Phase1 negotiation failed due to send error 500

Mikrotik phase 1 negotiation failed due to send error

Wed Mar 03, 2021 3:50 pm

I created dual WAN (with PCC) configuration on my Mikrotik according to instructions:
https://wiki.mikrotik.com/wiki/Manual:PCC

Now I’m trying to run VPN (https://grzegorzkowalik.com/mikrotik-od . -ipsec-07/) with IPsec but I have error

Re: VPN IPsec — phase1 negotiation failed due to send error

Thu Mar 04, 2021 2:18 pm

So the Mikrotik with the PCC configuration acts as an L2TP server, is that correct?

Can you post the complete export of your configuration, after a «non-destructive anonymisation» as suggested in my automatic signature below?

Re: VPN IPsec — phase1 negotiation failed due to send error

Thu Mar 04, 2021 3:13 pm

Re: VPN IPsec — phase1 negotiation failed due to send error

Thu Mar 04, 2021 4:35 pm

The whole problem is that the use of policy routing for packets generated by the router itself is a bit counter-intuitive.

When some process running on the router itself sends a packet, the first step is to find a route for that packet using routing table main, which consists of routes with no routing-mark assigned. Then, a source address is assigned to that packet, according to the properties of that route (the local address indicated in the pref-src parameter or, if no pref-src is specified, the local address associated to the output interface), unless the packet is a response. And only after successful completion of these two steps, the packet is processed by mangle rules in chain output. If the mangle rules eventually assign a routing-mark, the packet is routed once again, taking the routing-mark value into account (but not changing the source address of the packet). This step is called «routing adjustment» at the packet flow diagram.

Since you weren’t aware of this, you haven’t configured any default route in the routing table main, hence the first round of routing of the response IKE packets sent by the router itself to the address of the L2TP client device fails as no route is available for their destination address.

Routes to connected subnets (XX.XX.19.152/29, XX.XX.16.24/29, 192.168.0.0/24) are added to routing table main dynamically, which explains why the L2TP/IPsec sessions establishes fine when the client is in your LAN subnet.

So for the incoming connections through WAN, the remedy is simple, add a default route pointing to any gateway. You can create a bridge interface with no ports, which will thus stay up no matter what, and use the interface name as the gateway parameter of that route if you don’t want to make one of the two WAN gateways a default one in routing table main.

Источник

Mikrotik phase 1 negotiation failed due to send error

Tue Apr 02, 2019 12:13 pm

The RB4011 is behind a router with Port-Forwarding. This seems to work, but i think something with the packet handling is failing. On Client side ive tried with different devices, everytime the same error. On Mikrotik i am using a NAT for all outgoing traffic to the internet router. So the actual Internet IP of the packet is different than the internal. This could cause the client to invalidate the connection due different IP encryption or something like this.

I hope anybody has an idea how to solve this?

Re: L2TP/IPSec — phase 1 negotiation failed due to send error

Tue Apr 02, 2019 1:57 pm

Re: L2TP/IPSec — phase 1 negotiation failed due to send error

Tue Apr 02, 2019 2:26 pm

Re: L2TP/IPSec — phase 1 negotiation failed due to send error

Tue Apr 02, 2019 4:55 pm

Re: L2TP/IPSec — phase 1 negotiation failed due to send error

Wed Apr 03, 2019 10:48 am

Re: L2TP/IPSec — phase 1 negotiation failed due to send error

Sat Apr 06, 2019 8:20 pm

Well, «only the default routing table» and «they are configured to use for the same connection always the same WAN interface» don’t fit well to the same paragraph The default routing table is the one called «main», and a routing-mark is essentialy a synonym to a routing table name, except that no routing-mark assigned means «use the routing table called ‘main’, i.e. the default one».

But that’s not the merit. The merit is that your mangle rules in chains input and output do ensure that what came in via one of the WANs will be responded via the same WAN, and that the other rules in chain prerouting do not break this. So up to the routing-mark assignment, we are fine here. What you haven’t shown, though, are the actual marked routes, can you please show your /ip route export ? Since your load balancing works, I assume this part is fine too, but better double-check.

I’m not sure I understand the part you wrote about NAT. If you had in mind that you do some NAT on Mikrotik itself, the rules you’ve shown say this is not the case for output packets of ROS itself. So I assume you had in mind that the Fritz!box NATs the Mikrotik’s source address 192.168.55.1 to its own WAN address (and vice versa in internet->Mikrotik direction).

And now yes, at least the Windows embedded VPN client does not handle well a NAT on the IPsec responder («server») side. I’m not sure how about other (iOS, Android) embedded clients; the NAT-T mechanism itself does handle both initiator-side NAT and responder-side NAT, so a Mikrotik as an IPsec initiator («client») can handle that fine. In some versions of Windows you can tweak the registry settings to make the embedded client accept responder-side NAT, but I think it is better not to depend on this.

So what you need to do is to use an «un-NAT» on Mikrotik side: you create an /interface bridge on the Mikrotik, not assign any member interfaces to it, and assign the public IP address of the Fritz to it (with a /32 mask). Then you tell the /ip ipsec peer used by the L2TP server to use that address by setting its local-address parameter. Next, you add a chain=dstnat action=dst-nat protocol=udp dst-port=500,4500 to-addresses=the.public.ip.of.Fritz to the nat table, so that incoming connections to the IPsec UDP ports would be redirected to that public address. So the initiator will see the same address (the public one of the Fritz!box) as both the source of the responder’s traffic and the responder’s actual address in the NAT-T payload.

