| CARVIEW |
Navigation Menu
-
-
Notifications
You must be signed in to change notification settings - Fork 396
Feature/configure outbound msg retry #41
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Feature/configure outbound msg retry #41
Conversation
countsudoku
changed the title
|
Hi Countsudoku, |
Thanks for the info. You can also pick only some commits, since I've committed also some style fixes etc. I am not sure, if it would be better to re-generate the autogenerated files (util/configlexer.c, util/configparser.{h,c}) again with your flex version. Therefor I have committed these files in a separate commit. I am curious: Would it not be better to keep these autogenertaed files out of the repo and create them on demand? |
|
Is there any chance that this will be merged in the near future? |
|
Would love to see this merged as well. In the organization I work in, we sometimes have multiple unbound resolvers "chained" (one running on the client, then another one caching in front of the authorities), and the fact that they each retry 5 times by default on a THROWAWAY response means the latency at the client increases exponentially. If we have a lame server on our network, or elsewhere, this means the client can sometimes have to wait 1+ second(s) before getting a reply, which in some cases causes timeouts. We would love to drop this value to one on "intermediate" resolvers and leave it to the client to make the decision as to retry to not. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
The code looks good; have some smaller fixes I want to do. A lot of whitespace changes (of existing code), that is nicer in separate patches; diff that ignores whitespace can very quickly deal with the review of it. But it is okay here. I'll try to commit and fix the conflicts.
processQueryResponse takes an iterator env argument like other functions in the iterator, no colon in string for set_option, and some whitespace style, to make it similar to the rest.
|
Thanks for the patch! Merged it in the code repository. |
* nlnet/master: (118 commits) - Fix to add example.conf note for outbound-msg-retry. - Implement RFC8375: Special-Use Domain 'home.arpa.'. - Fix crosscompile script for the shared build flags. - Fix crosscompile windows to use libssp when it exists. - For the windows compile script disable gost. - Fix that on windows, use BIO_set_callback_ex instead of deprecated - Fix crosscompile shell syntax. - For crosscompile on windows, detect 64bit stackprotector library. - Fix crosscompile on windows to work with openssl 3.0.0 the link with ws2_32 needs -l:libssp.a for __strcpy_chk. Also copy results from lib64 directory if needed. - Fix more initialisation errors reported by gcc sanitizer. - Fix lock debug code for gcc sanitizer reports. - Fix initialisation errors reported by gcc sanitizer. - Fix root_anchor test to check with new icannbundle date. - Fix for NLnetLabs#41: change outbound retry to int to fix signed comparison warnings. - Small fixes for NLnetLabs#41: changelog, conflicts resolved, processQueryResponse takes an iterator env argument like other functions in the iterator, no colon in string for set_option, and some whitespace style, to make it similar to the rest. Changelog entry for NLnetLabs#538 - Fix NLnetLabs#538: Fix subnetcache statistics. Fix subnetcache statistics - Fix tcp fastopen failure when disabled, try normal connect instead. - Fix NLnetLabs#533: Negative responses get cached even when setting cache-max-negative-ttl: 1 - Fix asynclook unit test for setup of lockchecks before log. - Fix compile warning in libunbound for listen desetup routine. - Fix RPZ locks. Do not unlock zones lock if requested and rpz find zone does not find the zone. Readlock the clientip that is found for ipbased triggers. Unlock the nsdname zone lock when done. Unlock zone and ip in rpz nsip and nsdname callback. Unlock authzone and localzone if clientip found in rpz worker call. ...
I've added a option to configure the number of retries unbound will make in case of a THROWAWAY answer. In case of forward nameservers, this will restrict the number of retries for every nameserver (more precise for every target).
I try to use libunbound in a high performance application. Therefore it is important to restrict the number of request unbound will make.