries in > the RAM so that I could send the flush command every 10 minutes respectively > short before I wan't to draw a graph. Try using the local UNIX socket instead of a TCP socket. This way, you will be able to use absolute path names (starting with '/'). This should solve your issue. -- kevin brintnall =~ /[email protected]/ > Regards > Sascha > -- > View this message in context: http://n2.nabble.com/Problems-with-rrdcached-and-RRDs-graph-tp4480089p4481552.html > Sent from the RRDtool Developers Mailinglist mailing list archive at Nabble.com. > > _______________________________________________ > rrd-developers mailing list > [email protected] > https://lists.oetiker.ch/cgi-bin/listinfo/rrd-developers