Errors, bugs, questions - page 2692
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
Then I don't know...
Here I don't know... paid for an inexpensive VPS and 0 use, VPS from MQ doesn't give such a ping to this broker by the way, otherwise I might have been happy to use it.
Here I don't know... I pay for an inexpensive VPS and it does no good, VPS from MQ does not give such a ping to this broker, by the way, otherwise I might have been happy to use it.
Renat once explained that servers regulate the load on themselves in this way.
I.e., they fob clients off to other outlets if they themselves are overloaded.
Perhaps an automatic software switch to the right server would help.
Renat once explained that servers regulate the load on themselves in this way.
That is, they will kick clients to other locations if they are overloaded themselves.
Perhaps an automatic software switch to the right server would help.
Heh heh, wrong server implementation...
That is, the server doesn't scale vertically, but scales horizontally joined into a cluster, and stupidly redirects the client to another server.
Not the right implementation for this case where ping is an important priority.
Renat once explained that servers regulate the load on themselves in this way.
That is, they will kick clients to other locations if they are overloaded themselves.
Perhaps an automatic software switch to the right server would help.
Dear developers, please pay close attention, the problem is very serious.
I'll give you the logs in my personal if you need them.
Not the right implementation for this case, where the important priority is ping.
What joy is there in a 1ms ping if the server is so busy that there is a 50ms queue?
Not that I'm defending the current implementation, just reasoning.
What joy is there in a 1ms ping if the server is so busy that there is a 50ms queue?
Not that I'm defending the current implementation, just reasoning.
Too bad if the mt5 server can't cope.... We will have to move to a platform that can handle it.
Take a look at the terminal logs, please.
All reconnections, reasons for disconnection and network cluster rescans are shown there.
Re-connections are automatically evaluated every few hours, not every few seconds. Otherwise terminals would make a round robin of constant hopping between servers.
What joy is there in a 1ms ping if the server is so busy that there is a 50ms queue?
Not that I'm defending the current implementation, just reasoning.
In order to ensure that the server is not fully loaded, it is necessary to build a vertical scaling scheme for the server.
That is, during the peak load on the server, for example at 80% load, iron power is automatically added.
The provider has a lot of this power, and it is not realistic to choose the capacity of the provider by one or more servers.
When the peak load on the server decreases, then the added capacity of iron is reduced to the previous parameters.
With this scaling, there is no client redirect to another server.
A redirect to another server is an emergency case, if the server will be down or some failure, etc.
And MQ probably built their implementation on horizontal scaling, which is designed for web servers, where the redirect to another server is not critical.
when the text is under a grey (other colours are fine) rectangle, there is "erosion" of outline and colour
larger text:
Take a look at the terminal logs, please.
All reconnections, reasons for disconnection and network cluster rescans are shown there.
Thereconnection is automatically evaluated every few hours, not every few seconds. Otherwise the terminals would make a round of constant hopping between servers.
Yes, there is a loss of connection to server 1, yes, connected to server 3, but why it doesn't automatically connect to server 1 again is unclear.
In any case, modern algotrading is not trading once a quarter and it would be acceptable to have a loss of connection for a few hours. when server 3 is connected, the terminal freezes and the quotes are not updated for a few minutes.
maybe too big difference in server and client builds is critical? broker has server 2280, my terminal has 2363.