sync: auto-sync from GURU-5070 at 2026-07-18 06:12:00

Author: Mike Swanson
Machine: GURU-5070
Timestamp: 2026-07-18 06:12:00
This commit is contained in:
2026-07-18 06:12:30 -07:00
parent 5ffa384e69
commit 62eda5d3f6

View File

@@ -82,3 +82,34 @@ Earlier in the day (separate committed log `clients/azcomputerguru.com/session-l
- BB SharePoint: Syncro #32187 (id 109277420), comment id 424114287; Donor Services site /sites/donorservices, group `11d28c77-c259-4429-8aab-a3a3945f58fd`
- Jupiter archives: /mnt/user/Backups/websvr-archives/ — compoundfitness (befef13f…), grabbanddurando (f9a06482…), tucsonsafety (b78b0799…), devgrabblaw (92641a2f…)
- Related log: clients/azcomputerguru.com/session-logs/2026-07-16-mike-mailbox-security-investigation-exchange-tier-breakglass.md
---
## Update: 2026-07-18 06:10 PT — Ralph's transfer verified + cutover confirmed COMPLETE (mike)
Continuation of the Ralph's Live Transfer started 7/17. Verified the transfer landed cleanly, established the DNS cutover had already happened automatically, confirmed zero mail loss, and posted a team bot-alert. Ralph's migration is DONE; only the websvr-side decommission remains (deferred as a rollback cushion).
### What was done
1. **Transfer completion verified.** `view_transfer` = "Account ralphstransfer: Success / Session Complete" (7/17 19:26). IX account live, not suspended, 542 GB mail, all 25 mailboxes present (applications, armando, billing, claims, dispatch, info, kathleen.wicks, kathryn, kevin.razo, lw, lynette, managers, nts, office, olivia, phones, pvo15803, rob.wicks, sales, survey, taylor.chaidez, tc, test, warehouse, whse).
2. **DB-count flag resolved — false alarm.** Earlier "6 DBs on source vs 2 on IX" was a mis-read. Both servers have exactly **2 DBs** (`ralphstr_newwp`, `ralphstr_s597`) and **2 DB users** (`ralphstr_admin`, `ralphstr_s597`) — identical. Read `/var/cpanel/databases/ralphstransfer.json` on both to confirm. No gap.
3. **ralphsmoving.com "missing addon" flag resolved — vestigial.** ralphsmoving.com is in websvr `/etc/userdomains` + `DNS1=` in the user file, but **NOT** in `parked_domains`/`addon_domains` (userdata/main both empty) → no vhost, no local named zone, no docroot, no mail (all 25 mailboxes are on ralphstransfer.com). It did not (correctly) transfer to IX. The **actual** ralphsmoving.com behavior is a **Cloudflare 301 redirect**`https://ralphstransfer.com/` (NS crystal/bradley.ns.cloudflare.com; A 172.67.216.45 / 104.21.45.152; `Server: cloudflare` on the 301). Because the CF redirect targets the *hostname* ralphstransfer.com (not an IP), it kept working through the cutover with zero changes.
4. **Cutover confirmed ALREADY DONE — rode along with the restore.** `ralphstransfer.com` A = **72.194.62.5 (IX)**, MX = ralphstransfer.com (itself → IX), NS ns1/ns2.azcomputerguru.com. Both IX and websvr zone files were rewritten at the **exact same instant `2026-07-17 19:24:46`** with IX's IP — the Live Transfer restore wrote the zone pointing at IX and the DNS cluster (sync mode) propagated it. IX zone serial 2026071804.
5. **Zero mail lost — no delta needed.** websvr received **0** inbound since 19:00 7/17; its **newest message is 16:26 7/17** (before the transfer's mail copy AND before the flip), so everything it ever received was captured. IX has been the live server continuously — 95 deliveries in the current log, newest mail file 04:43 7/18 (rob.wicks, kathleen.wicks/docusign, lynette, etc.). The pre-compaction "final delta sync" concern is MOOT — the flip rode along with the restore, so there was never a split-delivery window.
6. **Source already suspended by the Transfer Tool.** `/var/cpanel/suspended/ralphstransfer` created 19:24:46, reason "User transferred to another server"; `whmapi accountsummary` = `suspended: 1`. So at 19:24:46 three things happened together: DNS→IX, source suspended, transfer complete. No manual suspend needed.
7. **Decommission deferred per Mike** — leave the suspended websvr copy as a rollback cushion for a few days (~7/21), then `removeacct user=ralphstransfer keepdns=1`.
8. **Bot-alert posted** to #bot-alerts (message_id 1528012554029502634) so Rob/team have the migration status.
### Key facts / decisions (this update)
- The cPanel Live Transfer "Copy an Account" auto-suspends the source AND rewrites the DNS zone to the destination IP on restore — in a shared DNS cluster (sync mode) that IS the cutover. No separate manual DNS flip was required (contrary to the pre-compaction plan).
- Decommission MUST use `keepdns=1` — a plain `removeacct` on websvr would propagate a zone deletion cluster-wide and kill ralphstransfer.com on IX. keepdns=1 leaves IX's authoritative zone intact.
- ralphsmoving.com needs nothing: Cloudflare owns its DNS + the redirect, which follows the ralphstransfer.com hostname.
### Commands / outputs (this update)
- IX check: `ssh -i ~/.ssh/id_ed25519 ix "echo <b64> | base64 -d | bash"` — DB map, `/etc/userdomains`, userdata/main (parked/addon), `/var/named/ralphstransfer.com.db` mtime+serial, exim_mainlog greps.
- websvr check: `plink -hostkey SHA256:qcaW8BWq5UyM0l0g6DS9JfYbMZN/LTXLs3BIEZV8BE0 -pw <vault infrastructure/websvr-legacy-hosting> -batch root@162.248.93.81 "echo <b64> | base64 -d | bash"` — same fields + suspend flag.
- Live DNS/HTTP: `nslookup -type=A ralphstransfer.com 8.8.8.8` → 72.194.62.5; `curl -sI https://ralphsmoving.com` → 301 Location: https://ralphstransfer.com/ Server: cloudflare.
- Bot-alert: `bash .claude/scripts/post-bot-alert.sh "<msg>" bot` → message_id 1528012554029502634.
### Pending (updated)
- **Ralph's websvr decommission (~2026-07-21):** `whmapi1 removeacct user=ralphstransfer keepdns=1` on websvr → reclaims 542 GB (websvr ~78% → ~35%). Suspended copy is the rollback cushion until then.
- (Unchanged from 7/17) Furrier upr.com + desertrat moves parked to next month; amtransit 6-vs-5 mailbox billing reconcile (Winter); BB SharePoint #32187 awaiting Annise's list; mike@ break-glass password rotation; Reliant GBP creation; wiki compiles.