agents wont connect.
i figured out what was wrong, the us server must have been down for maintenance, the european servers were showing in mt5, but the us ones were missing, they are back now and the agents have reconnected to the network.
i could have saved sometime had i known this, there is no log saying the servers were down or anything, so it looked like it was a problem on my end, when it was actually mql5 that had the problem.
it wasnt actually a heating problem the average temp of the pc is about 40c at full load, it was a computer running out of resources problem, it would use memory till thecomputer would shutdown with an error about lack of memory. any suggestion on how to fix this?? ill be putting another 32 gigs into the computer next week.
the problem with agents not connecting should be fixed now, the us server is back. not sure if its possible to run vpn with the agents, but if it is i would suggest a european location, they seem to get more work/users.
i figured out what was wrong, the us server must have been down for maintenance, the european servers were showing in mt5, but the us ones were missing, they are back now and the agents have reconnected to the network.
i could have saved sometime had i known this, there is no log saying the servers were down or anything, so it looked like it was a problem on my end, when it was actually mql5 that had the problem.
it wasnt actually a heating problem the average temp of the pc is about 40c at full load, it was a computer running out of resources problem, it would use memory till thecomputer would shutdown with an error about lack of memory. any suggestion on how to fix this?? ill be putting another 32 gigs into the computer next week.
the problem with agents not connecting should be fixed now, the us server is back. not sure if its possible to run vpn with the agents, but if it is i would suggest a european location, they seem to get more work/users.
can you point me or show me a screen shot in MT5 where i assign the ram??
atm they are using 10+ gigs each.
can you point me or show me a screen shot in MT5 where i assign the ram??
atm they are using 10+ gigs each.
yea thats the problem i am having, there seems to be a memory leak, the memory gets used then never freed, so next time the agent runs it uses another 3g of ram, even though it has 3g assigned to it in idle state.
if you look at the attached jpg, the 2 agents at the top of the window are idle, but they are using 2gigs+ each, one is using almost 5.5gigs that wasnt freed up when running a test.
i think its adding up and adding up till the computer is out of resources, this just started happening recently, ive run the cloud agent thing for almost a year now and never seen anything close to this. as of the 4040 update though, the computer hasnt been handing mql5 cloud very well, it was just formatted and installed with all new drivers etc.
yea thats the problem i am having, there seems to be a memory leak, the memory gets used then never freed, so next time the agent runs it uses another 3g of ram, even though it has 3g assigned to it in idle state.
if you look at the attached jpg, the 2 agents at the top of the window are idle, but they are using 2gigs+ each, one is using almost 5.5gigs that wasnt freed up when running a test.
i think its adding up and adding up till the computer is out of resources, this just started happening recently, ive run the cloud agent thing for almost a year now and never seen anything close to this. as of the 4040 update though, the computer hasnt been handing mql5 cloud very well, it was just formatted and installed with all new drivers etc.
When that happen to you again look your SSD... That happen to me each day. It look like the agents are uncompressing (unzipping) massive data in the SSD (over 250GB). Also you are right, memory is freed in the next job, and the SSD too. (or maybe the next job is the same job). Because it took more than an hour to fill my SSD with 240GB (They were in no hurry with this).
Look the images i attached in this post
https://www.mql5.com/es/forum/462700
Best Regard!!
- 2024.02.19
- www.mql5.com
yea thats the problem i am having, there seems to be a memory leak, the memory gets used then never freed, so next time the agent runs it uses another 3g of ram, even though it has 3g assigned to it in idle state.
if you look at the attached jpg, the 2 agents at the top of the window are idle, but they are using 2gigs+ each, one is using almost 5.5gigs that wasnt freed up when running a test.
i think its adding up and adding up till the computer is out of resources, this just started happening recently, ive run the cloud agent thing for almost a year now and never seen anything close to this. as of the 4040 update though, the computer hasnt been handing mql5 cloud very well, it was just formatted and installed with all new drivers etc.
