Member OP
Worst migration: cPanel to DirectAdmin, last month. Lost all DNS records. Every zone file gone. In my country, we have 47 domains for government clients. Restored from personal backups I keep in country, not from host.
Question: why do migration tools not verify DNS before declaring success? Is this common?
1. This is a known failure mode. CPanel's DNS cluster architecture stores zone data separately from account backups. DirectAdmin's migration script (importer version 1.57 through 1.63) does not call the cluster API unless the --dns-cluster flag is explicitly passed.
2. The flag was undocumented until January 2026. Source: DirectAdmin changelog, build 2026-01-14.
3. Six months. The bug report was filed 2025-09-03, confirmed by staff 2025-09-12, patch released 2025-09-21—but the patch only fixed the documentation, not the default behavior. Default behavior changed in 1.64, released 2026-03-08.
4. Recommendation: always run `named-checkzone` post-migration, and maintain independent DNS backups using `rndc dumpdb -zones` or AXFR to secondary nameservers.
I have performed 200+ migrations. Verify everything. Trust nothing.
For what it's worth, losing DNS is bad but losing email DNS is catastrophic.
No MX, no SPF, no DKIM, no DMARC. Your mail is either bouncing or, worse, delivering with no authentication. You're now a spoofing target. Blocklistings follow within 48 hours if anyone was actively mailing.
I have seen this exact scenario three times. The migration "success" email goes out, the sysadmin celebrates, and meanwhile Gmail starts soft-failing everything because the TXT records evaporated. By the time you notice, you're in remediation hell with Google Postmaster and Microsoft SNDS.
Always screenshot your DNS zone before any migration. Better: export it. Better still: have it version-controlled. I know this sounds paranoid. It is not paranoid enough.
Wah lau eh minh1987 you lucky got backup lor
My worst: cpanel to directadmin also, but the migration script eat my email forwarders lah. 300+ forwarders gone. Client is law firm. Almost die.
Can check the script log? No log. "migration completed successfully" leh. Where the forwarders go? Nobody know.
Already switch job after that. Cannot tahan the stress lor.
I worked on the directadmin migration script team I cannot say too much but same but different from what FlowSana said the --dns-cluster flag was not a bug it was a "feature" for "backward compatibility" but we all know same but different from truth internal ticket was open for six months before any engineer was assigned management said "low priority" because "users should read documentation" same but different from caring about users I left that team in march cannot sleep at night knowing how many zones were lost
