About this article: This article is reviewed against the cited sources. This is source research, not lab validation. Check the stated versions and your own requirements before making changes.
If a site-to-site IPsec tunnel never forms and the log shows NO_PROPOSAL_CHOSEN, the first thing to establish is which exchange rejected the proposal. When it is the first one, IKE_SA_INIT, the two gateways could not agree on a cipher, integrity or pseudo-random function (PRF), or key exchange group for the IKE security association itself. No IKE security association exists, and nothing else about the tunnel can work until they agree.
Compare what the peer offered with what the responder is configured to accept, then make one common choice on both sides.
What the error means
IKEv2 builds a tunnel in two steps. The first exchange, IKE_SA_INIT, negotiates the cryptographic suite for the IKE security association (the IKE SA) and swaps keying material and nonces. The second, IKE_AUTH, authenticates both ends and creates the first child SA that carries your traffic. RFC 7296 states that "the responder MUST accept a single proposal or reject them all and return an error. The error is given in a notification of type NO_PROPOSAL_CHOSEN." It defines the notification as "None of the proposed crypto suites was acceptable."
The same notification can appear later. RFC 7296 lists NO_PROPOSAL_CHOSEN among the errors in IKE_AUTH that "do not prevent an IKE SA from being set up", because the IKE SA already exists and only the child SA failed. That is a different fault. A tunnel whose IKE SA is up but whose child SA is missing is covered separately under a site-to-site IPsec tunnel that is up but passes no traffic.
Here is how to tell them apart:
- Rejected in IKE_SA_INIT: no IKE SA exists. RFC 7296 notes that in this case the responder's SPI in the reply is zero.
- Rejected in IKE_AUTH: the IKE SA is up and the child SA is missing.
The NetApp example later in this section labels the notification received unauthenticated v2N_NO_PROPOSAL_CHOSEN. Treat it as a pointer to where to look, and confirm the cause by reading the responder's own log.
Likely causes
| Cause | What you see | How to confirm |
|---|---|---|
| Cipher or integrity mismatch (for example AES-GCM offered, AES-CBC with HMAC configured) | Responder logs received proposals and configured proposals that differ |
Compare the two lines |
| Different key exchange group (DH group) | Cipher and PRF match, group differs | Compare the group in each proposal |
| IKEv1 on one peer and IKEv2 on the other | FortiGate reports a Phase 1 mismatch even when algorithms match | Check the IKE version on both gateways |
| FortiGate requires a zero SPI size in the IKE SA proposal (FortiOS 7.4.5 and later, fixed in 7.4.8) | no proposal found and no SA proposal chosen against a non-Fortinet peer |
Check the FortiOS build; try FortiGate as initiator |
| Wrong peer or wrong ID (other Phase 1 settings) | Fortinet lists remote gateway IP, network-id, local ID |
Verify peer address and ID settings on both sides |
| Responder preference picks a suite the initiator cannot use | Initiator lists several, responder accepts none | Compare ordered lists on both sides |
Prove the cause before changing anything
Read-only steps first. These commands have not been run or lab-tested.
On a strongSwan responder
List the IKE SAs. The command, documented as "lists established IKE SAs and their dependent CHILD SAs", is:
swanctl --list-sas
If the first exchange is failing, the list has no entry for this peer. If an IKE SA is listed but the child SA is missing, you have the IKE_AUTH case, not this one.
Next, raise the log level for the IKE and configuration subsystems. The strongSwan logging page documents the groups cfg ("Configuration management and plugins") and ike ("IKE_SA/ISAKMP SA"), and levels where 1 is "Generic control flow with errors, a good default to see whats going on" and 2 is "More detailed debugging control flow". A file logger in strongswan.conf looks like this:
charon {
filelog {
charon {
path = /var/log/charon.log
time_format = %b %e %T
ike_name = yes
default = 2
flush_line = yes
}
}
}
Level 2 is more detailed and writes more. The same page documents swanctl --log, which subscribes to the log events (levels 0 and 1) without changing the file logger.
Now provoke one attempt. The swanctl --initiate page documents swanctl --initiate [--child <name>] [--ike <name>] [--timeout <s>], where --ike means "initiate an IKE_SA, or name of child's parent". Either start the connection from the other gateway or run swanctl --initiate --ike <connection-name> locally, then read the log.