This way, except a minor brain damage which heals quickly, the only trouble will be with the clients running on public IP addresses themselves if the Fritz!box is unable to forward and NAT the ESP. The NAT-T mechanism switches over to encapsulation of ESP into UDP only if it detects NAT on at least one end of the connection, so if it doesn’t detect it on the responder side (which was our goal) nor on the client side, it won’t activate the ESP over UDP mode. So if you expect to have clients on public IPs, you have to make sure that the Fritz!box delivers ESP packets to the 4011 and to add another chain=dstnat rule which is the same like the one above except that you replace protocol=udp dst-port=500,4500 by protocol=esp .

Another stage of the quest is that the L2TP server normally auto-creates its own /ip ipsec peer and you cannot instruct it what parameters to use for that, except the secret . So to be able to set the local-address the way you need, you have to clone the auto-created peer (using /ip ipsec peer add copy-from=[find where dynamic and secret=your-secret] or something similar), and then do /interface l2tp-server server set use-ipsec=no to remove the dynamically created peer.

Источник

Mikrotik phase 1 negotiation failed due to send error

Fri Jan 12, 2018 11:06 am

Re: permanent «phase 1 negotiation failed»

Fri Jan 12, 2018 1:46 pm

Hi Sindy,
I have no access to a console but I managed to use Winbox to look at the IP IPsec configuration. Will it be the same?
This is the information that I collected. I have not replaced any address because it only has generic addresses:

• Policies
— *T
Src.Address ::/0
Src.Port Dst. Address ::/0
Protocol 0 (all)
Action encrypt
Level require
Tunnel no
• Groups
— *
Default
• Peers
—
Address 0.0.0.0
Port 500
Hash algoritm sha512
Encryption algorithm 3des aes-256
• Remote peers
—
Local address 127.0.0.1
Remote address 0.0.0.0
• Mode configs
— *
Name request-only
Address pool Address prefix Split-include Send DNS yes
• Proposals
— *
Name default
Auth algorithms sha1
Encr algorithms aes-128 cbc aes-192 cbc aes-256 cbc
Lifetime 00:30:00
PFS Group modp 1024
—
Name proposal1
Auth algorithms sha1 sha512
Encr algorithms 3des cbc aes-256 ctr
Lifetime 00:30:00
PFS Group none
• Installed SAs • Keys • Users

Re: permanent «phase 1 negotiation failed»

Fri Jan 12, 2018 5:36 pm

If you can connect using Winbox, you should be able to press the «console» button and get the command line window.

Now as you have just a single peer defined in the

Источник

IPsec

Internet Protocol Security (IPsec) is a set of protocols defined by the Internet Engineering Task Force (IETF) to secure packet exchange over unprotected IP/IPv6 networks such as the Internet.

IPsec protocol suite can be divided into the following groups:

• Internet Key Exchange (IKE) protocols. Dynamically generates and distributes cryptographic keys for AH and ESP.
• Authentication Header (AH) RFC 4302
• Encapsulating Security Payload (ESP) RFC 4303

Internet Key Exchange Protocol (IKE)

The Internet Key Exchange (IKE) is a protocol that provides authenticated keying material for the Internet Security Association and Key Management Protocol (ISAKMP) framework. There are other key exchange schemes that work with ISAKMP, but IKE is the most widely used one. Together they provide means for authentication of hosts and automatic management of security associations (SA).

Most of the time IKE daemon is doing nothing. There are two possible situations when it is activated:

There is some traffic caught by a policy rule which needs to become encrypted or authenticated, but the policy doesn’t have any SAs. The policy notifies the IKE daemon about that, and the IKE daemon initiates a connection to a remote host. IKE daemon responds to remote connection. In both cases, peers establish a connection and execute 2 phases:

• Phase 1 — The peers agree upon algorithms they will use in the following IKE messages and authenticate. The keying material used to derive keys for all SAs and to protect following ISAKMP exchanges between hosts is generated also. This phase should match the following settings:
  • authentication method
  • DH group
  • encryption algorithm
  • exchange mode
  • hash algorithm
  • NAT-T
  • DPD and lifetime (optional)

• Phase 2 — The peers establish one or more SAs that will be used by IPsec to encrypt data. All SAs established by the IKE daemon will have lifetime values (either limiting time, after which SA will become invalid, or the amount of data that can be encrypted by this SA, or both). This phase should match the following settings:
  • IPsec protocol
  • mode (tunnel or transport)
  • authentication method
  • PFS (DH) group
  • lifetime

There are two lifetime values — soft and hard. When SA reaches its soft lifetime threshold, the IKE daemon receives a notice and starts another phase 2 exchange to replace this SA with a fresh one. If SA reaches a hard lifetime, it is discarded.

Phase 1 is not re-keyed if DPD is disabled when the lifetime expires, only phase 2 is re-keyed. To force phase 1 re-key, enable DPD.

