Our nameservers stopped loading new zone data after an update-named change from "service named restart" to "service named reload" recently. Per cronmail, everything worked just fine: Fetching external item into 'zones' U zones/mozilla.com/svc/phx1 U zones/mozilla.com/svc/SOA Updated external to revision 40464. ... Reloading named: [ OK ] But it simply didn't load the new SOA data until a "service named restart" was run manually on the server. This affects all nameservers in SCL2 and PHX1.
Also ignores failure result code from named reload: Reloading named: [FAILED] After reload: total 56 ... Implementing puppet code to test sysadmins/dnsconfig-weave/bin/update-named.stage on ns1/2.scl2.stage.svc
Assignee: nobody → rsoderberg
Created attachment 638596 [details] [diff] [review] use rndc freeze+reload+thaw instead of rndc reload, so that dynamic zones are reloaded as well Root cause is that some or all of our zones are considered dynamic, and so "rndc reload" only reloads non-dynamic zones. Attempting to "rndc reload soatest.svc.mozilla.com" logs: rndc: 'reload' failed: dynamic zone to ensure that we reload zones correctly, the following alteration to update-named.stage repairs the issue: - /sbin/service named reload + "$RNDC freeze; $RNDC reload; $RNDC thaw;" This patch implements the above change, and is confirmed working in stage using "soatest.svc.mozilla.com" zone which exists only in stage configs.
deployed to SCL2, PHX1
Status: NEW → RESOLVED
Last Resolved: 6 years ago
Resolution: --- → FIXED
You need to log in before you can comment on or make changes to this bug.