5 registered members (AndrewAMD, ozgur, Ayumi, 2 invisible),
690
guests, and 10
spiders. |
Key:
Admin,
Global Mod,
Mod
|
|
|
Re: Dukascopy plugin (yet another version)
[Re: ags]
#474969
11/15/18 14:31
11/15/18 14:31
|
Joined: Jan 2017
Posts: 95 Bulgaria
kvm
OP
Junior Member
|
OP
Junior Member
Joined: Jan 2017
Posts: 95
Bulgaria
|
I can't comment on how Z12 behaves, but Zorro saves all trades in Data/*.trd files. And after a disconnect it tries to resume them. You should see something like this in the message window:
TEST compiling...........
!OK: JVM initialized
Login 0 vDukascopy-1.2.. at UTC 11-15 14:22
!GBP/USD: 1.27997 0.00012 1
Trade: TEST GBP/USD 2018-11-15
[GBP/USD::L9184] - resumed (1 lots)
and here is the test script:
function run() {
LookBack = 0;
BarPeriod = 5;
if(Bar == 1) enterLong();
}
|
|
|
Re: Dukascopy plugin (yet another version)
[Re: kvm]
#474989
11/16/18 07:20
11/16/18 07:20
|
Joined: Oct 2018
Posts: 90 Germanien
ags
Junior Member
|
Junior Member
Joined: Oct 2018
Posts: 90
Germanien
|
Hi kvm,
sorry, i have to express my question a little different, more positive.
During the 2 weeks period live with vDukascopy-1.1 i disconnected and reconnected a few times without any problem, all open trades where resumed perfectly.
It was only the last time that Zorro wasn't finding the open positions and did not resumed the trades, i guess because i changed the leverage of the account having open positions, but i'm not sure why.
Nevertheless, after closing all positions manually in JForex, yesterday i started Z12 again in Dukascopy and it continues working good.
I just get some communication messages from the API a few times during the day, also during the two weeks period it was connected like these:
!ERROR Primary session ping failed: java.until.concurrent.TimeoutException: [DDS2 Standalone Transport Client] Timeout while waiting for response, 10000, syncRec...
!ERROR Secondary session (and the same as with Primary)...
!ERROR An existing connection was forcibly closed by the remote host java.io.IOException: An existing connection was forcibly closed by the remote host.
But never got "operative" problems, all runs continuosly but just get these messages a few times during the day, any idea why ?? Can i just ignore them??
Thanks again for the plugin, testing live in the 3rd week.
|
|
|
Re: Dukascopy plugin (yet another version)
[Re: ags]
#475130
11/24/18 09:33
11/24/18 09:33
|
Joined: Oct 2018
Posts: 90 Germanien
ags
Junior Member
|
Junior Member
Joined: Oct 2018
Posts: 90
Germanien
|
I also tried to reconnect from my PC but got the same result, Zorro startet trading without finding the open trades...
For this pat at least, i understand zorro stores the open trades in files Data/*.trd; and read these files to resume trades. So if you have a breackdown or try to resume from another environment it won't work because it does not has the information on open trades...
|
|
|
Re: Dukascopy plugin (yet another version)
[Re: ags]
#475151
11/26/18 09:35
11/26/18 09:35
|
Joined: Jan 2017
Posts: 95 Bulgaria
kvm
OP
Junior Member
|
OP
Junior Member
Joined: Jan 2017
Posts: 95
Bulgaria
|
Hi kvm,
for that i need to learn a lot of Java... for the time being is only a link language for me...
Would be interesting since i really get some stability problems. In my 4th week live trading with Z12 in Dukascopy i got some disturbing connectivity issues. At some point i desconnected Z12 having open trades (dieconnected without closing the trades of course), and at reconnection the open trades were found and reconnected, no problem. At other 2 experiences and after some weird disconnection Zorro reconnected but did not found the old trades, which is fundamental for Z12 kind of longer time period trades scripts. I had to close them manually from the JForex platform from Dukascopy.
