From svogl at vodata.de Sat Jun 20 08:05:14 2020 From: svogl at vodata.de (Siegfried Vogl) Date: Sat, 20 Jun 2020 14:05:14 +0200 Subject: [Openswan Users] Openswan road warrior problem Message-ID: <79e016d9-496f-b357-01e7-12655807a571@vodata.de> Hello, i have a road warrior problem. Data flow: IPSEC-Client (Initiator) <-> IP-Gateway <-> Ipsec-Gateway(Responder) <-> Web-Server If the client is configured as a "road warrior" on the Ipsec gateway, the rekeying of the ISAKMP SA does not work on the responder side. The gateway receives the first negotiation packet itself (see log: "RZNV87 / Log.txt" (search:)-> "##$$ Pluto sends packet to himself ...") Attachments - a TGZ extracting as folder "Data":  - Networkplan.pdf: a sheme of the test bed  - RZNV87: Ipsec-Gateway (responder) data (search for comments in Log.txt: "##$$")  - R2_Testclient: Ipsec-Client (Initiator) data. Notes:  - "IpsecStatusCmd.txt" in both directories is the output of "ipsec status" after the tunnel was established and after the fist try of the responder to rekey the ISAKMP SA.  - After the tunnel is established, data transfer is successfull.  - All data files were filtered against readable DN's an FQDN's (although it's only a test bed anyway).  - RZNV89, RZNV87 and RZNV82 are VM's on a virtualbox host (RZN161).  - The test bed is simplified a little.    In production, the initiator (which has a static ip here to simplify things) is a DHCP client. Therefore i need the roadwarrior setup.    I shortened the timers for ikelifetime, salifetime and rekeymargin to get shorter test intervalls. But it is the same problem with default values.  - I did some debugging:    It seems that "st->st_remoteaddr" of the new created state element is not copied, although the "orient()" seems to find the correct assignment.    But because I don't know details of the implementation, this is only an assumption. Thanks in advance, Siegfried -------------- next part -------------- A non-text attachment was scrubbed... Name: Data.tar.gz Type: application/gzip Size: 210980 bytes Desc: not available URL: From svogl at vodata.de Sat Jun 20 08:25:37 2020 From: svogl at vodata.de (Siegfried Vogl) Date: Sat, 20 Jun 2020 14:25:37 +0200 Subject: [Openswan Users] Openswan road warrior problem Message-ID: <5ecdc684-18d1-c2d3-8da1-4bda8b0e4ce6@vodata.de> Hello, sorry, I've made a mistake. The "R2_Testclient/IpsecBarf.txt" is from a wrong device. I've packed the whole data once again, now with the corrected "IpsecBarfCorrected.txt" Sorry, Siegfried -----Ursprüngliche Nachricht----- *Von:* Siegfried Vogl  *Gesendet:* Samstag 20 Juni 2020 14:05 *An:* users at lists.openswan.org *Betreff:* Openswan road warrior problem Hello, i have a road warrior problem. Data flow: IPSEC-Client (Initiator) <-> IP-Gateway <-> Ipsec-Gateway(Responder) <-> Web-Server If the client is configured as a "road warrior" on the Ipsec gateway, the rekeying of the ISAKMP SA does not work on the responder side. The gateway receives the first negotiation packet itself (see log: "RZNV87 / Log.txt" (search:)-> "##$$ Pluto sends packet to himself ...") Attachments - a TGZ extracting as folder "Data":  - Networkplan.pdf: a sheme of the test bed  - RZNV87: Ipsec-Gateway (responder) data (search for comments in Log.txt: "##$$")  - R2_Testclient: Ipsec-Client (Initiator) data. Notes:  - "IpsecStatusCmd.txt" in both directories is the output of "ipsec status" after the tunnel was established and after the fist try of the responder to rekey the ISAKMP SA.  - After the tunnel is established, data transfer is successfull.  - All data files were filtered against readable DN's an FQDN's (although it's only a test bed anyway).  - RZNV89, RZNV87 and RZNV82 are VM's on a virtualbox host (RZN161).  - The test bed is simplified a little.    In production, the initiator (which has a static ip here to simplify things) is a DHCP client. Therefore i need the roadwarrior setup.    I shortened the timers for ikelifetime, salifetime and rekeymargin to get shorter test intervalls. But it is the same problem with default values.  - I did some debugging:    It seems that "st->st_remoteaddr" of the new created state element is not copied, although the "orient()" seems to find the correct assignment.    But because I don't know details of the implementation, this is only an assumption. Thanks in advance, Siegfried -------------- next part -------------- A non-text attachment was scrubbed... Name: Data_corrected.tar.gz Type: application/gzip Size: 203208 bytes Desc: not available URL: