novox/hq
3
0
Fork 0
Code Issues Pull Requests 2 Releases Wiki Activity

Commit Graph

  • 8783a13448 Merge pull request 'Design 28: the mesh runs on the new bus' (#155) from design/28-the-mesh-runs-on-nats into main main mesh-admin 2026-09-28 00:40:29 +00:00
  • a31cfcf461 Design 28: 5.3 is built and was used for the hand-over design/28-the-mesh-runs-on-nats jschoubben 2026-09-28 02:28:06 +02:00
  • 75b3861911 Design 28: the mesh runs on the new bus jschoubben 2026-09-28 02:27:49 +02:00
  • d7d6f2eda0 Design 28: the move needs a credential and a membership for machines already enrolled design/28-the-move-needs-a-membership-for-enrolled-nodes jschoubben 2026-09-27 23:47:45 +02:00
  • 694555214a Merge pull request 'Design 26: which assignment holds a seat is on record, and changes as one act' (#153) from design/26-a-seat-is-held-on-record into main jschoubben 2026-09-27 21:22:56 +00:00
  • a7249541df Design 26: which assignment holds a seat is on record, and changes as one act design/26-a-seat-is-held-on-record jschoubben 2026-09-27 23:20:36 +02:00
  • 95d8253f71 Merge pull request 'ADR 0131: everything on the mesh speaks to the broker seat, and AMQP is not a provision' (#152) from decision/0131-everything-speaks-to-the-broker-seat into main jschoubben 2026-09-27 21:06:02 +00:00
  • 784b487bf9 ADR 0131: everything on the mesh speaks to the broker seat, and AMQP is not a provision decision/0131-everything-speaks-to-the-broker-seat jschoubben 2026-09-27 23:03:05 +02:00
  • af6a397840 Issue 131: the trigger cannot be one build per changed module fix/131-nothing-tells-the-mesh-its-source-moved jschoubben 2026-09-27 21:08:57 +02:00
  • 748f95d417 Issue 131: nothing tells the mesh a source moved jschoubben 2026-09-27 19:51:09 +02:00
  • 7ae711ba0b Merge pull request 'Building the bus: the decisions the work needed, and what it taught back' (#150) from feat/nats-genesis into main jschoubben 2026-09-27 17:06:40 +00:00
  • 51739c4302 The order the repositories land in is part of the rollout feat/nats-genesis jschoubben 2026-09-27 19:04:45 +02:00
  • 5cac268457 Two checks were wrong about where a seat is judged jschoubben 2026-09-27 18:57:49 +02:00
  • ce6ae943b7 Merge main: renumber this branch's records around the trunk's jschoubben 2026-09-27 18:23:41 +02:00
  • 9ef9830dcf ADR 0122: the predecessor is ending, and its broker goes with it jschoubben 2026-09-27 18:11:11 +02:00
  • dfac01a6dd 5.2: the readiness half is in, and what a failed move actually costs jschoubben 2026-09-27 17:59:26 +02:00
  • 51f5ce5c0f 4.5 done: the catch-up needed nothing built, which was the answer jschoubben 2026-09-27 17:23:49 +02:00
  • fc64a2c1a4 4.4 done: a person's account and their client jschoubben 2026-09-27 17:07:28 +02:00
  • 9fc4e74b0d Merge pull request 'to-be 31: a module declares its fail2ban jail, mesh composes them per node' (#149) from design/a-module-declares-its-jail into main jschoubben 2026-09-27 15:00:16 +00:00
  • 225dfa9451 to-be 31: a module declares its fail2ban jail, mesh composes them per node jschoubben 2026-09-27 16:59:56 +02:00
  • bfefb1dbdb 4.3: the installer can raise a mesh on the new bus jschoubben 2026-09-27 16:39:48 +02:00
  • 85c0a3e567 4.2 done: a build is work submitted to a role, on both buses jschoubben 2026-09-27 16:02:17 +02:00
  • a39765c924 Merge pull request 'ADR 0122: a seat is data the controller owns; a rename is a database update' (#148) from design/a-seat-is-data into main jschoubben 2026-09-27 13:41:10 +00:00
