Commit Graph

87 Commits

Author SHA1 Message Date
Pierre Souchay 5b8a7d7127 Avoid to have infinite recursion in DNS lookups when resolving CNAMEs (#4918)
* Avoid to have infinite recursion in DNS lookups when resolving CNAMEs

This will avoid killing Consul when a Service.Address is using CNAME
to a Consul CNAME that creates an infinite recursion.

This will fix https://github.com/hashicorp/consul/issues/4907

* Use maxRecursionLevel = 3 to allow several recursions
2019-01-07 16:53:54 -05:00
Paul Banks 0962e95e85
bugfix: use ServiceTags to generate cache key hash (#4987)
* bugfix: use ServiceTags to generate cahce key hash

* update unit test

* update

* remote print log

* Update .gitignore

* Completely deprecate ServiceTag field internally for clarity

* Add explicit test for CacheInfo cases
2019-01-07 21:30:47 +00:00
Pierre Souchay a72f92cac6 dns: implements prefix lookups for DNS TTL (#4605)
This will fix https://github.com/hashicorp/consul/issues/4509 and allow forinstance lb-* to match services lb-001 or lb-service-007.
2018-10-19 08:41:04 -07:00
Pierre Souchay 42f250fa53 Added SOA configuration for DNS settings. (#4714)
This will allow to fine TUNE SOA settings sent by Consul in DNS responses,
for instance to be able to control negative ttl.

Will fix: https://github.com/hashicorp/consul/issues/4713

# Example

Override all settings:

* min_ttl: 0 => 60s
* retry: 600 (10m) => 300s (5 minutes),
* expire: 86400 (24h) => 43200 (12h)
* refresh: 3600 (1h) => 1800 (30 minutes)

```
consul agent -dev -hcl 'dns_config={soa={min_ttl=60,retry=300,expire=43200,refresh=1800}}'
```

Result:
```
dig +multiline @localhost -p 8600 service.consul

; <<>> DiG 9.12.1 <<>> +multiline @localhost -p 8600 service.consul
; (2 servers found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NXDOMAIN, id: 36557
;; flags: qr aa rd; QUERY: 1, ANSWER: 0, AUTHORITY: 1, ADDITIONAL: 1
;; WARNING: recursion requested but not available

;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;service.consul.		IN A

;; AUTHORITY SECTION:
consul.			0 IN SOA ns.consul. hostmaster.consul. (
				1537959133 ; serial
				1800       ; refresh (30 minutes)
				300        ; retry (5 minutes)
				43200      ; expire (12 hours)
				60         ; minimum (1 minute)
				)

;; Query time: 4 msec
;; SERVER: 127.0.0.1#8600(127.0.0.1)
;; WHEN: Wed Sep 26 12:52:13 CEST 2018
;; MSG SIZE  rcvd: 93
```
2018-10-10 15:50:56 -04:00
Matt Keeler 19d71c6eb4
Add ECS option to EDNS responses where appropriate (#4647)
This implements parts of RFC 7871 where Consul is acting as an authoritative name server (or forwarding resolver when recursors are configured)

If ECS opt is present in the request we will mirror it back and return a response with a scope of 0 (global) or with the same prefix length as the request (indicating its valid specifically for that subnet).

We only mirror the prefix-length (non-global) for prepared queries as those could potentially use nearness checks that could be affected by the subnet. In the future we could get more sophisticated with determining the scope bits and allow for better caching of prepared queries that don’t rely on nearness checks.

The other thing this does not do is implement the part of the ECS RFC related to originating ECS headers when acting as a intermediate DNS server (forwarding resolver). That would take a quite a bit more effort and in general provide very little value. Consul will currently forward the ECS headers between recursors and the clients transparently, we just don't originate them for non-ECS clients to get potentially more accurate "location aware" results.
2018-09-11 09:37:46 -04:00
Pierre Souchay 473e589d86 Implementation of Weights Data structures (#4468)
* Implementation of Weights Data structures

Adding this datastructure will allow us to resolve the
issues #1088 and #4198

This new structure defaults to values:
```
   { Passing: 1, Warning: 0 }
```

Which means, use weight of 0 for a Service in Warning State
while use Weight 1 for a Healthy Service.
Thus it remains compatible with previous Consul versions.

* Implemented weights for DNS SRV Records

* DNS properly support agents with weight support while server does not (backwards compatibility)

* Use Warning value of Weights of 1 by default

When using DNS interface with only_passing = false, all nodes
with non-Critical healthcheck used to have a weight value of 1.
While having weight.Warning = 0 as default value, this is probably
a bad idea as it breaks ascending compatibility.

Thus, we put a default value of 1 to be consistent with existing behaviour.

