Hi All,
In our setup running kamailio (alpine image) in stateless mode and have following TCP config set. After client registers to the backend see two TCP connections created - one towards the client and the other towards the backendserver. If there's a abrupt network disconnect either at the client side or at the backend side, do not see TCP reset happening from kamailio as per the config. Per below config expected after 120seconds( = [tcp_keepidle] + ( [tcp_keepintvl * tcp_keepcnt] ) instead it keeps resending TCP keepalive packets at varying intervals for varying duration length(like 6mins, 14mins..) before resetting the TCP connections. Also want to mention that we are not using TCPOPS module here.
I am trying to achieve consistent duration for TCP reset in kamailio. Could you please point me if I have missed anything here?
Setup:
Client ---- Kamailio (stateless mode) ---- Backend server
TCP Config:
tcp_connection_lifetime = 3600 tcp_accept_aliases = no tcp_async = yes tcp_connect_timeout = 10 tcp_crlf_ping = yes tcp_keepalive = yes tcp_keepcnt = 3 tcp_keepidle = 30 tcp_keepintvl = 30 # Total 2 mins - 30+(30*3)=120 sec tcp_linger2 = 30 tcp_rd_buf_size = 65536 tcp_send_timeout = 10
thanks, raj
Hello,
can you reproduce the case and then get the output of the next commands:
kamctl rpc core.tcp_list
kamctl rpc core.tcp_options
kamctl rpc core.tcp_info
Cheers, Daniel
On 06.06.19 18:53, R, Rajkumar (Raj) wrote:
Hi All,
In our setup running kamailio (alpine image) in stateless mode and have following TCP config set. After client registers to the backend see two TCP connections created – one towards the client and the other towards the backendserver. If there’s a abrupt network disconnect either at the client side or at the backend side, do not see TCP reset happening from kamailio as per the config. Per below config expected after 120seconds( = [tcp_keepidle] + ( [tcp_keepintvl * tcp_keepcnt] ) instead it keeps resending TCP keepalive packets at varying intervals for varying duration length(like 6mins, 14mins..) before resetting the TCP connections. Also want to mention that we are not using TCPOPS module here.
I am trying to achieve consistent duration for TCP reset in kamailio. Could you please point me if I have missed anything here?
*/Setup:/*
Client ---- Kamailio (stateless mode) ---- Backend server
*TCP Config:*
* *
*tcp_connection_lifetime = 3600*
tcp_accept_aliases = no
tcp_async = yes
tcp_connect_timeout = 10
tcp_crlf_ping = yes
*tcp_keepalive = yes*
*tcp_keepcnt = 3*
*tcp_keepidle = 30*
*tcp_keepintvl = 30 # Total 2 mins - 30+(30*3)=120 sec*
tcp_linger2 = 30
tcp_rd_buf_size = 65536
tcp_send_timeout = 10
thanks,
raj
Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users
Hi Daniel,
Please find attached stats output captured every 2 seconds for the entire duration of test (i.e., Start of registration from client till the connection towards client is reset by kamailio after 17mins approximately). I had used 'kamcmd' instead of 'kamctl' for same command arguments.
Setup:
10.27.117.55 (Client) --- Kamailio (10.1.4.55) ---- Backend server (10.2.8.70)
Summary of events executed/observed during the test:
* Once the client is registered successfully to the backend, turned off WiFi connection at the client side (time 13:07.05 UTC 2019). * After OPTIONS msg time out for response from the client, backend closes the TCP connection towards kamailio ( around time 13:10:39), * TCP connection towards client side is reset at 13:24:30 (i.e., 17 minutes since the WiFi is turned off) * In wireshark, see re-transmission of a TCP packet from kamailio for about 15 minutes until connection release.
thanks, raj
From: Daniel-Constantin Mierla miconda@gmail.com Sent: Friday, June 7, 2019 6:14 PM To: Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org; R, Rajkumar (Raj) rajkumaradass@avaya.com Subject: Re: [SR-Users] TCP reset behavior
Hello,
can you reproduce the case and then get the output of the next commands:
kamctl rpc core.tcp_list
kamctl rpc core.tcp_options
kamctl rpc core.tcp_info
Cheers, Daniel On 06.06.19 18:53, R, Rajkumar (Raj) wrote: Hi All,
In our setup running kamailio (alpine image) in stateless mode and have following TCP config set. After client registers to the backend see two TCP connections created - one towards the client and the other towards the backendserver. If there's a abrupt network disconnect either at the client side or at the backend side, do not see TCP reset happening from kamailio as per the config. Per below config expected after 120seconds( = [tcp_keepidle] + ( [tcp_keepintvl * tcp_keepcnt] ) instead it keeps resending TCP keepalive packets at varying intervals for varying duration length(like 6mins, 14mins..) before resetting the TCP connections. Also want to mention that we are not using TCPOPS module here.
I am trying to achieve consistent duration for TCP reset in kamailio. Could you please point me if I have missed anything here?
Setup:
Client ---- Kamailio (stateless mode) ---- Backend server
TCP Config:
tcp_connection_lifetime = 3600 tcp_accept_aliases = no tcp_async = yes tcp_connect_timeout = 10 tcp_crlf_ping = yes tcp_keepalive = yes tcp_keepcnt = 3 tcp_keepidle = 30 tcp_keepintvl = 30 # Total 2 mins - 30+(30*3)=120 sec tcp_linger2 = 30 tcp_rd_buf_size = 65536 tcp_send_timeout = 10
thanks, raj
_______________________________________________
Kamailio (SER) - Users Mailing List
sr-users@lists.kamailio.orgmailto:sr-users@lists.kamailio.org
https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-usershttps://urldefense.proofpoint.com/v2/url?u=https-3A__lists.kamailio.org_cgi-2Dbin_mailman_listinfo_sr-2Dusers&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=lQLR6IfVOO2CWDL4iEDldpjFjZfxQFyfKeFd1pGZkoY&e=
--
Daniel-Constantin Mierla -- www.asipto.comhttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=2-7juQBEXrjA_PjfpiIZ0JvAZxWKngyC0vcVyP1GmUs&e=
www.twitter.com/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=_qINzlBw2yRKQ3ke2UwXsjTQ7lKWI25XpjOlzw9gPqA&e= -- www.linkedin.com/in/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=TsFAXfvNfu1k1Kw-F4937svidQSEfPMgK4VmorRPn54&e=
Did the shared tcp stats help? Awaiting your inputs on this.
thanks, raj
From: sr-users sr-users-bounces@lists.kamailio.org On Behalf Of R, Rajkumar (Raj) Sent: Friday, June 7, 2019 10:44 PM To: miconda@gmail.com; Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org Subject: Re: [SR-Users] TCP reset behavior
Hi Daniel,
Please find attached stats output captured every 2 seconds for the entire duration of test (i.e., Start of registration from client till the connection towards client is reset by kamailio after 17mins approximately). I had used 'kamcmd' instead of 'kamctl' for same command arguments.