  • a8921fe737 ADR 0122: a seat is data the controller owns; a rename is a database update jschoubben 2026-09-27 15:40:55 +02:00
  • c8f430290e ADR 0121: a seat carries the protocol of its role jschoubben 2026-09-27 15:24:17 +02:00
  • d98d6fca11 Merge pull request 'to-be 30: the mesh updates itself on a push' (#147) from design/the-mesh-updates-itself into main jschoubben 2026-09-27 12:49:32 +00:00
  • b8cdfce16d to-be 30: the mesh updates itself on a push jschoubben 2026-09-27 14:49:02 +02:00
  • c4a8455e2e Issue 127 resolved; design 29 says what wildcards are and how the rule is checked jschoubben 2026-09-27 14:45:08 +02:00
  • bf7297b7ed Merge pull request 'ADR 0121: keep distribution, retire only verdaccio; node-* seats migrated' (#146) from design/0121-keep-distribution-fix into main jschoubben 2026-09-27 12:36:54 +00:00
  • 1bb0ef5658 ADR 0121: keep distribution, retire only verdaccio; node-* seats migrated jschoubben 2026-09-27 14:36:33 +02:00
  • 5a9917f802 Merge pull request 'ADR 0121: a system seat is named for its scope; a module may define its own' (#145) from design/system-seats-are-named-by-scope into main jschoubben 2026-09-27 12:15:22 +00:00
  • 8a6ee9177c ADR 0121: a system seat is named for its scope; a module may define its own jschoubben 2026-09-27 14:15:05 +02:00
  • dd577e9ebe 1.7 is done; 4.1 waits on nothing but the bed jschoubben 2026-09-27 03:20:42 +02:00
  • a9b8f41570 Design 25 §4: the server verifies no client certificate, and writes only accounts jschoubben 2026-09-27 02:51:40 +02:00
  • 92a5e8fc05 Merge pull request 'ADR 0120: a roster fact carries its format as a template; rewrite to-be 29' (#144) from design/roster-fact-is-a-template into main jschoubben 2026-09-26 23:51:21 +00:00
  • 8c91ba1cfa 1.7 in progress: the list is derived and the keys are kept jschoubben 2026-09-27 01:50:04 +02:00
  • 0f7f628730 ADR 0120: note the shared/region interaction with hq 128 jschoubben 2026-09-27 01:42:00 +02:00
  • 970da74136 1.7: the composer exists and the composition does not jschoubben 2026-09-27 01:39:20 +02:00
  • 4d4012cdf6 ADR 0120: a roster fact carries its format as a template; rewrite to-be 29 around it jschoubben 2026-09-27 01:33:12 +02:00
  • 0a61531c42 WBS 3.5 done; 4.1 waits on one thing, an enrolment user per token jschoubben 2026-09-27 01:32:26 +02:00
  • 4b0b659084 Merge pull request 'ADR 0119: a taken tunnel's predecessor is retired once the take is proven' (#143) from decision/0119-a-taken-tunnels-predecessor-is-retired into main jschoubben 2026-09-26 22:59:15 +00:00
  • 0042ca9258 0119: link ADR 0118 now that it is on main jochen 2026-09-27 00:58:55 +02:00
  • 63c19456b4 0119 review: rollback needs the private network unassigned first; a configuration written back is retired again with the first original kept; only the interface's own file, never a link jochen 2026-09-27 00:53:13 +02:00
  • 2f195d501e to-be 08: the found tunnel's configuration is retired once the take is proven (ADR 0119) jochen 2026-09-27 00:40:57 +02:00
  • 8a78ff4efe ADR 0119: a taken tunnel's predecessor is retired once the take is proven jochen 2026-09-27 00:40:31 +02:00
  • 762300a380 Merge pull request 'issue 130: undeclaring a service stops it, even one the mesh only reloads or keeps running' (#142) from issue/130-undeclaring-a-service-stops-it into main jschoubben 2026-09-26 22:58:43 +00:00
  • 248c99ca6c 0118/130: link ADR 0117 now that it is on main jochen 2026-09-27 00:58:17 +02:00