As I've said before, they haven't made significant changes to any of this for a very long time (years). Any changes in behavior that you're experiencing have to do with the type of jobs users are asking the MQL5 Cloud Network to process. You should be really happy that the biggest agent you're looking at is using only 5.5 GB of RAM—take a look at the attached screenshot to get an idea of what it could climb to! (And note the date in the screenshot—just to further confirm that this is not new behavior.) If your system is regularly running out of memory and crashing, add more RAM or reduce the number of agents running on your PC. And no, this is not happening due to a "memory leak" in the tester application; it is allocating memory based on the requirements of the incoming jobs (which are often coded poorly), and at the finish of a job pass, it leaves the agent set up so it will be ready should another pass from that same job come your way. Eventually, it will shut down the tester and automatically free all the allocated RAM and delete all the temporary files. Restart your PC or restart the individual agents if you wish to free this memory on demand, at the cost of forfeiting your place in the que to get future passes for that job.
As I've said before, they haven't made significant changes to any of this for a very long time (years). Any changes in behavior that you're experiencing have to do with the type of jobs users are asking the MQL5 Cloud Network to process. You should be really happy that the biggest agent you're looking at is using only 5.5 GB of RAM—take a look at the attached screenshot to get an idea of what it could climb to! (And note the date in the screenshot—just to further confirm that this is not new behavior.) If your system is regularly running out of memory and crashing, add more RAM or reduce the number of agents running on your PC. And no, this is not happening due to a "memory leak" in the tester application; it is allocating memory based on the requirements of the incoming jobs (which are often coded poorly), and at the finish of a job pass, it leaves the agent set up so it will be ready should another pass from that same job come your way. Eventually, it will shut down the tester and automatically free all the allocated RAM and delete all the temporary files. Restart your PC or restart the individual agents if you wish to free this memory on demand, at the cost of forfeiting your place in the que to get future passes for that job.
Wow!! That's awesome!!
I don't understand why it need sooo disk space!! How many SSDs have they broken?
I have a lot of RAM, but they never use it all. 28 agents and the maximum RAM used has been about 90GB
But such intensive use of the SSD worries me. The Achilles' heel of SSDs is precisely the writing of data.
If they write one or two TB per day the SSD will not live more than a year.
I tried to install the agents on a secondary SSD different from the operating system. But they didn't work. They immediately disconnected. (i don't want to reinstall the system)
I hope Metaquotes engineers realize that destroying SSD is not profitable.
Best regards!!
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
You agree to website policy and terms of use
im not sure whats up, i formatted my computer last week, and now i cant get the agents to connect to the cloud....
CS 0 13:17:20.605 Startup access rights to common directory successfully checked
CS 0 13:17:20.607 Startup service start initialized
CS 0 13:17:20.607 Startup create startup thread
CS 0 13:17:20.607 Startup thread successfully created
CS 0 13:17:20.607 Startup MetaTester 5 build 4041 (24 Oct 2023)
CS 0 13:17:20.607 Startup Windows 10 build 19045, 32 x Intel Xeon E5-2667 v4 @ 3.20GHz, AVX2, 42 / 47 Gb memory, 664 / 930 Gb disk, UAC, admin, GMT-5
CS 0 13:17:20.609 Server MetaTester 5 started on 0.0.0.0:2000
CS 0 13:17:20.616 Network server agent1.mql5.net ping 103 ms
CS 0 13:17:20.616 Network server agent2.mql5.net ping 29 ms
CS 0 13:17:20.616 Network server agent3.mql5.net ping 139 ms
CS 0 13:17:20.616 Network server agent4.mql5.net ping 29 ms
CS 0 13:17:20.618 Startup MQL5 account: cloud network mode is on
CS 0 13:17:20.618 Startup initialization finished
i dont see anything wrong, my setup works fine otherwise, just agents wont connect to the cloud to sell resources...
is there a log file somewhere with more information as to whats going on?
currently they have been in a disconnected state for over a week. no matter how many times i reinsatall the agents, reboot my computer, reboot my modem, they are refusing to connect.
since the 4040 update i have had nothing but problems with my agents. first they used more resources then were available, i thought something was wrong with my setup so i formatted, now they are refusing to work at all.
before the format they would use resources till the computer crashed.
the PR of my agents is 150-156.