Setup:
10.27.117.55 (Client) --- Kamailio (10.1.4.55) ---- Backend server (10.2.8.70)
Summary of events executed/observed during the test:
* Once the client is registered successfully to the backend, turned off WiFi connection at the client side (time 13:07.05 UTC 2019). * After OPTIONS msg time out for response from the client, backend closes the TCP connection towards kamailio ( around time 13:10:39), * TCP connection towards client side is reset at 13:24:30 (i.e., 17 minutes since the WiFi is turned off) * In wireshark, see re-transmission of a TCP packet from kamailio for about 15 minutes until connection release.
thanks, raj
From: Daniel-Constantin Mierla <miconda@gmail.commailto:miconda@gmail.com> Sent: Friday, June 7, 2019 6:14 PM To: Kamailio (SER) - Users Mailing List <sr-users@lists.kamailio.orgmailto:sr-users@lists.kamailio.org>; R, Rajkumar (Raj) <rajkumaradass@avaya.commailto:rajkumaradass@avaya.com> Subject: Re: [SR-Users] TCP reset behavior
Hello,
can you reproduce the case and then get the output of the next commands:
kamctl rpc core.tcp_list
kamctl rpc core.tcp_options
kamctl rpc core.tcp_info
Cheers, Daniel On 06.06.19 18:53, R, Rajkumar (Raj) wrote: Hi All,
In our setup running kamailio (alpine image) in stateless mode and have following TCP config set. After client registers to the backend see two TCP connections created - one towards the client and the other towards the backendserver. If there's a abrupt network disconnect either at the client side or at the backend side, do not see TCP reset happening from kamailio as per the config. Per below config expected after 120seconds( = [tcp_keepidle] + ( [tcp_keepintvl * tcp_keepcnt] ) instead it keeps resending TCP keepalive packets at varying intervals for varying duration length(like 6mins, 14mins..) before resetting the TCP connections. Also want to mention that we are not using TCPOPS module here.
I am trying to achieve consistent duration for TCP reset in kamailio. Could you please point me if I have missed anything here?
Setup:
Client ---- Kamailio (stateless mode) ---- Backend server
TCP Config:
tcp_connection_lifetime = 3600 tcp_accept_aliases = no tcp_async = yes tcp_connect_timeout = 10 tcp_crlf_ping = yes tcp_keepalive = yes tcp_keepcnt = 3 tcp_keepidle = 30 tcp_keepintvl = 30 # Total 2 mins - 30+(30*3)=120 sec tcp_linger2 = 30 tcp_rd_buf_size = 65536 tcp_send_timeout = 10
thanks, raj
_______________________________________________
Kamailio (SER) - Users Mailing List
sr-users@lists.kamailio.orgmailto:sr-users@lists.kamailio.org
https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-usershttps://urldefense.proofpoint.com/v2/url?u=https-3A__lists.kamailio.org_cgi-2Dbin_mailman_listinfo_sr-2Dusers&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=lQLR6IfVOO2CWDL4iEDldpjFjZfxQFyfKeFd1pGZkoY&e=
--
Daniel-Constantin Mierla -- www.asipto.comhttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=2-7juQBEXrjA_PjfpiIZ0JvAZxWKngyC0vcVyP1GmUs&e=
www.twitter.com/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=_qINzlBw2yRKQ3ke2UwXsjTQ7lKWI25XpjOlzw9gPqA&e= -- www.linkedin.com/in/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=TsFAXfvNfu1k1Kw-F4937svidQSEfPMgK4VmorRPn54&e=
Hello,
I didn't get the time to look at. It was a long weekend in Germany, with Monday being public holiday. Not sure if I get to it today, but hope to get time in the next few days.
Cheers, Daniel
On 11.06.19 07:56, R, Rajkumar (Raj) wrote:
Did the shared tcp stats help? Awaiting your inputs on this.
thanks,
raj
*From:*sr-users sr-users-bounces@lists.kamailio.org *On Behalf Of *R, Rajkumar (Raj) *Sent:* Friday, June 7, 2019 10:44 PM *To:* miconda@gmail.com; Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org *Subject:* Re: [SR-Users] TCP reset behavior
Hi Daniel,
Please find attached stats output captured every 2 seconds for the entire duration of test (i.e., Start of registration from client till the connection towards client is reset by kamailio after 17mins approximately). I had used ‘kamcmd’ instead of ‘kamctl’ for same command arguments.
Setup:
10.27.117.55 (Client) --- Kamailio (10.1.4.55) ---- Backend server (10.2.8.70)
Summary of events executed/observed during the test:
- Once the client is registered successfully to the backend, turned off WiFi connection at the client side (time 13:07.05 UTC 2019).
- After OPTIONS msg time out for response from the client, backend closes the TCP connection towards kamailio ( around time 13:10:39),
- TCP connection towards client side is reset at 13:24:30 (i.e., 17 minutes since the WiFi is turned off)
- In wireshark, see re-transmission of a TCP packet from kamailio for about 15 minutes until connection release.