  • 13208f0f48 0118: give a unit back the state it was found in — never-stop broke the converge rollback; process removal found and fixed jochen 2026-09-27 00:21:18 +02:00
  • ebd19c4c6c ADR 0118: undeclaring removes what the mesh made, gives back what it changed, leaves the machine's units as they are — resolves issue 130 jochen 2026-09-27 00:16:01 +02:00
  • dd4cbabffb 130: ADR 0117 named, not linked, until it is on main jochen 2026-09-27 00:08:23 +02:00
  • 5b76a09da6 issue 130: undeclaring a service stops it, even one the mesh only reloads or keeps running jochen 2026-09-27 00:08:09 +02:00
  • a363a605cb Merge pull request 'issue 129: nothing makes a machine trust the mesh's own certificate authority' (#141) from issue/129-nothing-makes-a-machine-trust-the-meshs-authority into main jschoubben 2026-09-26 22:57:30 +00:00
  • 24835ab710 Merge pull request 'issue 128: the machine's hosts file is written whole, and on a workstation it is shared' (#140) from issue/128-the-hosts-file-is-written-whole into main jschoubben 2026-09-26 22:57:02 +00:00
  • ba397d4cbe Merge pull request 'ADR 0117: a machine's uplink is a seat — the mesh configures the manager, never the link' (#139) from decision/0117-the-uplink-is-a-seat into main jschoubben 2026-09-26 22:56:27 +00:00
  • f555d523c7 WBS 3.4 is done both halves; issue 127 holds 4.2, 4.3 and catch-up jschoubben 2026-09-27 00:55:07 +02:00
  • 4f93d304d7 WBS: a person's account is done; the client is not blocked by step 3 jschoubben 2026-09-27 00:18:05 +02:00
  • d898bd87e8 Step 5.4 was wrong from 0119 onward; removed jschoubben 2026-09-27 00:16:45 +02:00
  • 7e4da874a9 Design 25 §2: the eaten reply address is verified, not assumed jschoubben 2026-09-27 00:15:00 +02:00
  • 53092020eb WBS: asking a tool is through the seam; a build is a different shape jschoubben 2026-09-27 00:11:44 +02:00
  • df4a3538c3 0117 review: a holder's service declares no state — the manager's lifecycle is the machine's; a start-only setting's gap on a fresh machine jochen 2026-09-27 00:07:42 +02:00
  • 00817fb9e3 Design 25: the store window, and what moving it into the server changes jschoubben 2026-09-27 00:06:15 +02:00
  • 90b8eeff6b 129 review: the example name is a routed name, not a doubled suffix jochen 2026-09-27 00:03:36 +02:00
  • a865fc7d79 128 review: located; the fix as built (block, at, never held, order) jochen 2026-09-27 00:03:30 +02:00
  • c3313f6e17 0117 review: seat table row + decision cited, networkd/dhcpcd lines match the modules, block placement, dispatcher scope, references jochen 2026-09-27 00:03:15 +02:00
  • 9946e852e1 WBS: 3.5's outbound half is in jschoubben 2026-09-27 00:01:56 +02:00
  • 60a53f9b18 issue 129: nothing makes a machine trust the mesh's own certificate authority jochen 2026-09-26 23:38:30 +02:00
  • ee31f9f761 issue 128: the machine's hosts file is written whole, and on a workstation it is shared jochen 2026-09-26 23:38:08 +02:00
  • 0e066473b3 ADR 0117 accepted; a manager that cannot reload takes the setting at its next start (dhcpcd, measured) jochen 2026-09-26 23:44:50 +02:00
  • 504adef221 ADR 0117: a machine's uplink is a seat — the mesh configures the manager, never the link jochen 2026-09-26 23:37:48 +02:00
  • 9f6aa7ea9c The bus is the mesh's centre, not a transport that replaced one jschoubben 2026-09-26 23:54:09 +02:00
  • 8deecf620a Merge pull request 'to-be 29: a node has operator accounts, and the mesh owns what lives under a home' (#138) from design/29-a-node-has-operator-accounts into main jschoubben 2026-09-26 21:53:39 +00:00