I'm attaching the log fle from today only, you may see the error messages and that 2 trades where open. Then wen reconnecting Zorro was unable to find the open trades.
Any idea ??
By the way, I'm connecting Zorro from an EC2 AWS instance, that also gives me some troubles from time to time it kind of freezes speed, getting really really slow... could this be a problem?
I also tried to reconnect from my PC but got the same result, Zorro startet trading without finding the open trades... Hi ags, Examining the log shows several timeout issues which can be releated with EC2 AWS slowliness, but the thing that bothers me the most is "OutOfMemoryError: unable to create new native thread". This means that Java VM is out of resources (threads) so it may hang and become unresponsive and it is due to unfreed resources inside the Dukascopy JForex library: [DDS2 Standalone Transport Client] Event did not execute in timeout time [1000] and is still executing, CRITICAL EXECUTION WAIT TIME: 5453ms, possible application problem or deadLock
The only thing that I can do is to keep the plugin in sync with the latest version of the library and hope the bugs will be fixed then. And one more thing: the best way to restart the plugin is to restart the Zorro itself, i.e the Java VM will be restarted too.
|
|
|
Re: Dukascopy plugin (yet another version)
[Re: ags]
#475152
11/26/18 09:44
11/26/18 09:44
|
Joined: Jan 2017
Posts: 95 Bulgaria
kvm
OP
Junior Member
|
OP
Junior Member
Joined: Jan 2017
Posts: 95
Bulgaria
|
I also tried to reconnect from my PC but got the same result, Zorro startet trading without finding the open trades...
For this pat at least, i understand zorro stores the open trades in files Data/*.trd; and read these files to resume trades. So if you have a breackdown or try to resume from another environment it won't work because it does not has the information on open trades... Exactly! I hope to start using the plugin next week with a real account, so you won't be a single tester anymore.
|
|
|
Re: Dukascopy plugin (yet another version)
[Re: kvm]
#475167
11/26/18 12:21
11/26/18 12:21
|
Joined: Oct 2018
Posts: 90 Germanien
ags
Junior Member
|
Junior Member
Joined: Oct 2018
Posts: 90
Germanien
|
Hi kvm,
thank you for your efforts with the plugin, let's try to collaborate to get this done!
Since yesterday and with the idea to break down issues, i actually have two Z12 Zorro instances working on two different live accounts from Dukascopy, in two different AWS instances.
#1 in a T2.micro instance, the one you already got the log file #2 in a M3.medium instance, which has a complete different profile and should cope with 10 Zorros simultaniusly (at least).
Just according to Murphy, today I got some troubles with the M3.medium isntance as follows:
Weekend ends on 25.11. 23:00 [677: Mon 18-11-26 00:22] 11174.54/11227.7911174.54/11203.59 -6.18 [678: Mon 18-11-26 04:22] 11203.59/11242.0911200.49/11238.89 -6.18 [679: Mon 18-11-26 08:22] 11238.89/11339.4211234.99/11332.39 -1.48 Enter 1L UK100 Stop 568.7 Trail 568.7 at 08:22:01 {UK100:HU:l7901} Long 1@7024.03 Risk 627 ts at 08:22:01 Stop 6455 Trail 7593 TP 0.0 Com 0000 Mrg 0000 Net 0 Enter 1L UK100 Stop 0.25747 at 08:22:01 !ERROR java.util.concurrent.ExecutionException: com.dukascopy.api.JFException: state is CANCELED
Caused by: com.dukascopy.api.JFException: state is CANCELED
!java.util.concurrent.ExecutionException !com.dukascopy.api.JFException: state is CANCELED Error 075: (UK100:NET:L) - can't open 1@7024 Stop 1756 at 08:22:01 UK100 Pool: 0 Virtual: 1
I'll give you all messages i get from the plugin from these two different instances. Hopefully this helps you in the development.
According the jcl's team, they have good experience with 3 Zorro instances working on a T2.micro, but not with Dukascopy and not with a java based plugin I suppose.
Last edited by ags; 11/26/18 12:23.
|
|
|
|