thanks,
raj
*From:*Daniel-Constantin Mierla <miconda@gmail.com mailto:miconda@gmail.com> *Sent:* Friday, June 7, 2019 6:14 PM *To:* Kamailio (SER) - Users Mailing List <sr-users@lists.kamailio.org mailto:sr-users@lists.kamailio.org>; R, Rajkumar (Raj) <rajkumaradass@avaya.com mailto:rajkumaradass@avaya.com> *Subject:* Re: [SR-Users] TCP reset behavior
Hello,
can you reproduce the case and then get the output of the next commands:
kamctl rpc core.tcp_list
kamctl rpc core.tcp_options
kamctl rpc core.tcp_info
Cheers, Daniel
On 06.06.19 18:53, R, Rajkumar (Raj) wrote:
Hi All, In our setup running kamailio (alpine image) in stateless mode and have following TCP config set. After client registers to the backend see two TCP connections created – one towards the client and the other towards the backendserver. If there’s a abrupt network disconnect either at the client side or at the backend side, do not see TCP reset happening from kamailio as per the config. Per below config expected after 120seconds( = [tcp_keepidle] + ( [tcp_keepintvl * tcp_keepcnt] ) instead it keeps resending TCP keepalive packets at varying intervals for varying duration length(like 6mins, 14mins..) before resetting the TCP connections. Also want to mention that we are not using TCPOPS module here. I am trying to achieve consistent duration for TCP reset in kamailio. Could you please point me if I have missed anything here? */Setup:/* Client ---- Kamailio (stateless mode) ---- Backend server *TCP Config:* * * *tcp_connection_lifetime = 3600* tcp_accept_aliases = no tcp_async = yes tcp_connect_timeout = 10 tcp_crlf_ping = yes *tcp_keepalive = yes* *tcp_keepcnt = 3* *tcp_keepidle = 30* *tcp_keepintvl = 30 # Total 2 mins - 30+(30*3)=120 sec* tcp_linger2 = 30 tcp_rd_buf_size = 65536 tcp_send_timeout = 10 thanks, raj _______________________________________________ Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org <mailto:sr-users@lists.kamailio.org> https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users <https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.kamailio.org_cgi-2Dbin_mailman_listinfo_sr-2Dusers&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=lQLR6IfVOO2CWDL4iEDldpjFjZfxQFyfKeFd1pGZkoY&e=>
-- Daniel-Constantin Mierla -- www.asipto.com https://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=2-7juQBEXrjA_PjfpiIZ0JvAZxWKngyC0vcVyP1GmUs&e= www.twitter.com/miconda https://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=_qINzlBw2yRKQ3ke2UwXsjTQ7lKWI25XpjOlzw9gPqA&e= -- www.linkedin.com/in/miconda https://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=TsFAXfvNfu1k1Kw-F4937svidQSEfPMgK4VmorRPn54&e=
Hello,
the connection state is OK in all outputs, so not seen as closed by Kamailio.
Can you get the output of the netstat for that connection after is supposed to be closed to see its state there?
Cheers, Daniel
On 11.06.19 12:17, Daniel-Constantin Mierla wrote:
Hello,
I didn't get the time to look at. It was a long weekend in Germany, with Monday being public holiday. Not sure if I get to it today, but hope to get time in the next few days.
Cheers, Daniel
On 11.06.19 07:56, R, Rajkumar (Raj) wrote:
Did the shared tcp stats help? Awaiting your inputs on this.
thanks,
raj
*From:*sr-users sr-users-bounces@lists.kamailio.org *On Behalf Of *R, Rajkumar (Raj) *Sent:* Friday, June 7, 2019 10:44 PM *To:* miconda@gmail.com; Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org *Subject:* Re: [SR-Users] TCP reset behavior
Hi Daniel,
Please find attached stats output captured every 2 seconds for the entire duration of test (i.e., Start of registration from client till the connection towards client is reset by kamailio after 17mins approximately). I had used ‘kamcmd’ instead of ‘kamctl’ for same command arguments.
Setup:
10.27.117.55 (Client) --- Kamailio (10.1.4.55) ---- Backend server (10.2.8.70)
Summary of events executed/observed during the test:
- Once the client is registered successfully to the backend, turned off WiFi connection at the client side (time 13:07.05 UTC 2019).
- After OPTIONS msg time out for response from the client, backend closes the TCP connection towards kamailio ( around time 13:10:39),
- TCP connection towards client side is reset at 13:24:30 (i.e., 17 minutes since the WiFi is turned off)
- In wireshark, see re-transmission of a TCP packet from kamailio for about 15 minutes until connection release.
thanks,
raj
*From:*Daniel-Constantin Mierla <miconda@gmail.com mailto:miconda@gmail.com> *Sent:* Friday, June 7, 2019 6:14 PM *To:* Kamailio (SER) - Users Mailing List <sr-users@lists.kamailio.org mailto:sr-users@lists.kamailio.org>; R, Rajkumar (Raj) <rajkumaradass@avaya.com mailto:rajkumaradass@avaya.com> *Subject:* Re: [SR-Users] TCP reset behavior
Hello,
can you reproduce the case and then get the output of the next commands:
kamctl rpc core.tcp_list
kamctl rpc core.tcp_options
kamctl rpc core.tcp_info
Cheers, Daniel
On 06.06.19 18:53, R, Rajkumar (Raj) wrote:
Hi All, In our setup running kamailio (alpine image) in stateless mode and have following TCP config set. After client registers to the backend see two TCP connections created – one towards the client and the other towards the backendserver. If there’s a abrupt network disconnect either at the client side or at the backend side, do not see TCP reset happening from kamailio as per the config. Per below config expected after 120seconds( = [tcp_keepidle] + ( [tcp_keepintvl * tcp_keepcnt] ) instead it keeps resending TCP keepalive packets at varying intervals for varying duration length(like 6mins, 14mins..) before resetting the TCP connections. Also want to mention that we are not using TCPOPS module here. I am trying to achieve consistent duration for TCP reset in kamailio. Could you please point me if I have missed anything here? */Setup:/* Client ---- Kamailio (stateless mode) ---- Backend server *TCP Config:* * * *tcp_connection_lifetime = 3600* tcp_accept_aliases = no tcp_async = yes tcp_connect_timeout = 10 tcp_crlf_ping = yes *tcp_keepalive = yes* *tcp_keepcnt = 3* *tcp_keepidle = 30* *tcp_keepintvl = 30 # Total 2 mins - 30+(30*3)=120 sec* tcp_linger2 = 30 tcp_rd_buf_size = 65536 tcp_send_timeout = 10 thanks, raj _______________________________________________ Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org <mailto:sr-users@lists.kamailio.org> https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users <https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.kamailio.org_cgi-2Dbin_mailman_listinfo_sr-2Dusers&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=lQLR6IfVOO2CWDL4iEDldpjFjZfxQFyfKeFd1pGZkoY&e=>
-- Daniel-Constantin Mierla -- www.asipto.com https://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=2-7juQBEXrjA_PjfpiIZ0JvAZxWKngyC0vcVyP1GmUs&e= www.twitter.com/miconda https://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=_qINzlBw2yRKQ3ke2UwXsjTQ7lKWI25XpjOlzw9gPqA&e= -- www.linkedin.com/in/miconda https://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=TsFAXfvNfu1k1Kw-F4937svidQSEfPMgK4VmorRPn54&e=
-- Daniel-Constantin Mierla -- www.asipto.com www.twitter.com/miconda -- www.linkedin.com/in/miconda
It remained in 'ESTABLISHED' state only. Please find netstat output below.