The NetApp ONTAP knowledge base entry on IPsec connection attempts that fail with No Proposals Chosen shows the lines to look for from a strongSwan (Charon) responder:
11[CFG] received proposals: IKE:AES_GCM_16_256/PRF_HMAC_SHA2_384/ECP_384
11[CFG] configured proposals: IKE:AES_CBC_256/HMAC_SHA2_384_192/PRF_HMAC_SHA2_384/ECP_384
The first line is what the peer offered. The second is what this side will accept. Read them in order, element by element: cipher, integrity, PRF, group. In that example the peer offers AES-GCM with no separate integrity algorithm and the responder is configured for AES-CBC with HMAC-SHA2-384-192, so they share a PRF and a group but no cipher. The visible part of the NetApp page does not state a cause. The difference is in the two lines. The initiator in that entry (Libreswan) logged STATE_PARENT_I1: received unauthenticated v2N_NO_PROPOSAL_CHOSEN, and a packet capture showed the reply "Notify Message Type: NO_PROPOSAL_CHOSEN (14)".
A good result is a received proposals line that contains at least one suite fully present in the configured proposals line. A bad result is any element with no overlap.
On a FortiGate
Fortinet's community article, Understanding message 'no proposal chosen' in IKE debug (last updated 21 September 2026), gives the debug sequence:
diagnose vpn ike log filter rem-addr4 <remote IP address>
diagnose debug application ike -1
diagnose debug console timestamp enable
diagnose debug enable
Stop with diag debug disable and diagnose debug reset. Risk: debug at -1 on a busy gateway is noisy and uses CPU, so filter by the remote address first and disable debug when you have the lines.
Fortinet's logs include ike 0:VPN:968190: processing notify type NO_PROPOSAL_CHOSEN and no SA proposal chosen. Fortinet says "no SA proposal chosen" is different from "no proposal chosen". The first is usually a Phase 1 mismatch of IKE version, IKE mode, or encryption and authentication algorithms. Fortinet also lists IKEv1 on one peer and IKEv2 on the other as a cause that applies "even if the encrypt/auth algorithm matches", and a network-id set on only one peer.
On a Palo Alto firewall
The PAN-OS 10.2 page, Troubleshoot Your IPSec VPN Tunnel Connection (updated 29 June 2026), lists these entries:
Received unencrypted notify payload (no proposal chosen) from IP x.x.x.x[500] to y.y.y.y[500], ignored...IKE phase-1 negotiation is failed. Unable to process peer's SA payload.The page's fix is to check the IKE Crypto profile and verify that the proposals on both sides have a common encryption, authentication and DH group.IKE phase-1 negotiation is failed as initiator, main mode. Failed SA: x.x.x.x[500]-y.y.y.y[500]. For this one the page's fix is to verify that the peer's public IP address is accurate in the IKE gateway.
To verify Phase 1, the page gives show vpn ike-sa gateway <gateway_name>. If the SA is absent from the output, read the system log for the cause.
Check the proposal order
strongSwan documents proposals as a "list of algorithm keywords separated by dashes". With IKEv2, "multiple algorithms of the same kind can be specified in a single proposal, their order defines the preference in descending order". Multiple proposals are separated by commas. An example is aes256gcm16-aes192gcm16-prfsha384-prfsha256-ecp384-modp3072. If no proposals are configured, strongSwan uses its defaults. For IKE SA proposals, the key exchange method is mandatory. It also says AEAD algorithms "can't be combined with classic encryption ciphers in the same proposal", so a GCM proposal needs an explicit PRF.
By default strongSwan prefers its own configured order. The page says this behaviour "can be changed by disabling charon.prefer_configured_proposals in strongswan.conf", after which the first received proposal that the configured list accepts is chosen.
Example (illustrative): gateway fw-a (strongSwan) is the responder for peer fw-b. Between fw-b at 203.0.113.10 and fw-a, the tunnel carries 10.0.10.0/24 to 10.0.20.0/24. On fw-a, swanctl --list-sas shows nothing, and the log reads received proposals: IKE:AES_GCM_16_256/PRF_HMAC_SHA2_384/ECP_384 followed by configured proposals: IKE:AES_CBC_256/HMAC_SHA2_384_192/PRF_HMAC_SHA2_384/ECP_384. Cipher is the only element that differs. Adding aes256gcm16-prfsha384-ecp384 to the proposal list on fw-a gives the two sides a suite in common.
Fix
Choose one suite and set it on both ends. Prefer changing the side whose list is wider, since it usually takes less change. Ask the other side's owner for their exact configured algorithms before changing yours. Do not widen the list to weak algorithms just to get a tunnel up.
