error - Connection reset by peer - IOException

[jkelany]

Hi,
I have issue in the service it give me "error - Connection reset by peer - IOException" some times not received any signal and some time drop some

my debug cases for this issues

• When i tried to restart the service it back to work well
• When the service totally hung i tried to test if the port is open or not i found that the log not recived any thing on this port
• My server is centos 7 and i give Java heap 16GB to run with it and i run it with 8300 tracker re-translated from wialon server

my warpper config file for the java heap memory

# Initial Java Heap Size (in MB)
wrapper.java.initmemory=0

# Maximum Java Heap Size (in MB)
wrapper.java.maxmemory=0


wrapper.java.additional.6=-Xmx16000m
wrapper.java.additional.7=-Xms12000m

this is sample of my log

2016-04-09 14:53:51  WARN: [6547F920] error - Value cannot be null. - QueryException (... < DataManager:152 < ConnectionManager:111 < *:72 < MainEventHandler:73 < ...)
2016-04-09 14:53:51  INFO: [13220B7F] id: null, time: 2016-04-09 14:37:09, lat: 24.78612, lon: 46.68182, speed: 0.0, course: 0.0
2016-04-09 14:53:51 DEBUG: [CE05E90B: 12056 < 10.10.20.38] HEX: 302e3030303030303b32313b4e413b303b4e413b4e413b4e413b626174746572795f6368617267653a313a300d0a
2016-04-09 14:53:51 DEBUG: [13220B7F: 12056 > 10.10.20.38] HEX: 23414423310d0a
2016-04-09 14:53:51  WARN: [A53E543D] error - Connection reset by peer - IOException (...)
2016-04-09 14:53:51  INFO: [A53E543D] disconnected
2016-04-09 14:53:51  WARN: [13220B7F] error - ClosedChannelException (... < WialonProtocolDecoder:65 < *:145 < ExtendedObjectDecoder:59 < ...)
2016-04-09 14:53:51 DEBUG: [8E68F038: 12056 > 10.10.20.38] HEX: 23414c23310d0a
2016-04-09 14:53:51  WARN: [99A2A008] error - Value cannot be null. - QueryException (... < DataManager:152 < ConnectionManager:111 < *:72 < MainEventHandler:73 < ...)
2016-04-09 14:53:51  WARN: [AFEF4169] error - Connection reset by peer - IOException (...)
2016-04-09 14:53:51  INFO: [AFEF4169] disconnected
2016-04-09 14:53:51  WARN: [5F2FD395] error - Value cannot be null. - QueryException (... < DataManager:152 < ConnectionManager:111 < *:72 < MainEventHandler:73 < ...)
2016-04-09 14:53:51  WARN: [8E68F038] error - ClosedChannelException (... < WialonProtocolDecoder:65 < *:132 < ExtendedObjectDecoder:59 < ...)
2016-04-09 14:53:51  WARN: [E5155804] error - Connection reset by peer - IOException (...)
2016-04-09 14:53:51  WARN: [87FE6E2D] error - Connection reset by peer - IOException (...)
2016-04-09 14:53:51  INFO: [E5155804] disconnected
2016-04-09 14:53:51  INFO: [87FE6E2D] disconnected

any help please, thanks

[tananaev]

"Connection reset by peer" basically means that device disconnected from server.

[jkelany]

I know that this issue if because the device disconnected from server but in this situation i tried to test my port via http://test.gpsgate.com/FirewallTest/index.aspx
to test if the port open or not, the tracker-server.log not received any thin, but in normal case when it tried to test the port is open or not the tracker-server.log received the signal

[tananaev]

What kind of signal do you usually receive in the log when you test port?

[jkelany]

This is the signal when i tested the port connection

2016-04-09 15:41:35  INFO: [89F82020] connected
2016-04-09 15:41:35 DEBUG: [89F82020: 12056 < 10.10.21.193] HEX: 48656c6c6f2066726f6d204770734761746521
2016-04-09 15:41:35  INFO: [89F82020] disconnected

[tananaev]

OK, so you don't see it anymore? Are you using UDP? What version of Traccar are you using?

[jkelany]

• Yes when the server is hung i don't see this message any more
• We are using TCP not UDP
• I use the last release Version: 3.4 and also i have complied the last source code and i have the same issue

[tananaev]

How do you know that the server is hung and it's not some network issue?

[jkelany]

• The re-translated signals is in the same local network and when i tested the port it's not received the signal that is mentioned before
• Two servers ping each other
• Tried to check if the port is still open

[root@localhost bin]# netstat -l | egrep 12056
tcp6      51      0 [::]:12056              [::]:*                  LISTEN

[tananaev]

If you give me access to the server I can check it. My email address is in profile.

[jkelany]

I have sent you an email with the server access

[dronsv]

Anton, I confirm the problem for GPS103 protocol. I have this problem on my development version and on your test server 3. It's something magical in pipeline. Pass StandardLoggingHandler but not switches to addSpecificHandlers added Handlers.
Device now registered on your test server3 by my email. Please don't play with control :)

[jkelany]

Anton first of all thanks for you help

• I have changed my Java version from OpenJDK to Oracle JDK it's really cases the problem i think i faced this problem before but i missed
• Also i have increase open files limit on Linux to fit my 8300 device as my limit was 4096 that not suitable to my devices connections (as Anton told me and this also fixed the issue)

Thanks