# netstat -natlp Active Internet connections (servers and established) Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name tcp 0 0 0.0.0.0:5060 0.0.0.0:* LISTEN 20/kamailio tcp 0 0 0.0.0.0:80 0.0.0.0:* LISTEN - tcp 0 0 0.0.0.0:411 0.0.0.0:* LISTEN - tcp 0 0 0.0.0.0:8443 0.0.0.0:* LISTEN - tcp 0 0 0.0.0.0:5056 0.0.0.0:* LISTEN 20/kamailio tcp 0 3777 10.1.4.55:5060 10.27.117.55:38115 ESTABLISHED 20/Kamailio
thanks, raj +918067153382
From: Daniel-Constantin Mierla miconda@gmail.com Sent: Wednesday, June 12, 2019 12:29 PM To: R, Rajkumar (Raj) rajkumaradass@avaya.com; Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org Subject: Re: [SR-Users] TCP reset behavior
Hello,
the connection state is OK in all outputs, so not seen as closed by Kamailio.
Can you get the output of the netstat for that connection after is supposed to be closed to see its state there?
Cheers, Daniel On 11.06.19 12:17, Daniel-Constantin Mierla wrote:
Hello,
I didn't get the time to look at. It was a long weekend in Germany, with Monday being public holiday. Not sure if I get to it today, but hope to get time in the next few days.
Cheers, Daniel On 11.06.19 07:56, R, Rajkumar (Raj) wrote: Did the shared tcp stats help? Awaiting your inputs on this.
thanks, raj
From: sr-users sr-users-bounces@lists.kamailio.orgmailto:sr-users-bounces@lists.kamailio.org On Behalf Of R, Rajkumar (Raj) Sent: Friday, June 7, 2019 10:44 PM To: miconda@gmail.commailto:miconda@gmail.com; Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.orgmailto:sr-users@lists.kamailio.org Subject: Re: [SR-Users] TCP reset behavior
Hi Daniel,
Please find attached stats output captured every 2 seconds for the entire duration of test (i.e., Start of registration from client till the connection towards client is reset by kamailio after 17mins approximately). I had used 'kamcmd' instead of 'kamctl' for same command arguments.
Setup:
10.27.117.55 (Client) --- Kamailio (10.1.4.55) ---- Backend server (10.2.8.70)
Summary of events executed/observed during the test:
* Once the client is registered successfully to the backend, turned off WiFi connection at the client side (time 13:07.05 UTC 2019). * After OPTIONS msg time out for response from the client, backend closes the TCP connection towards kamailio ( around time 13:10:39), * TCP connection towards client side is reset at 13:24:30 (i.e., 17 minutes since the WiFi is turned off) * In wireshark, see re-transmission of a TCP packet from kamailio for about 15 minutes until connection release.
thanks, raj
From: Daniel-Constantin Mierla <miconda@gmail.commailto:miconda@gmail.com> Sent: Friday, June 7, 2019 6:14 PM To: Kamailio (SER) - Users Mailing List <sr-users@lists.kamailio.orgmailto:sr-users@lists.kamailio.org>; R, Rajkumar (Raj) <rajkumaradass@avaya.commailto:rajkumaradass@avaya.com> Subject: Re: [SR-Users] TCP reset behavior
Hello,
can you reproduce the case and then get the output of the next commands:
kamctl rpc core.tcp_list
kamctl rpc core.tcp_options
kamctl rpc core.tcp_info
Cheers, Daniel On 06.06.19 18:53, R, Rajkumar (Raj) wrote: Hi All,
In our setup running kamailio (alpine image) in stateless mode and have following TCP config set. After client registers to the backend see two TCP connections created - one towards the client and the other towards the backendserver. If there's a abrupt network disconnect either at the client side or at the backend side, do not see TCP reset happening from kamailio as per the config. Per below config expected after 120seconds( = [tcp_keepidle] + ( [tcp_keepintvl * tcp_keepcnt] ) instead it keeps resending TCP keepalive packets at varying intervals for varying duration length(like 6mins, 14mins..) before resetting the TCP connections. Also want to mention that we are not using TCPOPS module here.
I am trying to achieve consistent duration for TCP reset in kamailio. Could you please point me if I have missed anything here?
Setup:
Client ---- Kamailio (stateless mode) ---- Backend server
TCP Config:
tcp_connection_lifetime = 3600 tcp_accept_aliases = no tcp_async = yes tcp_connect_timeout = 10 tcp_crlf_ping = yes tcp_keepalive = yes tcp_keepcnt = 3 tcp_keepidle = 30 tcp_keepintvl = 30 # Total 2 mins - 30+(30*3)=120 sec tcp_linger2 = 30 tcp_rd_buf_size = 65536 tcp_send_timeout = 10
thanks, raj
_______________________________________________
Kamailio (SER) - Users Mailing List
sr-users@lists.kamailio.orgmailto:sr-users@lists.kamailio.org
https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-usershttps://urldefense.proofpoint.com/v2/url?u=https-3A__lists.kamailio.org_cgi-2Dbin_mailman_listinfo_sr-2Dusers&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=lQLR6IfVOO2CWDL4iEDldpjFjZfxQFyfKeFd1pGZkoY&e=
--
Daniel-Constantin Mierla -- www.asipto.comhttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=2-7juQBEXrjA_PjfpiIZ0JvAZxWKngyC0vcVyP1GmUs&e=
www.twitter.com/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=_qINzlBw2yRKQ3ke2UwXsjTQ7lKWI25XpjOlzw9gPqA&e= -- www.linkedin.com/in/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=TsFAXfvNfu1k1Kw-F4937svidQSEfPMgK4VmorRPn54&e=
--
Daniel-Constantin Mierla -- www.asipto.comhttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=UMaxn1Owy1eHi2eAw29Y1v44igz-iYzUksNhxkR7em8&e=
www.twitter.com/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=nLdsS8oJ8kD-DcfH79nsBCyj95290vEG4AifFckLDcM&e= -- www.linkedin.com/in/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=lb0Xu4UZ9vgK9ovbzSApy8zFMVF67L6Hh2-rHcEGtt0&e=
--
Daniel-Constantin Mierla -- www.asipto.comhttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=UMaxn1Owy1eHi2eAw29Y1v44igz-iYzUksNhxkR7em8&e=
www.twitter.com/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=nLdsS8oJ8kD-DcfH79nsBCyj95290vEG4AifFckLDcM&e= -- www.linkedin.com/in/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=lb0Xu4UZ9vgK9ovbzSApy8zFMVF67L6Hh2-rHcEGtt0&e=
If it is established, then the other side didn't actually requested the close of the connection. It should be in close-wait state when the remote peer sent the close request.