PSK authentication was known to be vulnerable against Offline attacks in «aggressive» mode, however recent discoveries indicate that offline attack is possible also in the case of «main» and «ike2» exchange modes. A general recommendation is to avoid using the PSK authentication method.

IKE can optionally provide a Perfect Forward Secrecy (PFS), which is a property of key exchanges, that, in turn, means for IKE that compromising the long term phase 1 key will not allow to easily gain access to all IPsec data that is protected by SAs established through this phase 1. It means an additional keying material is generated for each phase 2.

The generation of keying material is computationally very expensive. Exempli Gratia, the use of the modp8192 group can take several seconds even on a very fast computer. It usually takes place once per phase 1 exchange, which happens only once between any host pair and then is kept for a long time. PFS adds this expensive operation also to each phase 2 exchange.

Diffie-Hellman Groups

Diffie-Hellman (DH) key exchange protocol allows two parties without any initial shared secret to create one securely. The following Modular Exponential (MODP) and Elliptic Curve (EC2N) Diffie-Hellman (also known as «Oakley») Groups are supported:

More on standards can be found here.

IKE Traffic

To avoid problems with IKE packets hit some SPD rule and require to encrypt it with not yet established SA (that this packet perhaps is trying to establish), locally originated packets with UDP source port 500 are not processed with SPD. The same way packets with UDP destination port 500 that are to be delivered locally are not processed in incoming policy checks.

Setup Procedure

To get IPsec to work with automatic keying using IKE-ISAKMP you will have to configure policy, peer, and proposal (optional) entries.

IPsec is very sensitive to time changes. If both ends of the IPsec tunnel are not synchronizing time equally(for example, different NTP servers not updating time with the same timestamp), tunnels will break and will have to be established again.

EAP Authentication methods

PAP
CHAP
MS-CHAP
MS-CHAPv2
EAP-MSCHAPv2
EAP-GTC
EAP-MD5
EAP-TLS

EAP-TLS on Windows is called «Smart Card or other certificates».

AH is a protocol that provides authentication of either all or part of the contents of a datagram through the addition of a header that is calculated based on the values in the datagram. What parts of the datagram are used for the calculation, and the placement of the header depends on whether tunnel or transport mode is used.

The presence of the AH header allows to verify the integrity of the message but doesn’t encrypt it. Thus, AH provides authentication but not privacy. Another protocol (ESP) is considered superior, it provides data privacy and also its own authentication method.

RouterOS supports the following authentication algorithms for AH:

Transport mode

In transport mode, the AH header is inserted after the IP header. IP data and header is used to calculate authentication value. IP fields that might change during transit, like TTL and hop count, are set to zero values before authentication.

Tunnel mode

In tunnel mode, the original IP packet is encapsulated within a new IP packet. All of the original IP packets are authenticated.

Encapsulating Security Payload (ESP)

Encapsulating Security Payload (ESP) uses shared key encryption to provide data privacy. ESP also supports its own authentication scheme like that used in AH.

ESP packages its fields in a very different way than AH. Instead of having just a header, it divides its fields into three components:

• ESP Header — Comes before the encrypted data and its placement depends on whether ESP is used in transport mode or tunnel mode.
• ESP Trailer — This section is placed after the encrypted data. It contains padding that is used to align the encrypted data.
• ESP Authentication Data — This field contains an Integrity Check Value (ICV), computed in a manner similar to how the AH protocol works, for when ESP’s optional authentication feature is used.

Transport mode

In transport mode, the ESP header is inserted after the original IP header. ESP trailer and authentication value are added to the end of the packet. In this mode only the IP payload is encrypted and authenticated, the IP header is not secured.

Tunnel mode

In tunnel mode, an original IP packet is encapsulated within a new IP packet thus securing IP payload and IP header.

Encryption algorithms

RouterOS ESP supports various encryption and authentication algorithms.

• MD5
• SHA1
• SHA2 (256-bit, 512-bit)

• AES — 128-bit, 192-bit, and 256-bit key AES-CBC, AES-CTR, and AES-GCM algorithms;
• Blowfish — added since v4.5
• Twofish — added since v4.5
• Camellia — 128-bit, 192-bit, and 256-bit key Camellia encryption algorithm added since v4.5
• DES — 56-bit DES-CBC encryption algorithm;
• 3DES — 168-bit DES encryption algorithm;

Hardware acceleration

Hardware acceleration allows doing a faster encryption process by using a built-in encryption engine inside the CPU.

List of devices with hardware acceleration is available here

* supported only 128 bit and 256 bit key sizes

** only manufactured since 2016, serial numbers that begin with number 5 and 7

*** AES-CBC and AES-CTR only encryption is accelerated, hashing done in software

**** DES is not supported, only 3DES and AES-CBC

IPsec throughput results of various encryption and hash algorithm combinations are published on the MikroTik products page.

Policies

The policy table is used to determine whether security settings should be applied to a packet.

Properties

Following parameters are used by template:

• group — name of the policy group to which this template is assigned;
• src-address, dst-address — Requested subnet must match in both directions(for example 0.0.0.0/0 to allow all);
• protocol — protocol to match, if set to all, then any protocol is accepted;
• proposal — SA parameters used for this template;
• level — useful when unique is required in setups with multiple clients behind NAT.

tunnel (yes | no; Default: no) Specifies whether to use tunnel mode.

