Re: BiblePay | Masternodes | Sanctuaries | POBH - ASIC Resistant | 10% ORPHANS
Hi all. One of the Slovak guys messaged me with an issue he is having that I couldn't help with. Perhaps someone can give mesome ideas to help?
Basically:
1. On a Windows server, same machine as Linux boxes he owns, all his cpu stats are 100% on Rosetta. Or in other words, all of his processing power is BOINC-ing.
2. On Linux server, with WCG exactly the same, all cpu stats are 100%.
3. On Linux server, with Rosetta, some cpu cores are 100%, others are much lower.
He is on Ubuntu 16.04 LTS.
So for some reason Rosetta on Linux is not downloading enough tasks for his machine?
He tried reinstalling and restarting server but no luck.
Any suggestions?
Thanks
Basically:
1. On a Windows server, same machine as Linux boxes he owns, all his cpu stats are 100% on Rosetta. Or in other words, all of his processing power is BOINC-ing.
2. On Linux server, with WCG exactly the same, all cpu stats are 100%.
3. On Linux server, with Rosetta, some cpu cores are 100%, others are much lower.
He is on Ubuntu 16.04 LTS.
So for some reason Rosetta on Linux is not downloading enough tasks for his machine?
He tried reinstalling and restarting server but no luck.
Any suggestions?
Thanks
What is the RAM utilization on the Linux server? Rosetta is much more RAM demanding, it may be throttling the number of concurrent tasks because of the RAM limit.
Nice suggestion, I forgot about that.
I checked and there is still 7Gb RAM free.
Is it just a case of wait for Rosetta to send more tasks?
The easiest thing to solve this is by checking this machine with the BOINC manager GUI. There you have all the settings, stats etc. Maybe for some reason it is not set to 100% CPU time and 100% CPU cores utilization. Or there are not enough tasks.
One thing that happened to me from time to time (resulting in exactly the same situation that not all cores are fully utilized) is that one or more Rosetta tasks were stuck. BOINC then keeps them forever in "running" mode although nothing happens. You can easiliy see this (again with the BOINC manager) by sorting the tasks for "elapsed" time. If there is a job that took many hours or even days, just abort it and everything should be fine.