Look at the traffic of the connection with wireshare, it shows also the tcp control packets.
On 12.06.19 09:57, R, Rajkumar (Raj) wrote:
It remained in ‘ESTABLISHED’ state only. Please find netstat output below.
# netstat -natlp
Active Internet connections (servers and established)
Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name
tcp 0 0 0.0.0.0:5060 0.0.0.0:* LISTEN 20/kamailio
tcp 0 0 0.0.0.0:80 0.0.0.0:* LISTEN -
tcp 0 0 0.0.0.0:411 0.0.0.0:* LISTEN -
tcp 0 0 0.0.0.0:8443 0.0.0.0:* LISTEN -
tcp 0 0 0.0.0.0:5056 0.0.0.0:* LISTEN 20/kamailio
tcp 0 3777 10.1.4.55:5060 10.27.117.55:38115 ESTABLISHED 20/Kamailio
thanks,
raj
+918067153382
*From:*Daniel-Constantin Mierla miconda@gmail.com *Sent:* Wednesday, June 12, 2019 12:29 PM *To:* R, Rajkumar (Raj) rajkumaradass@avaya.com; Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org *Subject:* Re: [SR-Users] TCP reset behavior
Hello,
the connection state is OK in all outputs, so not seen as closed by Kamailio.
Can you get the output of the netstat for that connection after is supposed to be closed to see its state there?
Cheers, Daniel
On 11.06.19 12:17, Daniel-Constantin Mierla wrote:
Hello, I didn't get the time to look at. It was a long weekend in Germany, with Monday being public holiday. Not sure if I get to it today, but hope to get time in the next few days. Cheers, Daniel On 11.06.19 07:56, R, Rajkumar (Raj) wrote: Did the shared tcp stats help? Awaiting your inputs on this. thanks, raj *From:*sr-users <sr-users-bounces@lists.kamailio.org> <mailto:sr-users-bounces@lists.kamailio.org> *On Behalf Of *R, Rajkumar (Raj) *Sent:* Friday, June 7, 2019 10:44 PM *To:* miconda@gmail.com <mailto:miconda@gmail.com>; Kamailio (SER) - Users Mailing List <sr-users@lists.kamailio.org> <mailto:sr-users@lists.kamailio.org> *Subject:* Re: [SR-Users] TCP reset behavior Hi Daniel, Please find attached stats output captured every 2 seconds for the entire duration of test (i.e., Start of registration from client till the connection towards client is reset by kamailio after 17mins approximately). I had used ‘kamcmd’ instead of ‘kamctl’ for same command arguments. Setup: 10.27.117.55 (Client) --- Kamailio (10.1.4.55) ---- Backend server (10.2.8.70) Summary of events executed/observed during the test: * Once the client is registered successfully to the backend, turned off WiFi connection at the client side (time 13:07.05 UTC 2019). * After OPTIONS msg time out for response from the client, backend closes the TCP connection towards kamailio ( around time 13:10:39), * TCP connection towards client side is reset at 13:24:30 (i.e., 17 minutes since the WiFi is turned off) * In wireshark, see re-transmission of a TCP packet from kamailio for about 15 minutes until connection release. thanks, raj *From:*Daniel-Constantin Mierla <miconda@gmail.com <mailto:miconda@gmail.com>> *Sent:* Friday, June 7, 2019 6:14 PM *To:* Kamailio (SER) - Users Mailing List <sr-users@lists.kamailio.org <mailto:sr-users@lists.kamailio.org>>; R, Rajkumar (Raj) <rajkumaradass@avaya.com <mailto:rajkumaradass@avaya.com>> *Subject:* Re: [SR-Users] TCP reset behavior Hello, can you reproduce the case and then get the output of the next commands: kamctl rpc core.tcp_list kamctl rpc core.tcp_options kamctl rpc core.tcp_info Cheers, Daniel On 06.06.19 18:53, R, Rajkumar (Raj) wrote: Hi All, In our setup running kamailio (alpine image) in stateless mode and have following TCP config set. After client registers to the backend see two TCP connections created – one towards the client and the other towards the backendserver. If there’s a abrupt network disconnect either at the client side or at the backend side, do not see TCP reset happening from kamailio as per the config. Per below config expected after 120seconds( = [tcp_keepidle] + ( [tcp_keepintvl * tcp_keepcnt] ) instead it keeps resending TCP keepalive packets at varying intervals for varying duration length(like 6mins, 14mins..) before resetting the TCP connections. Also want to mention that we are not using TCPOPS module here. I am trying to achieve consistent duration for TCP reset in kamailio. Could you please point me if I have missed anything here? */Setup:/* Client ---- Kamailio (stateless mode) ---- Backend server *TCP Config:* * * *tcp_connection_lifetime = 3600* tcp_accept_aliases = no tcp_async = yes tcp_connect_timeout = 10 tcp_crlf_ping = yes *tcp_keepalive = yes* *tcp_keepcnt = 3* *tcp_keepidle = 30* *tcp_keepintvl = 30 # Total 2 mins - 30+(30*3)=120 sec* tcp_linger2 = 30 tcp_rd_buf_size = 65536 tcp_send_timeout = 10 thanks, raj _______________________________________________ Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org <mailto:sr-users@lists.kamailio.org> https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users <https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.kamailio.org_cgi-2Dbin_mailman_listinfo_sr-2Dusers&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=lQLR6IfVOO2CWDL4iEDldpjFjZfxQFyfKeFd1pGZkoY&e=> -- Daniel-Constantin Mierla -- www.asipto.com <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=2-7juQBEXrjA_PjfpiIZ0JvAZxWKngyC0vcVyP1GmUs&e=> www.twitter.com/miconda <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=_qINzlBw2yRKQ3ke2UwXsjTQ7lKWI25XpjOlzw9gPqA&e=> -- www.linkedin.com/in/miconda <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=TsFAXfvNfu1k1Kw-F4937svidQSEfPMgK4VmorRPn54&e=> -- Daniel-Constantin Mierla -- www.asipto.com <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=UMaxn1Owy1eHi2eAw29Y1v44igz-iYzUksNhxkR7em8&e=> www.twitter.com/miconda <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=nLdsS8oJ8kD-DcfH79nsBCyj95290vEG4AifFckLDcM&e=> -- www.linkedin.com/in/miconda <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=lb0Xu4UZ9vgK9ovbzSApy8zFMVF67L6Hh2-rHcEGtt0&e=>
-- Daniel-Constantin Mierla -- www.asipto.com https://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=UMaxn1Owy1eHi2eAw29Y1v44igz-iYzUksNhxkR7em8&e= www.twitter.com/miconda https://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=nLdsS8oJ8kD-DcfH79nsBCyj95290vEG4AifFckLDcM&e= -- www.linkedin.com/in/miconda https://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=lb0Xu4UZ9vgK9ovbzSApy8zFMVF67L6Hh2-rHcEGtt0&e=
The scenario here is the other end's RESET/connection closure will never reach Kamailio, as the client has lost it's network connectivity. In such case expected Kamailio to reset the TCP connection after 120seconds based on the TCP keepalive config params mentioned earlier. From wireshark capture observe, Kamailio keeps resending TCP keepalive packets at varying intervals for varying duration length(like 6mins, 14mins..) before resetting the TCP connections.