Read-only properties

Policy order is important starting from v6.40. Now it works similarly to firewall filters where policies are executed from top to bottom (priority parameter is removed).

All packets are IPIP encapsulated in tunnel mode, and their new IP header’s src-address and dst-address are set to sa-src-address and sa-dst-address values of this policy. If you do not use tunnel mode (id est you use transport mode), then only packets whose source and destination addresses are the same as sa-src-address and sa-dst-address can be processed by this policy. Transport mode can only work with packets that originate at and are destined for IPsec peers (hosts that established security associations). To encrypt traffic between networks (or a network and a host) you have to use tunnel mode.

Statistics

This menu shows various IPsec statistics and errors.

Read-only properties

Proposals

Proposal information that will be sent by IKE daemons to establish SAs for certain policies.

Properties

Read-only properties

Groups

In this menu, it is possible to create additional policy groups used by policy templates.

Properties

Peers

Peer configuration settings are used to establish connections between IKE daemons. This connection then will be used to negotiate keys and algorithms for SAs. Exchange mode is the only unique identifier between the peers, meaning that there can be multiple peer configurations with the same remote-address as long as a different exchange-mode is used.

Properties

Read-only properties

Identities

Identities are configuration parameters that are specific to the remote peer. The main purpose of identity is to handle authentication and verify the peer’s integrity.

Properties

Read only properties

Profiles

Profiles define a set of parameters that will be used for IKE negotiation during Phase 1. These parameters may be common with other peer configurations.

Properties

Active Peers

This menu provides various statistics about remote peers that currently have established phase 1 connection.

Read only properties

Commands

Mode configs

ISAKMP and IKEv2 configuration attributes are configured in this menu.

Properties

Read-only properties

Not all IKE implementations support multiple split networks provided by the split-include option.

If RouterOS client is initiator, it will always send CISCO UNITY extension, and RouterOS supports only split-include from this extension.

It is not possible to use system-dns and static-dns at the same time.

Installed SAs

This menu provides information about installed security associations including the keys.

Read-only properties

Commands

This menu lists all imported public and private keys, that can be used for peer authentication. Menu has several commands to work with keys.

Properties

Read-only properties

Commands

Settings

Application Guides

RoadWarrior client with NAT

Consider setup as illustrated below. RouterOS acts as a RoadWarrior client connected to Office allowing access to its internal resources.

A tunnel is established, a local mode-config IP address is received and a set of dynamic policies are generated.

Currently, only packets with a source address of 192.168.77.254/32 will match the IPsec policies. For a local network to be able to reach remote subnets, it is necessary to change the source address of local hosts to the dynamically assigned mode config IP address. It is possible to generate source NAT rules dynamically. This can be done by creating a new address list that contains all local networks that the NAT rule should be applied. In our case, it is 192.168.88.0/24.

By specifying the address list under the mode-config initiator configuration, a set of source NAT rules will be dynamically generated.

When the IPsec tunnel is established, we can see the dynamically created source NAT rules for each network. Now every host in 192.168.88.0/24 is able to access Office’s internal resources.

Allow only IPsec encapsulated traffic

There are some scenarios where for security reasons you would like to drop access from/to specific networks if incoming/outgoing packets are not encrypted. For example, if we have L2TP/IPsec setup we would want to drop nonencrypted L2TP connection attempts.

There are several ways how to achieve this:

• Using IPsec policy matcher in firewall;
• Using generic IPsec policy with action set to drop and lower priority (can be used in Road Warrior setups where dynamic policies are generated);
• By setting DSCP or priority in mangle and matching the same values in firewall after decapsulation.

IPsec policy matcher

Let’s set up an IPsec policy matcher to accept all packets that matched any of the IPsec policies and drop the rest:

IPsec policy matcher takes two parameters direction, policy . We used incoming direction and IPsec policy. IPsec policy option allows us to inspect packets after decapsulation, so for example, if we want to allow only GRE encapsulated packet from a specific source address and drop the rest we could set up the following rules:

For L2TP rule set would be:

Using generic IPsec policy

The trick of this method is to add a default policy with an action drop. Let’s assume we are running an L2TP/IPsec server on a public 1.1.1.1 address and we want to drop all nonencrypted L2TP:

Now router will drop any L2TP unencrypted incoming traffic, but after a successful L2TP/IPsec connection dynamic policy is created with higher priority than it is on default static rule, and packets matching that dynamic rule can be forwarded.

Policy order is important! For this to work, make sure the static drop policy is below the dynamic policies. Move it below the policy template if necessary.

Manually specifying local-address parameter under Peer configuration

Using different routing table

IPsec, as any other service in RouterOS, uses the main routing table regardless of what local-address parameter is used for Peer configuration. It is necessary to apply routing marks to both IKE and IPSec traffic.

Consider the following example. There are two default routes — one in the main routing table and another in the routing table «backup». It is necessary to use the backup link for the IPsec site to site tunnel.

IPsec peer and policy configurations are created using the backup link’s source address, as well as the NAT bypass rule for IPsec tunnel traffic.

