Updates tailscale/corp#23782
Once the node key of the current exit node expires, we should stop attempting to forward DNS queries to it, because a connection cannot be established to an expired node. Here we change the logic in `exitNodeCanProxyDNS` so that it no longer attempts to use an invalid URL `/dns-query` as the DNS resolver if an IP is no longer available in the network map for the current exit node. Instead, we return false, and log that this issue occurred.
This doesn't fix the problem, but it at least mitigates it by allowing DNS queries in the broken state to go to the local default resolver (e.g. 192.168.1.1) instead of endlessly getting forwarded to an DNS server URL that doesn't exist.
Next steps include:
- add a user-facing health warning when this verifies, instead of just a log line
- understand why re-authenticating the exit node doesn't trigger a refresh of the DNS config.
Signed-off-by: Andrea Gottardo <andrea@gottardo.me>
Their callers using Range are all kinda clunky feeling. Iterators
should make them more readable.
Updates #12912
Change-Id: I93461eba8e735276fda4a8558a4ae4bfd6c04922
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
We don't need to error out and continuously reconcile if ProxyClass
has not (yet) been created, once it gets created the ProxyGroup
reconciler will get triggered.
Updates tailscale/tailscale#13406
Signed-off-by: Irbe Krumina <irbe@tailscale.com>
Ensure that .status.podIPs is used to select Pod's IP
in all reconcilers.
Updates tailscale/tailscale#13406
Signed-off-by: Irbe Krumina <irbe@tailscale.com>
As discussed in #13684, base the ProxyGroup's proxy definitions on the same
scaffolding as the existing proxies, as defined in proxy.yaml
Updates #13406
Signed-off-by: Tom Proctor <tomhjp@users.noreply.github.com>
Instead of converting our PortMap struct to a string during marshalling
for use as a key, convert the whole collection of PortMaps to a list of
PortMap objects, which improves the readability of the JSON config while
still keeping the data structure we need in the code.
Updates #13406
Signed-off-by: Tom Proctor <tomhjp@users.noreply.github.com>
Currently egress Services for ProxyGroup only work for Pods and Services
with IPv4 addresses. Ensure that it works on dual stack clusters by reading
proxy Pod's IP from the .status.podIPs list that always contains both
IPv4 and IPv6 address (if the Pod has them) rather than .status.podIP that
could contain IPv6 only for a dual stack cluster.
Updates tailscale/tailscale#13406
Signed-off-by: Irbe Krumina <irbe@tailscale.com>
The default ProxyClass can be set via helm chart or env var, and applies
to all proxies that do not otherwise have an explicit ProxyClass set.
This ensures proxies created by the new ProxyGroup CRD are consistent
with the behaviour of existing proxies
Nearby but unrelated changes:
* Fix up double error logs (controller runtime logs returned errors)
* Fix a couple of variable names
Updates #13406
Signed-off-by: Tom Proctor <tomhjp@users.noreply.github.com>
Rearrange conditionals to reduce indentation and make it a bit easier to read
the logic. Also makes some error message updates for better consistency
with the recent decision around capitalising resource names and the
upcoming addition of config secrets.
Updates #cleanup
Signed-off-by: Tom Proctor <tomhjp@users.noreply.github.com>
It is sometimes necessary to defer initialization steps until the first actual usage
or until certain prerequisites have been met. For example, policy setting and
policy source registration should not occur during package initialization.
Instead, they should be deferred until the syspolicy package is actually used.
Additionally, any errors should be properly handled and reported, rather than
causing a panic within the package's init function.
In this PR, we add DeferredInit, to facilitate the registration and invocation
of deferred initialization functions.
Updates #12687
Signed-off-by: Nick Hill <mykola.khyl@gmail.com>
To avoid warning:
find: warning: you have specified the global option -maxdepth after the argument -type, but global options are not positional, i.e., -maxdepth affects tests specified before it as well as those specified after it. Please specify global options before other arguments.
Fixestailscale/corp#23689
Change-Id: I91ee260b295c552c0a029883d5e406733e081478
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Implements the controller for the new ProxyGroup CRD, designed for
running proxies in a high availability configuration. Each proxy gets
its own config and state Secret, and its own tailscale node ID.