thanks, raj
From: Daniel-Constantin Mierla miconda@gmail.com Sent: Wednesday, June 12, 2019 1:36 PM To: R, Rajkumar (Raj) rajkumaradass@avaya.com; Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org Subject: Re: [SR-Users] TCP reset behavior
If it is established, then the other side didn't actually requested the close of the connection. It should be in close-wait state when the remote peer sent the close request.
Look at the traffic of the connection with wireshare, it shows also the tcp control packets. On 12.06.19 09:57, R, Rajkumar (Raj) wrote: It remained in 'ESTABLISHED' state only. Please find netstat output below.
# netstat -natlp Active Internet connections (servers and established) Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name tcp 0 0 0.0.0.0:5060 0.0.0.0:* LISTEN 20/kamailio tcp 0 0 0.0.0.0:80 0.0.0.0:* LISTEN - tcp 0 0 0.0.0.0:411 0.0.0.0:* LISTEN - tcp 0 0 0.0.0.0:8443 0.0.0.0:* LISTEN - tcp 0 0 0.0.0.0:5056 0.0.0.0:* LISTEN 20/kamailio tcp 0 3777 10.1.4.55:5060 10.27.117.55:38115 ESTABLISHED 20/Kamailio
thanks, raj +918067153382
From: Daniel-Constantin Mierla miconda@gmail.commailto:miconda@gmail.com Sent: Wednesday, June 12, 2019 12:29 PM To: R, Rajkumar (Raj) rajkumaradass@avaya.commailto:rajkumaradass@avaya.com; Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.orgmailto:sr-users@lists.kamailio.org Subject: Re: [SR-Users] TCP reset behavior
Hello,
the connection state is OK in all outputs, so not seen as closed by Kamailio.
Can you get the output of the netstat for that connection after is supposed to be closed to see its state there?
Cheers, Daniel On 11.06.19 12:17, Daniel-Constantin Mierla wrote:
Hello,
I didn't get the time to look at. It was a long weekend in Germany, with Monday being public holiday. Not sure if I get to it today, but hope to get time in the next few days.
Cheers, Daniel On 11.06.19 07:56, R, Rajkumar (Raj) wrote: Did the shared tcp stats help? Awaiting your inputs on this.
thanks, raj
From: sr-users sr-users-bounces@lists.kamailio.orgmailto:sr-users-bounces@lists.kamailio.org On Behalf Of R, Rajkumar (Raj) Sent: Friday, June 7, 2019 10:44 PM To: miconda@gmail.commailto:miconda@gmail.com; Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.orgmailto:sr-users@lists.kamailio.org Subject: Re: [SR-Users] TCP reset behavior
Hi Daniel,
Please find attached stats output captured every 2 seconds for the entire duration of test (i.e., Start of registration from client till the connection towards client is reset by kamailio after 17mins approximately). I had used 'kamcmd' instead of 'kamctl' for same command arguments.
Setup:
10.27.117.55 (Client) --- Kamailio (10.1.4.55) ---- Backend server (10.2.8.70)
Summary of events executed/observed during the test:
* Once the client is registered successfully to the backend, turned off WiFi connection at the client side (time 13:07.05 UTC 2019). * After OPTIONS msg time out for response from the client, backend closes the TCP connection towards kamailio ( around time 13:10:39), * TCP connection towards client side is reset at 13:24:30 (i.e., 17 minutes since the WiFi is turned off) * In wireshark, see re-transmission of a TCP packet from kamailio for about 15 minutes until connection release.
thanks, raj
From: Daniel-Constantin Mierla <miconda@gmail.commailto:miconda@gmail.com> Sent: Friday, June 7, 2019 6:14 PM To: Kamailio (SER) - Users Mailing List <sr-users@lists.kamailio.orgmailto:sr-users@lists.kamailio.org>; R, Rajkumar (Raj) <rajkumaradass@avaya.commailto:rajkumaradass@avaya.com> Subject: Re: [SR-Users] TCP reset behavior
Hello,
can you reproduce the case and then get the output of the next commands:
kamctl rpc core.tcp_list
kamctl rpc core.tcp_options
kamctl rpc core.tcp_info
Cheers, Daniel On 06.06.19 18:53, R, Rajkumar (Raj) wrote: Hi All,
In our setup running kamailio (alpine image) in stateless mode and have following TCP config set. After client registers to the backend see two TCP connections created - one towards the client and the other towards the backendserver. If there's a abrupt network disconnect either at the client side or at the backend side, do not see TCP reset happening from kamailio as per the config. Per below config expected after 120seconds( = [tcp_keepidle] + ( [tcp_keepintvl * tcp_keepcnt] ) instead it keeps resending TCP keepalive packets at varying intervals for varying duration length(like 6mins, 14mins..) before resetting the TCP connections. Also want to mention that we are not using TCPOPS module here.
I am trying to achieve consistent duration for TCP reset in kamailio. Could you please point me if I have missed anything here?
