Our repair steps suppress error output by convention, which keeps them quiet but meant we had been inferring failures from side effects. For this investigation we added a…
Read more →A follow-up repair did not run on a test server. The cause was a mismatch with the convention used elsewhere in the same script: other repair steps tolerate…
Read more →Our diagnostics ruled out several common explanations for a directory that would not accept an ownership change, without pointing to a new one. At that stage we chose…
Read more →The diagnostics showed that the preconditions for our repair were all satisfied, and that one of the two directories it touches had been updated while the other had…
Read more →A repair we had added did not appear to take effect on a test mail server, even when re-run. Rather than ship another blind change, we extended the…
Read more →The first real use of the new diagnostics page paid off immediately: webmail’s log directory had the package’s default ownership rather than the one the setup intended, so…
Read more →Phase 142’s key fix was real: the key length was confirmed, corrected and released, and the self-repair step ran successfully against the Mail Only satellite. Reloading the webmail…
Read more →Straight on from Phase 141: with the routing fixed, “Open Webmail” on a Mail Only satellite stopped returning an access error and started returning a server error. The…
Read more →During readiness testing on a multi-server stack, opening webmail for an account whose mail runs on a linked Mail Only server returned a 403. The webmail application was…
Read more →Same bug family as the mail-DNS fix from earlier that day, found while checking that fix live on a split mail and web install: the Email Deliverability page’s…
Read more →Two related pieces of DNS work, both triggered by watching a freshly created Mail Only satellite take over mail for a domain and asking what a proper hosting-panel…
Read more →Linking up a Mail Only server and DNS Only servers alongside a Main install surfaced something that had been missing since Mail Only first shipped: it never got…
Read more →Linking a Mail Only server and switching the active mail server has always resynced every existing mailbox’s routing and auth data immediately. What it never did, and never…
Read more →A small one: the static “nothing here yet” page nginx serves for any domain pointed at a server with no matching hosting account was still showing a generic…
Read more →A batch of real usage reports from a test install, all fixed in one sitting rather than piecemeal: nine separate issues spanning account creation, dashboards, safety guardrails and…
Read more →A different kind of audit from the fifteen before it. Every earlier one reviewed new code written since the previous pass. This one went back and checked the…
Read more →Phase 131 built the plumbing: the admin alert service, six new health-alert events, and webhook formats to carry them anywhere. What it did not build was one place…
Read more →The last item of the parallel build alongside Phase 131: the File Manager could list, download, upload, rename and delete files, but never let anyone see or change…
Read more →Audit #15 (Phase 130) closed out with the panel clean. Before the next release, one gap kept surfacing across almost every health-adjacent feature: detection without delivery. WAF health…
Read more →The first audit run explicitly as a readiness check ahead of a release, rather than only a post-feature diff. Scope widened past the usual code review to include…
Read more →