Increasing the original value retention on RRD's

[MSPk]

Hi Team,

I already see a large chunck of my Memory being used as cache.

My Nagios server has the below config
8 core CPU, 16G RAM and 200G HDD

and my current memory Utilisation is

procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu-----
r b swpd free buff cache si so bi bo in cs us sy id wa st
0 0 0 6007576 638136 8412248 0 0 0 3 0 0 0 0 100 0 0

a large part of the memory is being used for cache and i'm not even done with a quqrter of my configurations. Will the cache be freed as I add more servers for monitoring?

[abrist]

I don't have a good answer for you. The disk cache will always use what it needs and what it thinks it needs. There is a good reason why the default configuration exists in its current form. RRDs are bins, and the full bin must be loaded into memory in order to add checks. So these changes may be too much for your system. All I can do is suggest you try this in a test environment first. We have never attempted to set the RRA values anywhere near as high as you are trying to do. Interesting use case to say the least.

[jericho_g]

Is there any way to preserve the original history AND increase the retention values? As in exporting the RRD database prior to the change, then re-importing back.

[abrist]

Not easily. You can dump the original RRD to XML, but if you recreate the rrd from that XML it will be recreated with the old rra steps and values. If you change the XML to have the new values, the rrd will most likely fail creation, though you are welcome to try. Looking the following url:
https://oss.oetiker.ch/rrdtool/doc/rrdtune.en.html
It may be possible to tune some parameters. Have you tried that?

[MSPk]

Please close this post as solved