Setup:
Client ---- Kamailio (stateless mode) ---- Backend server
TCP Config:
tcp_connection_lifetime = 3600 tcp_accept_aliases = no tcp_async = yes tcp_connect_timeout = 10 tcp_crlf_ping = yes tcp_keepalive = yes tcp_keepcnt = 3 tcp_keepidle = 30 tcp_keepintvl = 30 # Total 2 mins - 30+(30*3)=120 sec tcp_linger2 = 30 tcp_rd_buf_size = 65536 tcp_send_timeout = 10
thanks, raj
_______________________________________________
Kamailio (SER) - Users Mailing List
sr-users@lists.kamailio.orgmailto:sr-users@lists.kamailio.org
https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-usershttps://urldefense.proofpoint.com/v2/url?u=https-3A__lists.kamailio.org_cgi-2Dbin_mailman_listinfo_sr-2Dusers&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=lQLR6IfVOO2CWDL4iEDldpjFjZfxQFyfKeFd1pGZkoY&e=
--
Daniel-Constantin Mierla -- www.asipto.comhttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=2-7juQBEXrjA_PjfpiIZ0JvAZxWKngyC0vcVyP1GmUs&e=
www.twitter.com/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=_qINzlBw2yRKQ3ke2UwXsjTQ7lKWI25XpjOlzw9gPqA&e= -- www.linkedin.com/in/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=TsFAXfvNfu1k1Kw-F4937svidQSEfPMgK4VmorRPn54&e=
--
Daniel-Constantin Mierla -- www.asipto.comhttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=UMaxn1Owy1eHi2eAw29Y1v44igz-iYzUksNhxkR7em8&e=
www.twitter.com/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=nLdsS8oJ8kD-DcfH79nsBCyj95290vEG4AifFckLDcM&e= -- www.linkedin.com/in/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=lb0Xu4UZ9vgK9ovbzSApy8zFMVF67L6Hh2-rHcEGtt0&e=
--
Daniel-Constantin Mierla -- www.asipto.comhttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=UMaxn1Owy1eHi2eAw29Y1v44igz-iYzUksNhxkR7em8&e=
www.twitter.com/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=nLdsS8oJ8kD-DcfH79nsBCyj95290vEG4AifFckLDcM&e= -- www.linkedin.com/in/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=lb0Xu4UZ9vgK9ovbzSApy8zFMVF67L6Hh2-rHcEGtt0&e=
--
Daniel-Constantin Mierla -- www.asipto.comhttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=4Ir7EzpigornPG5nIqPd6gt8_t8XTDJYOUR2UmwPjdk&s=i6DE7_KD6kIHLqHWtqMJQ6uNlWBkQyizhdR1Q8_vrTk&e=
www.twitter.com/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=4Ir7EzpigornPG5nIqPd6gt8_t8XTDJYOUR2UmwPjdk&s=TL4A8Sp38pb4_DQjx_0ojc8-TMKtaMO9ujATPs2Hfik&e= -- www.linkedin.com/in/micondahttps://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=4Ir7EzpigornPG5nIqPd6gt8_t8XTDJYOUR2UmwPjdk&s=0XaBVrYfnAMsdKSTrTizX2MzPJJiIpLCGQRENRMqm2M&e=
Kamailio just sets the keepalive options to the socket, sending and handling of the response (or lack of) is done at OS/kernel layer. Kamailio is notified if the connection changes the state.
Check the sysctl values for tcp keepalive and try to change them there to see if if makes any difference.
Cheers, Daniel
On 12.06.19 13:20, R, Rajkumar (Raj) wrote:
The scenario here is the other end’s RESET/connection closure will never reach Kamailio, as the client has lost it’s network connectivity.
In such case expected Kamailio to reset the TCP connection after 120seconds based on the TCP keepalive config params mentioned earlier. From wireshark capture observe, Kamailio keeps resending TCP keepalive packets at varying intervals for varying duration length(like 6mins, 14mins..) before resetting the TCP connections.
thanks,
raj
*From:*Daniel-Constantin Mierla miconda@gmail.com *Sent:* Wednesday, June 12, 2019 1:36 PM *To:* R, Rajkumar (Raj) rajkumaradass@avaya.com; Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org *Subject:* Re: [SR-Users] TCP reset behavior
If it is established, then the other side didn't actually requested the close of the connection. It should be in close-wait state when the remote peer sent the close request.
Look at the traffic of the connection with wireshare, it shows also the tcp control packets.
On 12.06.19 09:57, R, Rajkumar (Raj) wrote:
It remained in ‘ESTABLISHED’ state only. Please find netstat output below. # netstat -natlp Active Internet connections (servers and established) Proto Recv-Q Send-Q Local Address Foreign Address State PID/Program name tcp 0 0 0.0.0.0:5060 0.0.0.0:* LISTEN 20/kamailio tcp 0 0 0.0.0.0:80 0.0.0.0:* LISTEN - tcp 0 0 0.0.0.0:411 0.0.0.0:* LISTEN - tcp 0 0 0.0.0.0:8443 0.0.0.0:* LISTEN - tcp 0 0 0.0.0.0:5056 0.0.0.0:* LISTEN 20/kamailio tcp 0 3777 10.1.4.55:5060 10.27.117.55:38115 ESTABLISHED 20/Kamailio thanks, raj +918067153382 *From:*Daniel-Constantin Mierla <miconda@gmail.com> <mailto:miconda@gmail.com> *Sent:* Wednesday, June 12, 2019 12:29 PM *To:* R, Rajkumar (Raj) <rajkumaradass@avaya.com> <mailto:rajkumaradass@avaya.com>; Kamailio (SER) - Users Mailing List <sr-users@lists.kamailio.org> <mailto:sr-users@lists.kamailio.org> *Subject:* Re: [SR-Users] TCP reset behavior Hello, the connection state is OK in all outputs, so not seen as closed by Kamailio. Can you get the output of the netstat for that connection after is supposed to be closed to see its state there? Cheers, Daniel On 11.06.19 12:17, Daniel-Constantin Mierla wrote: Hello, I didn't get the time to look at. It was a long weekend in Germany, with Monday being public holiday. Not sure if I get to it today, but hope to get time in the next few days. Cheers, Daniel On 11.06.19 07:56, R, Rajkumar (Raj) wrote: Did the shared tcp stats help? Awaiting your inputs on this. thanks, raj *From:*sr-users <sr-users-bounces@lists.kamailio.org> <mailto:sr-users-bounces@lists.kamailio.org> *On Behalf Of *R, Rajkumar (Raj) *Sent:* Friday, June 