* Added documentation for new weight field in service description

* Better documentation about weights as suggested by @banks

* Return weight = 1 for unknown Check states as suggested by @banks

* Fixed typo (of -> or) in error message as requested by @mkeeler

* Fixed unstable unit test TestRetryJoin

* Fixed unstable tests

* Fixed wrong Fatalf format in `testrpc/wait.go`

* Added notes regarding DNS SRV lookup limitations regarding number of instances

* Documentation fixes and clarification regarding SRV records with weights as requested by @banks

* Rephrase docs
2018-09-07 15:30:47 +01:00
Siva Prasad dcd7d9b015
DNS : Fixes recursors answering the DNS query to properly return the correct response. (#4461)
* Fixes the DNS recursor properly resolving the requests

* Added a test case for the recursor bug

* Refactored code && added a test case for all failing recursors

* Inner indentation moved into else if check
2018-08-02 10:12:52 -04:00
Matt Keeler 86ce52d0d3 Merge remote-tracking branch 'origin/master' into bugfix/prevent-multi-cname 2018-07-10 10:26:45 -04:00
Matt Keeler a26deb44cf Ensure TXT RRs always end up in the Additional section except for ANY or TXT queries
This also changes where the enforcement of the enable_additional_node_meta_txt configuration gets applied.

formatNodeRecord returns the main RRs and the meta/TXT RRs in separate slices. Its then up to the caller to add to the appropriate sections or not.
2018-07-09 12:30:11 -04:00
Matt Keeler e9390fb5c7 Refactor to make this much less confusing 2018-07-03 11:04:19 -04:00
Matt Keeler 4d1bdd8fdb Add a bunch of comments about preventing multi-cname
Hopefully this a bit clearer as to the reasoning
2018-07-03 10:32:52 -04:00
Matt Keeler 22cc44877d Fix some edge cases and add some tests. 2018-07-02 16:58:52 -04:00
Matt Keeler e3859b4f04 Only allow 1 CNAME when querying for a service.
This just makes sure that if multiple services are registered with unique service addresses that we don’t blast back multiple CNAMEs for the same service DNS name and keeps us within the DNS specs.
2018-07-02 16:12:06 -04:00
mkeeler 1da3c42867 Merge remote-tracking branch 'connect/f-connect' 2018-06-25 19:42:51 +00:00
Kyle Havlovitz f3089a6647 connect/ca: undo the interface changes and use sign-self-issued in Vault 2018-06-25 12:25:42 -07:00
Matt Keeler b9d1e7042a Make filtering out TXT RRs only apply when they would end up in Additional section
ANY queries are no longer affected.
2018-06-19 10:08:16 -04:00
Mitchell Hashimoto a5fe6204d5
agent: working DNS for Connect queries, I think, but have to
implement Health endpoints to be sure
2018-06-14 09:41:47 -07:00
Matt Keeler 6604828009 Add configuration entry to control including TXT records for node meta in DNS responses
If set to false, the only way to retrieve TXT records for node meta is to specifically query for TXT records.
2018-06-11 11:49:04 -04:00
Matt Keeler 1c577b2012
Merge pull request #4131 from pierresouchay/enable_full_dns_compression
Enable full dns compression
2018-06-01 10:42:03 -04:00
Pierre Souchay 544acdf04e Fixed comments for max DNS records returned as requested by @mkeeler 2018-05-31 18:15:52 +02:00
Wim cea77e8825 Do reverse service lookup only if address doesn't match node 2018-05-21 22:27:41 +02:00
Wim 88514d6a82 Add support for reverse lookup of services 2018-05-19 19:39:02 +02:00
Pierre Souchay 74cbe5ac85 Ensure to never send messages more than 64k 2018-05-16 12:47:35 +02:00
Pierre Souchay 5f529c9ea7 Fixed unit tests and updated limits 2018-05-16 12:11:49 +02:00
Pierre Souchay 7e8878df0b Re-Enable compression while computing Len(), so we can send more answers
This will fix https://github.com/hashicorp/consul/issues/4071
2018-05-16 11:00:51 +02:00
Kyle Havlovitz 60307ef328
Remove deprecated metric names 2018-05-08 16:23:15 -07:00
Pierre Souchay a3b028d1d7 Removed unecessary copy of Extra and index 2018-04-20 22:51:04 +02:00
Pierre Souchay 7db49828bd Fixed sync of Extra in binarySearch 2018-04-18 14:17:44 +02:00
Pierre Souchay f2fc163b92 Added Unit tests + fixed boudary limit 2018-04-17 09:31:30 +02:00
Pierre Souchay 146152170f Added comment for function dnsBinaryTruncate 2018-04-17 01:10:52 +02:00
Pierre Souchay 9a819b5b29 Perform a binary search to find optimal size of DNS responses
Will fix https://github.com/hashicorp/consul/issues/4036

Instead of removing one by one the entries, find the optimal
size using binary search.

