I run a 32 gig server which only has 4 cores. My hidden service is running two workstations and obviously one gateway.
If I select two cores to run on one hidden service 1 on the gateway and 1 on the other workstation will my host machine be able to handle the workload, bearing in mind it only has 4 cores total?
Might work but its a close call. If anything stresses all cores at once it may starve your host. You should never over commit CPU cores however because it will cause system lockups.
So even if I added an extra two or four cores, if I tried to get the web server to use 2 or 3 of those it would risk locking up the system?
The reason I ask is because my nginx conf asks for how many worker process I have, and 1 is the default but I worry if the website receives a lot of traffic it may slow down the visitors.
Do you think 4 cores is not enough to run two workstations then? Even if I leave the default setting at 1.
Good day,
Giving you a “correct” answer is impossible in this case to be honest. There are so many reasons which determine how many machines you should be able to run parallel without issue that focusing on pure core count is prohibitive.
Things like the architecture used, SMT support, support for enhanced acceleration features, clock frequency, pipeline depth, instruction sets, the frequency at which your CPU runs (though like the core count, isn’t as important as it may seem), optimization and generally how well your host OS has been set-up all can have a massive impact on this.
This can go such a long way that, in certain scenarios, I am able to run multiple machines (1 Gateway and 3 Workstations) on a low voltage, low clockspeed and low TDP Skylake Dual Core CPU without many issues.
So, in conclusion, due to the complexity of processor technology, your question can not be answered that easily.
Have a nice day,
Ego
So even if I added an extra two or four cores, if I tried to get the web server to use 2 or 3 of those it would risk locking up the system?
If it works great but I would do a stress test before deploying into production.
Do you think 4 cores is not enough to run two workstations then? Even if I leave the default setting at 1.
Should be more than enough but understand I adjust settings according to most common uses. I have no idea how busy your service is or will become and really that is something you’ll have to plan for.