On 05/06/15 15:34, Carmelo Ponti (CSCS) wrote:
I tried it already and doesn't help much. The only trick I found (by
chance and I don't know the real reason) to reduce GET_INFO_FS is to
stop robinhood, umount and mount lustre again and restart robinhood. I
did it and this is the result:
It sounds familiar...
Lustre metadata cache get slower as its is populated.
Do you run the commands I gave you in my mail of 04/24:
(on the robinhood node)
- set a lower lru_size (after each mount):
lctl set_param ldlm.namespaces.*.lru_size=400
- run regularly in cron (e.g. hourly):
lctl set_param ldlm.namespaces.*.lru_size=clear
------------------------------------------------------------------------------
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
_______________________________________________
robinhood-support mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/robinhood-support