  • 64bbdc30c1 to-be 29: a node has operator accounts, and the mesh owns what lives under a home jschoubben 2026-09-26 23:53:13 +02:00
  • 672c994afa WBS: 3.4's seam is in, outbound half through it jschoubben 2026-09-26 23:47:40 +02:00
  • 2f9bb73685 WBS: the first fixtures are in, and what byte-for-byte means jschoubben 2026-09-26 23:41:15 +02:00
  • 5c193b3f54 WBS: 3.8's check is written jschoubben 2026-09-26 23:34:42 +02:00
  • fbf9440e8e WBS: 3.7 and 3.8 done, with the one check 3.8 still owes jschoubben 2026-09-26 23:34:04 +02:00
  • 9510bf5311 WBS: 3.2 done jschoubben 2026-09-26 23:33:10 +02:00
  • d940e14ec8 Design 19: the protocol on NATS jschoubben 2026-09-26 23:32:48 +02:00
  • 80456981be WBS: 3.6 done, and the certificate constraint it surfaced jschoubben 2026-09-26 23:29:56 +02:00
  • d2ed3152d3 Seat renames done; 0118 was wrong that it was a migration jschoubben 2026-09-26 23:06:55 +02:00
  • 24d99ddd24 WBS: 3.9 done, and 1.4's client with it jschoubben 2026-09-26 22:28:59 +02:00
  • b5730525fe Merge pull request 'issue 127: a declaration that shrinks to empty is skipped, not sent' (#137) from issue/127-shrink-to-empty into main jschoubben 2026-09-26 20:25:19 +00:00
  • bb334e138b issue 127: a declaration that shrinks to empty is skipped, so the node keeps what it should drop jschoubben 2026-09-26 22:24:58 +02:00
  • b759e36bfd Design 29: tools, not serves; WBS 3.9 partly done jschoubben 2026-09-26 22:17:17 +02:00
  • 0b8e84334f Why module events share one stream, checked against the server jschoubben 2026-09-26 21:49:55 +02:00
  • 39c802cbd4 Step 2: adoption recreates the bus once, on purpose jschoubben 2026-09-26 21:40:02 +02:00
  • 86a084b7ff The mesh bus is required, not ambient jschoubben 2026-09-26 21:17:21 +02:00
  • 85f972749a AMQP is a provision, not the bus jschoubben 2026-09-26 21:08:40 +02:00
  • 7abb268de6 Step 1 done but for its bed jschoubben 2026-09-26 21:03:35 +02:00
  • 78a2274baf Designs 25 and 29 disagreed about the subject space; implementing found it jschoubben 2026-09-26 21:02:18 +02:00
  • 3f9b316015 Design 25: a scoped inbox needs allow_responses, or nothing can answer jschoubben 2026-09-26 20:58:34 +02:00
  • e05825a881 Design 29: versioning, provisioning and secrets on the bus jschoubben 2026-09-26 20:44:16 +02:00
  • 7b4916e9ec Modules declare their own seats; the mesh reserves mesh-* jschoubben 2026-09-26 20:34:32 +02:00
  • 3f14264c7b Merge pull request '112 is fixed: a carried peer is nameable, and the resolver answers for all four machines' (#136) from issue/112-fixed into main jschoubben 2026-09-26 18:13:53 +00:00
  • 25d599094f 112 is fixed: a carried peer is nameable, and the resolver answers for all four machines jschoubben 2026-09-26 20:13:37 +02:00
  • b5b68e8852 Design 25: a host directory bind, not a named volume jschoubben 2026-09-26 19:34:54 +02:00
  • 814c9e563f The bus is the only broker; step 1 starts jschoubben 2026-09-26 19:26:07 +02:00
  • db4ca9b043 Merge pull request 'The bus in five steps: a decomposition, a breakdown, and progressive insight' (#134) from feat/the-bus-in-five-steps into main jschoubben 2026-09-26 17:05:43 +00:00
  • 5a6d0e111d Merge remote-tracking branch 'origin/main' into feat/the-bus-in-five-steps jschoubben 2026-09-26 19:05:36 +02:00
  • c94e2ece53 Merge pull request 'Give 0115 a home, and match its batch's status — main is failing both checks' (#135) from fix/0115-checks-on-main into main jschoubben 2026-09-26 17:04:50 +00:00