Hello: On Tue, Oct 19, 2010 at 9:41 AM, schleppy <[email protected]> wrote:
> I have to create hundreds of thousands of rrd files (one per user) to save > various statistics. In order to help with the disk I/O I was hoping to use > another system (perhaps Redis) for retrieval of the rrd files, which would > then exist in memory. Is it possible to feed the contents of the file from > memory to `rrdtool fetch` or `rrdtool update` commands? I would recommend you do a quick search of the mail archives for "rrdtool io" or similar search terms, this has been quite extensively discussed in the past. Some options for improving I/O are rrdcached, placing files in tmpfs, placing files in a loopback mountpoint of a single file, etc. Your suggestion of putting it into Redis may also be an interesting solution which may be worth investigating. However, I can tell you that with the above mentioned suggestions, you should be able to scale to millions of individual files. Cheers, Bernard _______________________________________________ rrd-users mailing list [email protected] https://lists.oetiker.ch/cgi-bin/listinfo/rrd-users