Currently, we see «phase1 negotiation failed due to time up» errors in the log. It is because IPsec tries to reach the remote peer using the main routing table with an incorrect source address. It is necessary to mark UDP/500, UDP/4500, and ipsec-esp packets using Mangle:

Using the same routing table with multiple IP addresses

Consider the following example. There are multiple IP addresses from the same subnet on the public interface. Masquerade rule is configured on out-interface. It is necessary to use one of the IP addresses explicitly.

IPsec peer and policy configuration is created using one of the public IP addresses.

Currently, the phase 1 connection uses a different source address than we specified, and «phase1 negotiation failed due to time up» errors are shown in the logs. This is because masquerade is changing the source address of the connection to match the pref-src address of the connected route. The solution is to exclude connections from the public IP address from being masqueraded.

Application examples

Site to Site IPsec (IKEv1) tunnel

Consider setup as illustrated below. Two remote office routers are connected to the internet and office workstations are behind NAT. Each office has its own local subnet, 10.1.202.0/24 for Office1 and 10.1.101.0/24 for Office2. Both remote offices need secure tunnels to local networks behind routers.

Site 1 configuration

Start off by creating a new Phase 1 profile and Phase 2 proposal entries using stronger or weaker encryption parameters that suit your needs. It is advised to create separate entries for each menu so that they are unique for each peer in case it is necessary to adjust any of the settings in the future. These parameters must match between the sites or else the connection will not establish.

Continue by configuring a peer . Specify the address of the remote router. This address should be reachable through UDP/500 and UDP/4500 ports, so make sure appropriate actions are taken regarding the router’s firewall. Specify the name for this peer as well as the newly created profile .

The next step is to create an identity . For a basic pre-shared key secured tunnel, there is nothing much to set except for a strong secret and the peer to which this identity applies.

If security matters, consider using IKEv2 and a different auth-method .

Lastly, create a policy that controls the networks/hosts between whom traffic should be encrypted.

Site 2 configuration

Office 2 configuration is almost identical to Office 1 with proper IP address configuration. Start off by creating a new Phase 1 profile and Phase 2 proposal entries:

Next is the peer and identity:

When it is done, create a policy :

At this point, the tunnel should be established and two IPsec Security Associations should be created on both routers:

NAT and Fasttrack Bypass

At this point if you try to send traffic over the IPsec tunnel, it will not work, packets will be lost. This is because both routers have NAT rules (masquerade) that are changing source addresses before a packet is encrypted. A router is unable to encrypt the packet because the source address does not match the address specified in the policy configuration. For more information see the IPsec packet flow example.

To fix this we need to set up IP/Firewall/NAT bypass rule.

Office 1 router:

Office 2 router:

If you previously tried to establish an IP connection before the NAT bypass rule was added, you have to clear the connection table from the existing connection or restart both routers.

It is very important that the bypass rule is placed at the top of all other NAT rules.

Another issue is if you have IP/Fasttrack enabled, the packet bypasses IPsec policies. So we need to add accept rule before FastTrack.

However, this can add a significant load to the router’s CPU if there is a fair amount of tunnels and significant traffic on each tunnel.

The solution is to use IP/Firewall/Raw to bypass connection tracking, that way eliminating the need for filter rules listed above and reducing the load on CPU by approximately 30%.

Site to Site GRE tunnel over IPsec (IKEv2) using DNS

This example explains how it is possible to establish a secure and encrypted GRE tunnel between two RouterOS devices when one or both sites do not have a static IP address. Before making this configuration possible, it is necessary to have a DNS name assigned to one of the devices which will act as a responder (server). For simplicity, we will use RouterOS built-in DDNS service IP/Cloud .

Site 1 (server) configuration

This is the side that will listen to incoming connections and act as a responder. We will use mode config to provide an IP address for the second site, but first, create a loopback (blank) bridge and assign an IP address to it that will be used later for GRE tunnel establishment.

Continuing with the IPsec configuration, start off by creating a new Phase 1 profile and Phase 2 proposal entries using stronger or weaker encryption parameters that suit your needs. Note that this configuration example will listen to all incoming IKEv2 requests, meaning the profile configuration will be shared between all other configurations (e.g. RoadWarrior).

Next, create a new mode config entry with responder=yes . This will provide an IP configuration for the other site as well as the host (loopback address) for policy generation.

It is advised to create a new policy group to separate this configuration from any existing or future IPsec configuration.

Now it is time to set up a new policy template that will match the remote peers new dynamic address and the loopback address.

The next step is to create a peer configuration that will listen to all IKEv2 requests. If you already have such an entry, you can skip this step.

Lastly, set up an identity that will match our remote peer by pre-shared-key authentication with a specific secret .

The server side is now configured and listening to all IKEv2 requests. Please make sure the firewall is not blocking UDP/4500 port.

The last step is to create the GRE interface itself. This can also be done later when an IPsec connection is established from the client-side.

Configure IP address and route to remote network through GRE interface.

Site 2 (client) configuration

Similarly to server configuration, start off by creating a new Phase 1 profile and Phase 2 proposal configurations. Since this site will be the initiator, we can use a more specific profile configuration to control which exact encryption parameters are used, just make sure they overlap with what is configured on the server-side.

Next, create a new mode config entry with responder=no . This will make sure the peer requests IP and split-network configuration from the server.