We are currently mounting all of the config secrets into the container,
but will stop mounting them and instead read them directly from the kube
API once #13578 is implemented.
Updates #13406
Signed-off-by: Tom Proctor <tomhjp@users.noreply.github.com>
Like we do for the ones on iOS.
As a bonus, this removes a caller of tsaddr.IsTailscaleIP which we
want to revamp/remove soonish.
Updates #13687
Change-Id: Iab576a0c48e9005c7844ab52a0aba5ba343b750e
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Per my investigation just now, the $HOME environment variable is unset
on the macsys (standalone macOS GUI) variant, but the current working
directory is valid. Look for the environment variable file in that
location in addition to inside the home directory.
Updates #3707
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: I481ae2e0d19b316244373e06865e3b5c3a9f3b88
Extend safeweb.Config with the ability to pass a http.Server that
safeweb will use to server traffic.
Updates corp#8207
Signed-off-by: Patrick O'Doherty <patrick@tailscale.com>
Adds a new reconciler that reconciles ExternalName Services that define a
tailnet target that should be exposed to cluster workloads on a ProxyGroup's
proxies.
The reconciler ensures that for each such service, the config mounted to
the proxies is updated with the tailnet target definition and that
and EndpointSlice and ClusterIP Service are created for the service.
Adds a new reconciler that ensures that as proxy Pods become ready to route
traffic to a tailnet target, the EndpointSlice for the target is updated
with the Pods' endpoints.
Updates tailscale/tailscale#13406
Signed-off-by: Irbe Krumina <irbe@tailscale.com>
The AddSNATRuleForDst rule was adding a new rule each time it was called including:
- if a rule already existed
- if a rule matching the destination, but with different desired source already existed
This was causing issues especially for the in-progress egress HA proxies work,
where the rules are now refreshed more frequently, so more redundant rules
were being created.
This change:
- only creates the rule if it doesn't already exist
- if a rule for the same dst, but different source is found, delete it
- also ensures that egress proxies refresh firewall rules
if the node's tailnet IP changes
Updates tailscale/tailscale#13406
Signed-off-by: Irbe Krumina <irbe@tailscale.com>
Add the ability to specify Strict-Transport-Security options in response
to BrowserMux HTTP requests in safeweb.
Updates https://github.com/tailscale/corp/issues/23375
Signed-off-by: Patrick O'Doherty <patrick@tailscale.com>
Updates tailscale/tailscale#3363
We know `log.tailscale.io` supports TLS 1.3, so we can enforce its usage in the client to shake some bytes off the TLS handshake each time a connection is opened to upload logs.
Signed-off-by: Andrea Gottardo <andrea@gottardo.me>
Not just its code.
Updates tailscale/corp#23584
Change-Id: I8001a675372fe15da797adde22f04488d8683448
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
The new logging in 2dd71e64ac is spammy at shutdown:
Receive func ReceiveIPv6 exiting with error: *net.OpError, read udp [::]:38869: raw-read udp6 [::]:38869: use of closed network connection
Receive func ReceiveIPv4 exiting with error: *net.OpError, read udp 0.0.0.0:36123: raw-read udp4 0.0.0.0:36123: use of closed network connection
Skip it if we're in the process of shutting down.
Updates #10976
Change-Id: I4f6d1c68465557eb9ffe335d43d740e499ba9786
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
We were selectively uploading it, but we were still gathering it,
which can be a waste of CPU.
Also remove a bunch of complexity that I don't think matters anymore.
And add an envknob to force service collection off on a single node,
even if the tailnet policy permits it.
Fixes#13463
Change-Id: Ib6abe9e29d92df4ffa955225289f045eeeb279cf
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
When an Exit Node is used, we create a WFP rule to block all inbound and outbound traffic,
along with several rules to permit specific types of traffic. Notably, we allow all inbound and
outbound traffic to and from LocalRoutes specified in wgengine/router.Config. The list of allowed
routes always includes routes for internal interfaces, such as loopback and virtual Hyper-V/WSL2
interfaces, and may also include LAN routes if the "Allow local network access" option is enabled.
However, these permitting rules do not allow link-local multicast on the corresponding interfaces.