7, 2019 10:44 PM *To:* miconda@gmail.com <mailto:miconda@gmail.com>; Kamailio (SER) - Users Mailing List <sr-users@lists.kamailio.org> <mailto:sr-users@lists.kamailio.org> *Subject:* Re: [SR-Users] TCP reset behavior Hi Daniel, Please find attached stats output captured every 2 seconds for the entire duration of test (i.e., Start of registration from client till the connection towards client is reset by kamailio after 17mins approximately). I had used ‘kamcmd’ instead of ‘kamctl’ for same command arguments. Setup: 10.27.117.55 (Client) --- Kamailio (10.1.4.55) ---- Backend server (10.2.8.70) Summary of events executed/observed during the test: * Once the client is registered successfully to the backend, turned off WiFi connection at the client side (time 13:07.05 UTC 2019). * After OPTIONS msg time out for response from the client, backend closes the TCP connection towards kamailio ( around time 13:10:39), * TCP connection towards client side is reset at 13:24:30 (i.e., 17 minutes since the WiFi is turned off) * In wireshark, see re-transmission of a TCP packet from kamailio for about 15 minutes until connection release. thanks, raj *From:*Daniel-Constantin Mierla <miconda@gmail.com <mailto:miconda@gmail.com>> *Sent:* Friday, June 7, 2019 6:14 PM *To:* Kamailio (SER) - Users Mailing List <sr-users@lists.kamailio.org <mailto:sr-users@lists.kamailio.org>>; R, Rajkumar (Raj) <rajkumaradass@avaya.com <mailto:rajkumaradass@avaya.com>> *Subject:* Re: [SR-Users] TCP reset behavior Hello, can you reproduce the case and then get the output of the next commands: kamctl rpc core.tcp_list kamctl rpc core.tcp_options kamctl rpc core.tcp_info Cheers, Daniel On 06.06.19 18:53, R, Rajkumar (Raj) wrote: Hi All, In our setup running kamailio (alpine image) in stateless mode and have following TCP config set. After client registers to the backend see two TCP connections created – one towards the client and the other towards the backendserver. If there’s a abrupt network disconnect either at the client side or at the backend side, do not see TCP reset happening from kamailio as per the config. Per below config expected after 120seconds( = [tcp_keepidle] + ( [tcp_keepintvl * tcp_keepcnt] ) instead it keeps resending TCP keepalive packets at varying intervals for varying duration length(like 6mins, 14mins..) before resetting the TCP connections. Also want to mention that we are not using TCPOPS module here. I am trying to achieve consistent duration for TCP reset in kamailio. Could you please point me if I have missed anything here? */Setup:/* Client ---- Kamailio (stateless mode) ---- Backend server *TCP Config:* * * *tcp_connection_lifetime = 3600* tcp_accept_aliases = no tcp_async = yes tcp_connect_timeout = 10 tcp_crlf_ping = yes *tcp_keepalive = yes* *tcp_keepcnt = 3* *tcp_keepidle = 30* *tcp_keepintvl = 30 # Total 2 mins - 30+(30*3)=120 sec* tcp_linger2 = 30 tcp_rd_buf_size = 65536 tcp_send_timeout = 10 thanks, raj _______________________________________________ Kamailio (SER) - Users Mailing List sr-users@lists.kamailio.org <mailto:sr-users@lists.kamailio.org> https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-users <https://urldefense.proofpoint.com/v2/url?u=https-3A__lists.kamailio.org_cgi-2Dbin_mailman_listinfo_sr-2Dusers&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=lQLR6IfVOO2CWDL4iEDldpjFjZfxQFyfKeFd1pGZkoY&e=> -- Daniel-Constantin Mierla -- www.asipto.com <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=2-7juQBEXrjA_PjfpiIZ0JvAZxWKngyC0vcVyP1GmUs&e=> www.twitter.com/miconda <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=_qINzlBw2yRKQ3ke2UwXsjTQ7lKWI25XpjOlzw9gPqA&e=> -- www.linkedin.com/in/miconda <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=m1fGDJFl23Vq-awSnjgML3TymzSQRBOSTCALAsyAHNM&s=TsFAXfvNfu1k1Kw-F4937svidQSEfPMgK4VmorRPn54&e=> -- Daniel-Constantin Mierla -- www.asipto.com <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=UMaxn1Owy1eHi2eAw29Y1v44igz-iYzUksNhxkR7em8&e=> www.twitter.com/miconda <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=nLdsS8oJ8kD-DcfH79nsBCyj95290vEG4AifFckLDcM&e=> -- www.linkedin.com/in/miconda <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=lb0Xu4UZ9vgK9ovbzSApy8zFMVF67L6Hh2-rHcEGtt0&e=> -- Daniel-Constantin Mierla -- www.asipto.com <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=UMaxn1Owy1eHi2eAw29Y1v44igz-iYzUksNhxkR7em8&e=> www.twitter.com/miconda <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=nLdsS8oJ8kD-DcfH79nsBCyj95290vEG4AifFckLDcM&e=> -- www.linkedin.com/in/miconda <https://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=OCP_rR5_9SQpkMRB1w_kOQYN3ooVOlU65du-BxCLfdM&s=lb0Xu4UZ9vgK9ovbzSApy8zFMVF67L6Hh2-rHcEGtt0&e=>
-- Daniel-Constantin Mierla -- www.asipto.com https://urldefense.proofpoint.com/v2/url?u=http-3A__www.asipto.com&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=4Ir7EzpigornPG5nIqPd6gt8_t8XTDJYOUR2UmwPjdk&s=i6DE7_KD6kIHLqHWtqMJQ6uNlWBkQyizhdR1Q8_vrTk&e= www.twitter.com/miconda https://urldefense.proofpoint.com/v2/url?u=http-3A__www.twitter.com_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=4Ir7EzpigornPG5nIqPd6gt8_t8XTDJYOUR2UmwPjdk&s=TL4A8Sp38pb4_DQjx_0ojc8-TMKtaMO9ujATPs2Hfik&e= -- www.linkedin.com/in/miconda https://urldefense.proofpoint.com/v2/url?u=http-3A__www.linkedin.com_in_miconda&d=DwMD-g&c=BFpWQw8bsuKpl1SgiZH64Q&r=LDE-f1bLxMPmcrsp8ONITcznNqEIlhe9ffKEZPeB7EI&m=4Ir7EzpigornPG5nIqPd6gt8_t8XTDJYOUR2UmwPjdk&s=0XaBVrYfnAMsdKSTrTizX2MzPJJiIpLCGQRENRMqm2M&e=