It is also advised to create a new policy group to separate this configuration from any existing or future IPsec configuration.

Create a new policy template on the client-side as well.

Move on to peer configuration. Now we can specify the DNS name for the server under the address parameter. Obviously, you can use an IP address as well.

Lastly, create an identity for our newly created peers.

If everything was done properly, there should be a new dynamic policy present.

A secure tunnel is now established between both sites which will encrypt all traffic between 192.168.99.2 192.168.99.1 addresses. We can use these addresses to create a GRE tunnel.

Configure IP address and route to remote network through GRE interface.

Road Warrior setup using IKEv2 with RSA authentication

This example explains how to establish a secure IPsec connection between a device connected to the Internet (road warrior client) and a device running RouterOS acting as a server.

RouterOS server configuration

Before configuring IPsec, it is required to set up certificates. It is possible to use a separate Certificate Authority for certificate management, however in this example, self-signed certificates are generated in RouterOS System/Certificates menu. Some certificate requirements should be met to connect various devices to the server:

• Common name should contain IP or DNS name of the server;
• SAN (subject alternative name) should have IP or DNS of the server;
• EKU (extended key usage) tls-server and tls-client are required.

Considering all requirements above, generate CA and server certificates:

Now that valid certificates are created on the router, add a new Phase 1 profile and Phase 2 proposal entries with pfs-group=none:

Mode config is used for address distribution from IP/Pools:

Since that the policy template must be adjusted to allow only specific network policies , it is advised to create a separate policy group and template.

Create a new IPsec peer entry that will listen to all incoming IKEv2 requests.

Identity configuration

The identity menu allows to match specific remote peers and assign different configurations for each one of them. First, create a default identity, that will accept all peers, but will verify the peer’s identity with its certificate.

If the peer’s ID (ID_i) is not matching with the certificate it sends, the identity lookup will fail. See remote-id in the identities section.

For example, we want to assign a different mode config for user «A», who uses certificate «rw-client1» to authenticate itself to the server. First of all, make sure a new mode config is created and ready to be applied for the specific user.

It is possible to apply this configuration for user «A» by using the match-by=certificate parameter and specifying his certificate with remote-certificate .

(Optional) Split tunnel configuration

Split tunneling is a method that allows road warrior clients to only access a specific secured network and at the same time send the rest of the traffic based on their internal routing table (as opposed to sending all traffic over the tunnel). To configure split tunneling, changes to mode config parameters are needed.

For example, we will allow our road warrior clients to only access the 10.5.8.0/24 network.

It is also possible to send a specific DNS server for the client to use. By default, system-dns=yes is used, which sends DNS servers that are configured on the router itself in IP/DNS . We can force the client to use a different DNS server by using the static-dns parameter.

While it is possible to adjust the IPsec policy template to only allow road warrior clients to generate policies to network configured by split-include parameter, this can cause compatibility issues with different vendor implementations (see known limitations ). Instead of adjusting the policy template, allow access to a secured network in IP/Firewall/Filter and drop everything else.

Split networking is not a security measure. The client (initiator) can still request a different Phase 2 traffic selector.

Generating client certificates

To generate a new certificate for the client and sign it with a previously created CA.

PKCS12 format is accepted by most client implementations, so when exporting the certificate, make sure PKCS12 is specified.

A file named cert_export_rw-client1.p12 is now located in the routers System/File section. This file should be securely transported to the client’s device.

Typically PKCS12 bundle contains also a CA certificate, but some vendors may not install this CA, so a self-signed CA certificate must be exported separately using PEM format.

A file named cert_export_ca.crt is now located in the routers System/File section. This file should also be securely transported to the client’s device.

PEM is another certificate format for use in client software that does not support PKCS12. The principle is pretty much the same.

Three files are now located in the routers Files section: cert_export_ca.crt , cert_export_rw-client1.crt and cert_export_rw-client1.key which should be securely transported to the client device.

Known limitations

Here is a list of known limitations by popular client software IKEv2 implementations.

• Windows will always ignore networks received by split-include and request policy with destination 0.0.0.0/0 (TSr). When IPsec-SA is generated, Windows requests DHCP option 249 to which RouterOS will respond with configured split-include networks automatically.

• Both Apple macOS and iOS will only accept the first split-include network.

• Both Apple macOS and iOS will use the DNS servers from system-dns and static-dns parameters only when 0.0.0.0/0 split-include is used.

• While some implementations can make use of different PFS group for phase 2, it is advised to use pfs-group=none under proposals to avoid any compatibility issues.

RouterOS client configuration

Import a PKCS12 format certificate in RouterOS.

There should now be the self-signed CA certificate and the client certificate in the Certificate menu. Find out the name of the client certificate.

cert_export_RouterOS_client.p12_0 is the client certificate.

It is advised to create a separate Phase 1 profile and Phase 2 proposal configurations to not interfere with any existing IPsec configuration.

While it is possible to use the default policy template for policy generation, it is better to create a new policy group and template to separate this configuration from any other IPsec configuration.

Create a new mode config entry with responder=no that will request configuration parameters from the server.

Lastly, create peer and identity configurations.

Verify that the connection is successfully established.