This results in broken mDNS/LLMNR, and potentially other similar issues, whenever an exit node is used.
In this PR, we update (*wf.Firewall).UpdatePermittedRoutes() to create rules allowing outbound and
inbound link-local multicast traffic to and from the permitted IP ranges, partially resolving the mDNS/LLMNR
and *.local name resolution issue.
Since Windows does not attempt to send mDNS/LLMNR queries if a catch-all NRPT rule is present,
it is still necessary to disable the creation of that rule using the disable-local-dns-override-via-nrpt nodeAttr.
Updates #13571
Signed-off-by: Nick Khyl <nickk@tailscale.com>
Of tests we wish we could easily add. One day.
Updates #13038
Change-Id: If44646f8d477674bbf2c9a6e58c3cd8f94a4e8df
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Updates tailscale/tailscale#6148
This is the result of some observations we made today with @raggi. The DNS over HTTPS client currently doesn't cap the number of connections it uses, either in-use or idle. A burst of DNS queries will open multiple connections. Idle connections remain open for 30 seconds (this interval is defined in the dohTransportTimeout constant). For DoH providers like NextDNS which send keep-alives, this means the cellular modem will remain up more than expected to send ACKs if any keep-alives are received while a connection remains idle during those 30 seconds. We can set the IdleConnTimeout to 10 seconds to ensure an idle connection is terminated if no other DNS queries come in after 10 seconds. Additionally, we can cap the number of connections to 1. This ensures that at all times there is only one open DoH connection, either active or idle. If idle, it will be terminated within 10 seconds from the last query.
We also observed all the DoH providers we support are capable of TLS 1.3. We can force this TLS version to reduce the number of packets sent/received each time a TLS connection is established.
Signed-off-by: Andrea Gottardo <andrea@gottardo.me>
1eaad7d3de regressed some tests in another repo that were starting up
a control server on `http://127.0.0.1:nnn`. Because there was no https
running, and because of a bug in 1eaad7d3de (which ended up checking
the recently-dialed-control check twice in a single dial call), we
ended up forcing only the use of TLS dials in a test that only had
plaintext HTTP running.
Instead, plumb down support for explicitly disabling TLS fallbacks and
use it only when running in a test and using `http` scheme control
plane URLs to 127.0.0.1 or localhost.
This fixes the tests elsewhere.
Updates #13597
Change-Id: I97212ded21daf0bd510891a278078daec3eebaa6
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
I noticed while debugging a test failure elsewhere that our failure
logs (when verbosity is cranked up) were uselessly attributing dial
failures to failure to dial an invalid IP address (this IPv6 address
we didn't have), rather than showing me the actual IPv4 connection
failure.
Updates #13597 (tangentially)
Change-Id: I45ffbefbc7e25ebfb15768006413a705b941dae5
Signed-off-by: Brad Fitzpatrick <bradfitz@tailscale.com>
Updates tailscale/tailscale#1634
Updates tailscale/tailscale#13265
Captive portal detection uses a custom `net.Dialer` in its `http.Client`. This custom Dialer ensures that the socket is bound specifically to the Wi-Fi interface. This is crucial because without it, if any default routes are set, the outgoing requests for detecting a captive portal would bypass Wi-Fi and go through the default route instead.
The Dialer did not have a Timeout property configured, so the default system timeout was applied. This caused issues in #13265, where we attempted to make captive portal detection requests over an IPsec interface used for Wi-Fi Calling. The call to `connect()` would fail and remain blocked until the system timeout (approximately 1 minute) was reached.
In #13598, I simply excluded the IPsec interface from captive portal detection. This was a quick and safe mitigation for the issue. This PR is a follow-up to make the process more robust, by setting a 3 seconds timeout on any connection establishment on any interface (this is the same timeout interval we were already setting on the HTTP client).
Signed-off-by: Andrea Gottardo <andrea@gottardo.me>
We were previously not checking that the external IP that we got back
from a UPnP portmap was a valid endpoint; add minimal validation that
this endpoint is something that is routeable by another host.
Updates tailscale/corp#23538
Signed-off-by: Andrew Dunham <andrew@du.nham.ca>
Change-Id: Id9649e7683394aced326d5348f4caa24d0efd532
