@@PRODUCT@@

From DirectAdmin, cPanel or Plesk to CoreCP

You know hosting. What you do not know yet is where CoreCP keeps the things you can find blindfolded in DirectAdmin, cPanel/WHM or Plesk. This article is that translation guide: first the concepts, then per everyday task the screen in CoreC

Written for: Administrator

You know hosting. What you do not know yet is where CoreCP keeps the things you can find blindfolded in DirectAdmin, cPanel/WHM or Plesk. This article is that translation guide: first the concepts, then per everyday task the screen in CoreCP, the route you can paste into the address bar, and the command in the terminal.

The biggest difference first: in your old panels one panel is one server. In CoreCP one panel runs all your servers. An account lives on one server, but you manage it from the same panel as all the others, and a command for one account is typed on the server that account lives on.

The concepts

CoreCPDirectAdmincPanel / WHMPlesk
administrator of the platformAdminroot / WHMAdministrator
a group with a reseller package, under GroupsResellerResellerReseller
account — one hosting account on one serverUserAccountSubscription
a customer with their people, under People——Customer
package, under PackagesPackagePackageService plan
websiteDomainmain domain, addon domaindomain
extra name of a websiteDomain pointerparked domain / aliasdomain alias
a website of its own (or an extra name)Subdomainsubdomainsubdomain
a server in your fleet, under ServersServerserverserver

A few consequences you will meet in the first week:

  • A reseller is not a kind of account in CoreCP but a group: the reseller and their customers are in it, and the reseller package caps the whole group. A customer package applies to one account.
  • A Plesk subscription becomes one account at import. You pick it by its main domain; the account is named after the subscription's system user.
  • There is no separate "subdomain" screen: a subdomain is a website, and a name that only points elsewhere is an extra name of an existing website.

The everyday tasks

In the route, :account stands for the account's name and :node for the server's name; fill them in and you are there. Where a panel command names an account, you write it as server/account. Commands are typed on the server the account lives on (corectl) or on the panel (corecp-panel).

TaskWhere in CoreCPRouteFrom the terminal
accountAccounts → Create account/accounts/newcorectl account add customer1
domainaccount → Websites → Add website/accounts/:account/websitescorectl domain add example.com --account customer1
mailaccount → Mail → Add → Mailbox/accounts/:account/mail/addressescorectl mailbox add info@example.com
DNSaccount → DNS → Add record/accounts/:account/dnscorectl dns record add example.com www A 192.0.2.10
databaseaccount → Databases → Create a database/accounts/:account/databasescorectl db add customer1 shop
SSLaccount → Security → SSL/TLS → Request certificate/accounts/:account/security/sslcorectl ssl issue example.com
backup/restoreaccount → Backups → Restore/accounts/:account/backups/restorecorectl backup restore account customer1 --dry-run
cronaccount → Developer → Scheduled tasks → New task/accounts/:account/developer/croncorectl cron add --account customer1 --schedule @daily --command "php cron.php"
PHP versionaccount → Websites → website → Set up PHP/accounts/:account/websitescorectl php set example.com 8.4
FTP/SSHaccount → Developer → Access/accounts/:account/developer/accesscorectl ftp add customer1 upload@example.com
reseller/packagePackages → Create package/packages/newcorecp-panel packages --assign web1.example.com/customer1 --to basic
migrationMigration → New migration/migrations/newcorectl import cpanel-server root@old-server --check

For SSH, add a public key in the same Access place with Add key; in the terminal that is corectl sshkey add --account customer1 --file key.pub. A reseller is created as a group under Groups (/organizations), and you attach a reseller package with corecp-panel packages --bind <group> --to <package>.

What you look for more often as an administrator

What you knowWhere in CoreCPRoute
Service Monitor (DirectAdmin), chkservd (WHM)the server → Configuration: the service list with its Watch column/nodes/:node/config
Server and security checksthe server → advisor/nodes/:node/advisor
CSF/ConfigServer Firewall, cPHulkthe server → firewall/nodes/:node/firewall
Mail Queue ManagerMail queue/mail/queue
Mail delivery reports, Track DeliveryMail delivery/mail/tracker
MultiPHP Manager, PHP Selectorthe server → PHP policy/nodes/:node/php-policy
Update Preferences, system updatesMaintenance/maintenance
Security Advisor, security noticesComponents & security/components
File Manageraccount → Files/accounts/:account/files
Statistics (AWStats)Statistics/statistics
Logsaccount → Developer → Logs/accounts/:account/developer/logs
Transfer Tool, Plesk MigratorMigration/migrations

What you did in an old panel with one click as root — restart a service, edit a configuration file — you do in CoreCP on the server itself with corectl, or in the server's web terminal. The panel asks the server; it runs nothing on your servers except through the agent.

Services that stop: the service watcher

DirectAdmin's Service Monitor restarts a service that is down, and WHM's chkservd does the same. CoreCP does this too: every server checks its own services every minute and starts one that stopped unexpectedly within a few minutes, without the panel or the network. It is on by default on every server. Where it differs:

  • A deliberate stop is respected. A service you stopped through CoreCP stays stopped until you start it again — the panel, corectl node service stop or the API, it is the same stop. The service list says who stopped it and when, and it raises no "service down" alert. A service that stops outside CoreCP (a bare systemctl stop, a crash) is restarted and reported.
  • A pause always ends. For maintenance you pause the watcher for at most 24 hours — the whole server or one service — and it resumes by itself. There is no switch that leaves a server unwatched and forgotten; turning it off for a whole server asks you to type the server's name.
  • A crash loop is not hidden. At most three attempts in fifteen minutes, with a growing wait between them. Then the watcher gives up on that service and the alert says so, with the attempt count and the service's last log lines, instead of restarting it for ever. It tries again after you start it or after an hour.
  • It stands aside while CoreCP itself changes the server, while apt or an automatic update runs, and while the server boots or shuts down.
  • SSH, the agent and the firewall are always watched; they cannot be excluded or paused.

How to exclude a service and pause the watcher is in The service watcher.

Moving over

How to move your customers off the old servers is in Migrating customers in waves: one wave per source server, with a trial migration, cutover and fallback. What comes along per source panel and what does not is in Migrating from cPanel and Migrating from Plesk.

From the terminal

On a server, two commands show what is there:

corectl account list
corectl domain list

And on the panel, which packages exist and who is on them:

corecp-panel packages --list

See also

  • Migrating customers in waves
  • Packages and limits
  • Creating a hosting account
  • Giving somebody access
  • The service watcher