Forum Discussion

StanI_87551's avatar
StanI_87551
Icon for Nimbostratus rankNimbostratus
Jun 08, 2015

Chrome (ERR_CONNECTION_RESET)

I have a developer who is experiencing an issue only with Chrome & here is his description:

 

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

 

I believe the issue is google chrome is throwing a connection reset error: (ERR_CONNECTION_RESET) with the F5 load balancer. I have not seen this error occur with Firefox or IE 11.

 

Which leads me to think it is only related to chrome.

 

I believe the issue is with chrome and the F5, because I can connect directly to the server using without issue.

 

In order to reproduce this error, I clear all history/cache/cookies etc. in chrome, and connect to our load balancer site Spotfire.aviallinc.com -- It connects without issue the majority of the time, however if I repeat this process of clearing history and re-hitting the website this error will eventually appear, usually within 5 or 6 cycles. -- The error will resolve itself, after a few seconds, sometimes within 1 sometimes up to 10, but in general it will correct itself and repoint.

 

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

 

The server for the load-balancer is a new version so I’m assuming the RFC5746 already exists? https://support.f5.com/kb/en-us/solutions/public/13000/800/sol13860.html

 

application delivery
LTM

7 Replies

Sort By
  • nitass's avatar
    nitass
    Icon for Employee rankEmployee
    Jun 11, 2015

    i think it may be useful if you can capture packet when issue is happening.

     tcpdump -nni 0.0:nnn -s0 -w /var/tmp/output.pcap host x.x.x.x or host y.y.y.y or host z.z.z.z -v

x.x.x.x is client ip
y.y.y.y and z.z.z.z are pool member ip
  • StanI's avatar
    StanI
    Icon for Nimbostratus rankNimbostratus
    Jun 18, 2015

    Ran the trace looked @ the pcap file and most of the traffic is between F5 Float IP on the vlan the servers are on and the server (secondary server was Offline (Enabled)) so this is a single threadded VS_ to Pool_ to Server on port 443.

    Virtual Server:

    vs_spotfire.aviallinc.com_80

    Destination: 172.18.19.30:80

    with iRule = http_to_https 
        when HTTP_REQUEST {
    HTTP::redirect https://[HTTP::host][HTTP::uri]
    }

    Virtual Server:

    vs_spotfire.aviallinc.com

    Destination: 172.18.19.30:443

    Pool: spotfire.aviallinc.com

    Pool Member: wspw-prd-1:443

    Pool Member: wspw-prd-2:443
  • StanI's avatar
    StanI
    Icon for Nimbostratus rankNimbostratus
    Jun 24, 2015

    Nothing that I can see, the capture from the F5 shows lots of duplicates

     

    62% of the tcp errors = Duplicate Acks

     

    22% of the tcp errors Retransmissions

     

    13% Resets

     

  • boneyard's avatar
    boneyard
    Icon for MVP rankMVP
    Jun 24, 2015

    i would be worried about the 111 resets, specially if that happens we you see the issue occur, do they come from the client or the big-ip? you will have to bring those two things together and see if you can find a common cause. the logging shows nothing at these moments?