- strongSwan: edit the
proposalsline in the connection inswanctl.confto include a suite the peer offers, written in the dash-separated form above. Put the preferred suite first. Reload with your normal swanctl procedure. Disruptive step: reloading connection configuration can drop IKE SAs for connections whose settings changed, and existing tunnels to other peers should be left alone. Take a copy ofswanctl.conffirst and restore it to back out. - FortiGate: edit the Phase 1 proposal on the tunnel so that at least one encryption and authentication combination matches the peer, and make sure both ends use the same IKE version. Disruptive step: changing Phase 1 settings on a tunnel in use restarts that tunnel. Record the existing settings so you can restore them.
- FortiOS 7.4.5 and later against a non-Fortinet peer: the community article says IKE negotiation can fail with
no proposal foundandno SA proposal chosenbecause FortiGate requires the IKE SA proposal SPI size to be zero. It says this is fixed in 7.4.8 and that setting FortiGate as the IKE initiator is a workaround. - Palo Alto: open the IKE Crypto profile used by the IKE gateway and make sure its encryption, authentication and DH group settings include one the peer also uses. Risk: a changed profile applies to every gateway that uses it, so check which gateways share it before you commit.
- If you cannot change the other side,
charon.prefer_configured_proposalscan be turned off on a strongSwan responder so the initiator's order wins. Risk: this changes the selection rule for every connection on that daemon, so use it only if you understand all its peers.
Confirm the fix
- Start the tunnel again from either side.
swanctl --list-sason strongSwan should now list the IKE SA with its child SAs. A listed IKE SA with no child SA is a new fault, in IKE_AUTH.- The log should no longer print
received proposalsandconfigured proposalsfor this peer. - On a FortiGate, the debug output should no longer show
no SA proposal chosen. On a Palo Alto,show vpn ike-sa gateway <gateway_name>should list the SA. - Pass traffic across the tunnel from a host in each subnet.
What to watch afterwards
- Change windows. A cipher change on one side without the other is the usual cause of a tunnel that worked last week.
- Firmware upgrades. The FortiOS 7.4.5 to 7.4.8 case is a reminder that a software update can change what a gateway accepts.
- Peers that change on their side. If the other gateway belongs to another organisation, ask to be told before they change algorithms.
Read next
Related articles you may want next.
- TroubleshootingIPsec VPN Tunnel Up but No Traffic: Read-Only DiagnosisAn IKEv2 tunnel can be up while no data flows. Check child SAs, selectors, NAT, routing and packet size on both strongSwan gateways before any change.Read article
- CloudAzure VPN Gateway: the site-to-site tunnel is up but traffic does not passA Connected status only proves the key exchange worked. Check routes, traffic selectors, gateway subnet settings and packet size, in that order.Read article
- CloudAWS Site-to-Site VPN: BGP against static, tunnel failover and asymmetric returnHow AWS picks a tunnel, what BGP gives that static routing does not, what happens during failover, and how to handle asymmetric return traffic.Read article
Risks and limits
- NetApp publishes the log lines in a public part of its page and keeps the cause and fix behind a sign-in. The cipher mismatch in the example is worked out from the two log lines themselves.
- The strongSwan documentation does not print the
received proposalsandconfigured proposalslog lines. They come from the NetApp entry, which describes ONTAP's strongSwan-based responder. - The Fortinet source is a community article, not the Fortinet documentation library.
- IKEv2 is described. IKEv1 uses different exchanges and different wording.
- Debug logging on a production gateway has a cost. Turn it off when you have the lines.
Sources
- RFC 7296: Internet Key Exchange Protocol Version 2 (IKEv2) — IETF, October 2014, accessed 17 November 2026
- IPsec connection attempt fails with No Proposals Chosen — NetApp knowledge base, last updated 22 July 2025, accessed 17 November 2026
- Algorithm Proposals (Cipher Suites) — strongSwan documentation, date not shown, accessed 17 November 2026
- Logging — strongSwan documentation, date not shown, accessed 17 November 2026
- swanctl --list-sas — strongSwan documentation, date not shown, accessed 17 November 2026
- swanctl --initiate — strongSwan documentation, date not shown, accessed 17 November 2026
- Troubleshooting Tip: Understanding message 'no proposal chosen' in IKE debug — Fortinet community, last updated 21 September 2026, accessed 17 November 2026
- Troubleshoot Your IPSec VPN Tunnel Connection (PAN-OS 10.2) — Palo Alto Networks, updated 29 June 2026, accessed 17 November 2026
This article was originally published on .