Hello everyone.. I have a RS 2000 G11 where I run a validator NODE. The problem is that every now and then, when the application requires a bit of processing, the server kills the process giving me OUT OF MEMORY. I contacted support but they send me back saying that it is not their problem... generally computers when they have to process a process process the process in the potential they have and do not close it, while the Netcup product if the process uses resources excessively even for a short time kills the process. I wanted to ask if this thing is solvable or not.. I am fine with netcup but if I can not solve this problem I have to cancel the product.. thanks
OUT OF MEMORY RS 2000 G11
-
atomico -
4. März 2025 um 17:40 -
Erledigt
-
I contacted support but they send me back saying that it is not their problem...
Yes, that is correct.
You got a Server with fixed and guaranteed Ressources, that your OS can use.
If your App needs more Memory, you could
- buy a larger Server
- fix your app to use only the provided amount of Memory
- use swap-Memory, if that is not too slow
In short. Your Server - Your Problem.
-
Hello everyone.. I have a RS 2000 G11 where I run a validator NODE. The problem is that every now and then, when the application requires a bit of processing, the server kills the process giving me OUT OF MEMORY. I contacted support but they send me back saying that it is not their problem... generally computers when they have to process a process process the process in the potential they have and do not close it, while the Netcup product if the process uses resources excessively even for a short time kills the process. I wanted to ask if this thing is solvable or not.. I am fine with netcup but if I can not solve this problem I have to cancel the product.. thanks
atomico: It might make more sense to change the title from "OUT OF MEMORY RS 2000 G11" to "out of CPU time on my RS 2000 G11", because I understand it so, you mean the CPU time and not his memory.
According to my information, we are allowed to utilize all CPU threads up to 100 percent at any time. And this permanently.
-
Most probably your process gets killed by the OOM-Killer (Out of memory killer) which is part of your OS. It does this, because there is not enough free RAM to run all processes. Either your program has a memory leak or the server is simply not big enough (not enough RAM) to run it. It's an unmanaged root server, so netcup is only responsible for technical problems, not for the programs you run on it or the resources they need.
It's clear, that your RS 2000 will not get a single byte more than 16 GB of RAM, not even for a nanosecond. Maybe creating a swapfile or swap partition may help to keep your process alive, it depends on how much RAM is really needed and how crucial the processing speed is. In the end it may be better to use a server with more RAM.
-
Molto probabilmente il tuo processo viene ucciso dall'OOM-Killer (Out of memory killer) che fa parte del tuo sistema operativo. Lo fa perché non c'è abbastanza RAM libera per eseguire tutti i processi. O il tuo programma ha una perdita di memoria o il server semplicemente non è abbastanza grande (non ha abbastanza RAM) per eseguirlo. È un server root non gestito, quindi netcup è responsabile solo dei problemi tecnici, non dei programmi che esegui su di esso o delle risorse di cui hanno bisogno.
È chiaro che il tuo RS 2000 non riceverà un singolo byte in più di 16 GB di RAM, nemmeno per un nanosecondo. Forse creare uno swapfile o una partizione di swap potrebbe aiutare a mantenere in vita il tuo processo, dipende da quanta RAM è realmente necessaria e da quanto è cruciale la velocità di elaborazione. Alla fine potrebbe essere meglio usare un server con più RAM.
Thanks... yes.. that's definitely it.... thanks for the reply
-
generally computers when they have to process a process process the process in the potential they have and do not close it
Ähm, nope. This task usually falls to the OS Kernel and between Windows, Unix and Linux this is handled differently.
For Linux there are different settings on how the Kernel behaves - and for every system ressource except for RAM the kernel scheduler takes care of organizing the resources and assigning a processor to your process.
With RAM it is more complicated, since the Kernel has other means of organising more available RAM (Swap, closing buffers, killing other processes etc.).
Usually RAM is requested by calling the malloc syscall which reserves RAM, but also increases the pressure on the kernel to free more memory.
At some point the Kernel kills your process, since it is using a big chunk of memory. (And the OOM Killer often does not behave optimal / has bugs).
What kind of validator node are you running? If this is something for the blockchain, often a part of the chain has to be loaded into memory, for which 16 GiB is far from enough...