Enabling dynamic source NAT rule generation

If we look at the generated dynamic policies, we see that only traffic with a specific (received by mode config) source address will be sent through the tunnel. But a router in most cases will need to route a specific device or network through the tunnel. In such case, we can use source NAT to change the source address of packets to match the mode config address. Since the mode config address is dynamic, it is impossible to create a static source NAT rule. In RouterOS, it is possible to generate dynamic source NAT rules for mode config clients.

For example, we have a local network 192.168.88.0/24 behind the router and we want all traffic from this network to be sent over the tunnel. First of all, we have to make a new IP/Firewall/Address list which consists of our local network

When it is done, we can assign the newly created IP/Firewall/Address list to the mode config configuration.

Verify correct source NAT rule is dynamically generated when the tunnel is established.

Make sure the dynamic mode config address is not a part of a local network.

Windows client configuration

Open PKCS12 format certificate file on the Windows computer. Install the certificate by following the instructions. Make sure you select the Local Machine store location. You can now proceed to Network and Internet settings -> VPN and add a new configuration. Fill in the Connection name, Server name, or address parameters. Select IKEv2 under VPN type. When it is done, it is necessary to select «Use machine certificates». This can be done in Network and Sharing Center by clicking the Properties menu for the VPN connection. The setting is located under the Security tab.

Currently, Windows 10 is compatible with the following Phase 1 ( profiles ) and Phase 2 ( proposals ) proposal sets:

macOS client configuration

Open the PKCS12 format certificate file on the macOS computer and install the certificate in the «System» keychain. It is necessary to mark the CA certificate as trusted manually since it is self-signed. Locate the certificate macOS Keychain Access app under the System tab and mark it as Always Trust.

You can now proceed to System Preferences -> Network and add a new configuration by clicking the + button. Select Interface: VPN, VPN Type: IKEv2 and name your connection. Remote ID must be set equal to common-name or subjAltName of server’s certificate. Local ID can be left blank. Under Authentication Settings select None and choose the client certificate. You can now test the connectivity.

Currently, macOS is compatible with the following Phase 1 ( profiles) and Phase 2 ( proposals) proposal sets:

iOS client configuration

Typically PKCS12 bundle contains also a CA certificate, but iOS does not install this CA, so a self-signed CA certificate must be installed separately using PEM format. Open these files on the iOS device and install both certificates by following the instructions. It is necessary to mark the self-signed CA certificate as trusted on the iOS device. This can be done in Settings -> General -> About -> Certificate Trust Settings menu. When it is done, check whether both certificates are marked as «verified» under the Settings -> General -> Profiles menu.

You can now proceed to Settings -> General -> VPN menu and add a new configuration. Remote ID must be set equal to common-name or subjAltName of server’s certificate. Local ID can be left blank.

Currently, iOS is compatible with the following Phase 1 ( profiles) and Phase 2 ( proposals) proposal sets:

If you are connected to the VPN over WiFi, the iOS device can go into sleep mode and disconnect from the network.

Android (strongSwan) client configuration

Currently, there is no IKEv2 native support in Android, however, it is possible to use strongSwan from Google Play Store which brings IKEv2 to Android. StrongSwan accepts PKCS12 format certificates, so before setting up the VPN connection in strongSwan, make sure you download the PKCS12 bundle to your Android device. When it is done, create a new VPN profile in strongSwan, type in the server IP, and choose «IKEv2 Certificate» as VPN Type. When selecting a User certificate, press Install and follow the certificate extract procedure by specifying the PKCS12 bundle. Save the profile and test the connection by pressing on the VPN profile.

It is possible to specify custom encryption settings in strongSwan by ticking the «Show advanced settings» checkbox. Currently, strongSwan by default is compatible with the following Phase 1 ( profiles) and Phase 2 ( proposals) proposal sets:

Linux (strongSwan) client configuration

Download the PKCS12 certificate bundle and move it to /etc/ipsec.d/private directory.

Add exported passphrase for the private key to /etc/ipsec.secrets file where «strongSwan_client.p12» is the file name and «1234567890» is the passphrase.

Add a new connection to /etc/ipsec.conf file

You can now restart (or start) the ipsec daemon and initialize the connection

Road Warrior setup using IKEv2 with EAP-MSCHAPv2 authentication handled by User Manager (RouterOS v7)

This example explains how to establish a secure IPsec connection between a device connected to the Internet (road warrior client) and a device running RouterOS acting as an IKEv2 server and User Manager. It is possible to run User Manager on a separate device in network, however in this example both User Manager and IKEv2 server will be configured on the same device (Office).

RouterOS server configuration

Requirements

For this setup to work there are several prerequisites for the router:

1. Router’s IP address should have a valid public DNS record — IP Cloud could be used to achieve this.
2. Router should be reachable through port TCP/80 over the Internet — if the server is behind NAT, port forwarding should be configured.
3. User Manager package should be installed on the router.

Generating Let’s Encrypt certificate

During the EAP-MSCHAPv2 authentication, TLS handshake has to take place, which means the server has to have a certificate that can be validated by the client. To simplify this step, we will use Let’s Encrypt certificate which can be validated by most operating systems without any intervention by the user. To generate the certificate, simply enable SSL certificate under the Certificates menu. By default the command uses the dynamic DNS record provided by IP Cloud, however a custom DNS name can also be specified. Note that, the DNS record should point to the router.

