Output limit?

[danilo.souza]

Hi,

I run minemeld (standalone) in a virtual machine with 2 cpu, 6GB RAM and 40GB hd. My config has 63 miners (mainly youtube miners and ransomware trackers), 13 aggregators and 30 output nodes. The miners start the job, but when it reach the band "85k-95k indicators", the Minemeld stops mining. The miners get the status "started" and few of them "stopped". It doesn't restart the service each x sec, it just stop mining. I know it stops mining, because one of the Youtube channels has more than 30k videos and the miner only gets few URLs, the same with the ransonware IP trackers.

I noticed that my Minemeld hardly aged-out or removes indicators. For example, in a typical dashboard showed in many articles, the monitor presents the number of aged-out or removed indicators as a parable (half sine). In my case, figure bellow, I have flat lines. I thought it could be something related with NTP leading Minemeld to be to out of resources with so many indicators, but the time configuration in my server is perfect.

Finally, some time later (there is not a specific interval) the dashbord shows 0 indicators.

Could someone give me any tip, advice, help?

Thanks in advanced.

[danilo.souza]

Hi,

I run minemeld (standalone) in a virtual machine with 2 cpu, 6GB RAM and 40GB hd. My config has 63 miners (mainly youtube miners and ransomware trackers), 13 aggregators and 30 output nodes. The miners start the job, but when it reach the band "85k-95k indicators", the Minemeld stops mining. The miners get the status "started" and few of them "stopped". It doesn't restart the service each x sec, it just stop mining. I know it stops mining, because one of the Youtube channels has more than 30k videos and the miner only gets few URLs, the same with the ransonware IP trackers.

I noticed that my Minemeld hardly aged-out or removes indicators. For example, in a typical dashboard showed in many articles, the monitor presents the number of aged-out or removed indicators as a parable (half sine). In my case, figure bellow, I have flat lines. I thought it could be something related with NTP leading Minemeld to be to out of resources with so many indicators, but the time configuration in my server is perfect.

Finally, some time later (there is not a specific interval) the dashbord shows 0 indicators.

Could someone give me any tip, advice, help?

Thanks in advanced.

[danilo.souza]

Hi,

I'm have some information to share with you guys.

Last friday (27/04) I commited my configuration trying to get the details of the problem. Today (30/04), I get from netstat (image 1) that a lot of process have bytes in the queue to be processed (green lines). The process 64058, for example has more than 900kB to be procees. Since friday!

If you see the answer from the top command (image 2), this procees is sleeping.

From the strace command (image 3) I got that the process is wating for something. But I don't know what it is waiting for, this is a exclusive MM machine, I don't have any other application stealing resources from MM.

If someone could help me with this, I appreciate.

Best regards.