How RiseCRM actually sends mail
RiseCRM handles outbound email in two distinct stages, and most delivery problems trace back to a misunderstanding of that split. The first stage is composition: when a user sends an invoice, assigns a task, or triggers a client-area notification, RiseCRM builds the message and either sends it immediately or writes it into an internal queue. The second stage is transport: the actual handoff to a mail server using the credentials you define under Settings → Email settings. When messages vanish silently, the fault is almost always in one of these two places rather than in the content itself.
On a Managed Cloud shared plan running LiteSpeed and CloudLinux, the default PHP mail() function is a poor choice for RiseCRM. Mail sent through the local sendmail binary carries the server hostname rather than your domain, frequently fails SPF alignment, and lands in spam or gets rejected outright by Gmail and Outlook. RiseCRM anticipates this by offering a native SMTP transport, which is the only configuration you should run in production. The application stores these settings in the database (the settings table) rather than in a flat config file, so you change them through the AdminCP, not through File Manager.
The queue adds a second layer. RiseCRM batches larger sends — notification digests, recurring invoice reminders, and bulk client emails — and drains that batch on a schedule. If the scheduler never runs, those messages sit in the database indefinitely while one-off transactional emails may still go out instantly. That mismatch produces the classic complaint: "password resets work but invoice reminders never arrive." Before touching SMTP credentials, decide which symptom you have. No mail at all points to transport. Some mail but not scheduled mail points to the cron and queue layer.
Configuring the SMTP transport correctly
Open the AdminCP and navigate to Settings → Email settings. