Le 22/06/2015 11:52, Carmelo Ponti (CSCS) a écrit :
> I cannot understand how 30805 records can fill the device but maybe I'm
> misunderstanding something.
Your rates seem not so high.
It looks like more the total number or records created on this 
filesystem was huge. Is this possible? Is it old?

Or you just hit a bug/corruption in the llog on-disk structure which 
makes it misbehave. It seems the bug even prevent you from removing the 
llog from the MDT.
When we faced a similar corruption, we analyzed the llog on-disk 
content, fixed it, then removing and restart a new scan+changelog. In 
the meanwhile we deployed some lustre fixes.
At some point in the past we had (for some reasons) more than 1billion 
waiting records. It seems this is far from your case :)


Aurélien

------------------------------------------------------------------------------
Monitor 25 network devices or servers for free with OpManager!
OpManager is web-based network management software that monitors 
network devices and physical & virtual servers, alerts via email & sms 
for fault. Monitor 25 devices for free with no restriction. Download now
http://ad.doubleclick.net/ddm/clk/292181274;119417398;o
_______________________________________________
robinhood-support mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/robinhood-support

Reply via email to