For SRV records, with 5k nodes, duration of DNS lookups is
divided by 4 or more.
2018-04-17 00:50:00 +02:00
Kyle Havlovitz be10300d06
Update make static-assets goal and run format 2018-04-13 09:57:25 -07:00
Matt Keeler d604642792 GH-3798: More PR Updates
Update docs a little
Update/add tests. Make sure all the various ways of determining the source IP work
Update X-Forwarded-For header parsing. This can be a comma separated list with the first element being the original IP so we now handle csv data there.
Got rid of error return from sourceAddrFromRequest
2018-04-12 10:40:46 -04:00
Matt Keeler 3a0f7789ec GH-3798: A few more PR updates 2018-04-11 20:32:35 -04:00
Matt Keeler de3a9be3d0 GH-3798: Updates for PR
Allow DNS peer IP as the source IP.
Break early when the right node was found for executing the preapred query.
Update docs
2018-04-11 17:02:04 -04:00
Matt Keeler 89cd24aeca GH-3798: Add near=_ip support for prepared queries 2018-04-10 14:50:50 -04:00
Matt Keeler c0b1fb6ede
Merge pull request #3948 from pierresouchay/fix_tcp_dns_limit
[BUGFIX] do not break when TCP DNS answer exceeds 64k
2018-03-30 16:25:23 -04:00
Matt Keeler fd9297ad8f Formatting update 2018-03-27 16:31:27 -04:00
Matt Keeler 2d8a68cce9 GH-3854: Warn when node name isnt a valid DNS label 2018-03-27 15:00:33 -04:00
Pierre Souchay 871b9907cb Optimize size for SRV records, should improve performance a bit
Stricter Unit tests that checks if truncation was OK.
2018-03-09 18:25:29 +01:00
Pierre Souchay c3713dbbf1 Performance optimization for services having more than 2k records 2018-03-08 00:26:41 +01:00
Pierre Souchay 1085d5a7b4 Avoid issue with compression of DNS messages causing overflow 2018-03-07 23:33:41 +01:00
Pierre Souchay b672707552 64000 max limit to DNS messages since there is overhead
Added debug log to give information about truncation.
2018-03-07 16:14:41 +01:00
Pierre Souchay 06afb4d02c [BUGFIX] do not break when TCP DNS answer exceeds 64k
It will avoid having discovery broken when having large number
of instances of a service (works with SRV and A* records).

Fixes https://github.com/hashicorp/consul/issues/3850
2018-03-07 10:08:06 +01:00
Pierre Souchay 09970479b5 Allow to control the number of A/AAAA Record returned by DNS
This allows to have randomized resource records (i.e. each
answer contains only one IP, but the IP changes every request) for
A, AAAA records.

It will fix https://github.com/hashicorp/consul/issues/3355 and
https://github.com/hashicorp/consul/issues/3937

See https://github.com/hashicorp/consul/issues/3937#issuecomment-370610509
for details.

It basically add a new option called `a_record_limit` and will not
return more than a_record_limit when performing A, AAAA or ANY DNS
requests.

The existing `udp_answer_limit` option is still working but should
be considered as deprecated since it works only with DNS clients
not supporting EDNS.
2018-03-06 02:07:42 +01:00
Andrei Burd dbb010c865 adding human readability for dns requests debug log (#3751) 2018-02-11 09:02:28 -06:00
Veselkov Konstantin c2395d9bd0 fix refactoring 2018-01-28 22:53:30 +04:00
Veselkov Konstantin 05666113a4 remove golint warnings 2018-01-28 22:40:13 +04:00
Frank Schröder d26b0406e4 dns: return NXDOMAIN if datacenter is invalid (#3200) (#3596)
Queries to the DNS server can contain an optional datacenter
name in the query name. You can query for 'foo.service.consul'
or 'foo.service.dc.consul' to get a response for either the
default or a specific datacenter.

Datacenter names cannot have dots, therefore the datacenter
name can refer to only one element in the DNS query name.

The DNS server allowed extra labels between the optional
datacenter name and the domain and returned a valid response
instead of returning NXDOMAIN. For example, if the domain
is set to '.consul' then 'foo.service.dc1.extra.consul'
should return NXDOMAIN because of 'extra' being between
the datacenter name 'dc1' and the domain '.consul'.

Fixes #3200
2017-10-20 16:49:17 -07:00
Ryan Slade 6f05ea91a3 Replace time.Now().Sub(x) with time.Since(x) 2017-10-17 20:38:24 +02:00