If the certificate generation succeeded, you should see the Let’s Encrypt certificate installed under the Certificates menu.

Configuring User Manager

First of all, allow receiving RADIUS requests from the localhost (the router itself):

Enable the User Manager and specify the Let’s Encrypt certificate (replace the name of the certificate to the one installed on your device) that will be used to authenticate the users.

Lastly add users and their credentials that clients will use to authenticate to the server.

Configuring RADIUS client

For the router to use RADIUS server for user authentication, it is required to add a new RADIUS client that has the same shared secret that we already configured on User Manager.

IPsec (IKEv2) server configuration

Add a new Phase 1 profile and Phase 2 proposal entries with pfs-group=none:

Mode config is used for address distribution from IP/Pools.

Since that the policy template must be adjusted to allow only specific network policies, it is advised to create a separate policy group and template.

Create a new IPsec peer entry which will listen to all incoming IKEv2 requests.

Lastly create a new IPsec identity entry that will match all clients trying to authenticate with EAP. Note that generated Let’s Encrypt certificate must be specified.

(Optional) Split tunnel configuration

Split tunneling is a method that allows road warrior clients to only access a specific secured network and at the same time send the rest of the traffic based on their internal routing table (as opposed to sending all traffic over the tunnel). To configure split tunneling, changes to mode config parameters are needed.

For example, we will allow our road warrior clients to only access the 10.5.8.0/24 network.

It is also possible to send a specific DNS server for the client to use. By default, system-dns=yes is used, which sends DNS servers that are configured on the router itself in IP/DNS . We can force the client to use a different DNS server by using the static-dns parameter.

Split networking is not a security measure. The client (initiator) can still request a different Phase 2 traffic selector.

(Optional) Assigning static IP address to user

Static IP address to any user can be assigned by use of RADIUS Framed-IP-Address attribute.

To avoid any conflicts, the static IP address should be excluded from the IP pool of other users, as well as shared-users should be set to 1 for the specific user.

(Optional) Accounting configuration

To keep track of every user’s uptime, download and upload statistics, RADIUS accounting can be used. By default RADIUS accounting is already enabled for IPsec, but it is advised to configure Interim Update timer that sends statistic to the RADIUS server regularly. If the router will handle a lot of simultaneous sessions, it is advised to increase the update timer to avoid increased CPU usage.

Basic L2TP/IPsec setup

This example demonstrates how to easily set up an L2TP/IPsec server on RouterOS for road warrior connections (works with Windows, Android, iOS, macOS, and other vendor L2TP/IPsec implementations).

RouterOS server configuration

The first step is to enable the L2TP server:

use-ipsec is set to required to make sure that only IPsec encapsulated L2TP connections are accepted.

Now what it does is enables an L2TP server and creates a dynamic IPsec peer with a specified secret.

Care must be taken if static IPsec peer configuration exists.

The next step is to create a VPN pool and add some users.

Now the router is ready to accept L2TP/IPsec client connections.

RouterOS client configuration

For RouterOS to work as L2TP/IPsec client, it is as simple as adding a new L2TP client.

It will automatically create dynamic IPsec peer and policy configurations.

Troubleshooting/FAQ

Phase 1 Failed to get a valid proposal

Peers are unable to negotiate encryption parameters causing the connection to drop. To solve this issue, enable IPSec to debug logs and find out which parameters are proposed by the remote peer, and adjust the configuration accordingly.

In this example, the remote end requires SHA1 to be used as a hash algorithm, but MD5 is configured on the local router. Setting before the column symbol (:) is configured on the local side, parameter after the column symbol (:) is configured on the remote side.

«phase1 negotiation failed due to time up» what does it mean?

There are communication problems between the peers. Possible causes include — misconfigured Phase 1 IP addresses; firewall blocking UDP ports 500 and 4500; NAT between peers not properly translating IPsec negotiation packets. This error message can also appear when a local-address parameter is not used properly. More information available here.

Random packet drops or connections over the tunnel are very slow, enabling packet sniffer/torch fixes the problem?

Problem is that before encapsulation packets are sent to Fasttrack/FastPath, thus bypassing IPsec policy checking. The solution is to exclude traffic that needs to be encapsulated/decapsulated from Fasttrack, see configuration example here.

How to enable ike2?

For basic configuration enabling ike2 is very simple, just change exchange-mode in peer settings to ike2.

fatal NO-PROPOSAL-CHOSEN notify message?

Remote peer sent notify that it cannot accept proposed algorithms, to find the exact cause of the problem, look at remote peers debug logs or configuration and verify that both client and server have the same set of algorithms.

I can ping only in one direction?

A typical problem in such cases is strict firewall, firewall rules allow the creation of new connections only in one direction. The solution is to recheck firewall rules, or explicitly accept all traffic that should be encapsulated/decapsulated.

Can I allow only encrypted traffic?

Yes, you can, see » Allow only IPsec encapsulated traffic» examples.

I enable IKEv2 REAUTH on StrongSwan and got the error ‘initiator did not reauthenticate as requested’

RouterOS does not support rfc4478, reauth must be disabled on StrongSwan.

Источник