Michał Kępień commented on merge request !12436 at ISC Open Source Projects / BIND
Okay, 596cbe3a is indeed too complicated, it becomes really hard to grasp what's happening there. Please drop it.
Posts from every feed QUEST has found across the web — blogs, news, podcasts, channels and more.
Okay, 596cbe3a is indeed too complicated, it becomes really hard to grasp what's happening there. Please drop it.
Mind you the %d in the format strings should be %u as the value is unsigned.
There should already be a log message: if ((lctx->options & DNS_MASTER_CHECKTTL) != 0 && lctx->ttl > lctx->maxttl) { (callbacks->error)(callbacks, "dns_master_load: %s:%lu: " "TTL %d exceeds configured " "max-zone-ttl %d", source, line, lctx->ttl, lctx->maxttl); result = ISC_R_RANGE; goto log_and_cleanup; } and a quick look at the system tests shows that it is emitting a message. [ant:~/git/bind9] marka% git grep "exceeds configured" bin/tests/system/checkconf/tests.sh:grep 'TTL 900 exceeds…
Description When (finally!) converting to DNSSEC policy, named startup failed with many errors like these on every zone. Oct 24 16:16:43 hagrid named[5833]: zone mobile.litts.net/IN/internal: loading from master file /opt/var/named/MOBILE_LITTS_NET.DB failed: out of range Oct 24 16:16:43 hagrid named[5833]: zone mobile.litts.net/IN/internal: not loaded due to errors. Oct 24 16:16:43 hagrid named[5833]: zone sd.litts.net/IN/internal: loading from master file /opt/var/named/SD_LITTS_NET.DB…
Michal Nowak (06f5046e) at 24 Oct 20:15
Breakage in the RPM packaging sources is currently only discovered when a release is being packaged. Build the packages in CI as well, from the "tarball-create" artifacts, using the packaging sources and the mock-driving Makefile from the public bind-rpms repository, and smoke-test them by installing them into a pristine chroot and running the installed binaries. Each job downloads several hundred megabytes from external services, so the jobs only run automatically in scheduled pipelines and…
Michal Nowak (31e50bae) at 24 Oct 20:14 [9.20] new: ci: Add jobs building RPM packages in mock ... and 1 more commit
Suggest to move this to backlog. Note it is possible that ALPN becomes at least recommended even with current RADIUS over TLS.
Need to check if there is a new Heimdal: if not this must be addressed before 3.4 release.
Michal Nowak (06f5046e) at 24 Oct 19:22 Add jobs building RPM packages in mock
Warnings ⚠️ This merge request does not have the Review label set. Please set it if you would like the merge request to be reviewed. Generated by Hazard against 06f5046e
bind-team-rw-api-rw-repo (1c061884) at 24 Oct 19:18 Add jobs building RPM packages in mock ... and 3944 more commits
bind-9.20 backport This MR is labeled for backporting into bind-9.20. An automated backport is ready for review: 👀 Continue work in this new MR 👉 !13010 (merged) git fetch --depth=1 origin bind-9.20 From https://gitlab.isc.org/isc-projects/bind9 * branch bind-9.20 -> FETCH_HEAD * [new branch] bind-9.20 -> origin/bind-9.20 git worktree add -b backport-mnowak/rpm-building-ci-jobs-v2-9.20 /tmp/mr12621-bind-9.20 FETCH_HEAD Preparing worktree (new branch 'bac…
Breakage in the RPM packaging sources is currently only discovered when a release is being packaged. Build the packages in CI as well, from the "tarball-create" artifacts, using the packaging sources and the mock-driving Makefile from the public bind-rpms repository, and smoke-test them by installing them into a pristine chroot and running the installed binaries. Each job downloads several hundred megabytes from external services, so the jobs only run automatically in scheduled pipelines and…
Okay, I've now a added test to check that a failure in dns_zone_validatezonemd() prevents a transfer from completing, so that's working. Note that for the moment it's only testing initial AXFR, using a misconfigured trust anchor. We'll also want to have tests that use a fetched parent DS, and IXFR. (@ondrej, perhaps you could ask your robot companion to help identify more test cases?)
Evan Hunt (e6369ae1) at 20 Oct 21:59 fixup! Add a DS zonefetch mechanism
Hello Paul: I should have caught this faster considering I fixed it before under this issue #4176 (closed), which was included in Kea 3.1.5. It has been backported to v3_0 and will be included in 3.0.5 due out end of next week. When trying to replicate the issue I pulled new repos and thought I had changed the branch to Kea-3.0.3 but had not, so I was actually testing master. Thus my inability to reproduce your issue. Once running on actual 3.0.3 code, it reproduced immediately so no need for…
1 Warning ⚠️ GitLab issue #4282 (closed) in msg of commit 99582fb5 and issue #4828 from branch 4828-add-ddns-behavioral-parameters-to-pools-in-config-back-end do not match Generated by 🚫 Danger
Thomas Markwalder (c7a4f03d) at 20 Oct 18:16 [#4828] Fixed constant spelling
Colin Vidal (ceb72e6e) at 20 Oct 16:44 Hacky experiments around EDT
@alessio, I've added one more extra commit on top of your branch that saves a lot of copying on unsuccessful paths. Could you retest on your better perflab instance?
should this check be done after checking parter connectivity? aka switch to 'partner-down' before we switch to 'down' state? I think not...
1 Warning ⚠️ GitLab issue #4282 (closed) in msg of commit 99582fb5 and issue #4828 from branch 4828-add-ddns-behavioral-parameters-to-pools-in-config-back-end do not match Generated by 🚫 Danger
Thomas Markwalder (45dfa328) at 20 Oct 16:10 [#4828] Addressed review comments
https://docs.gitlab.com/user/markdown/#escape-color-codes Sample: #4760
Fine so there was no need to update them, closing.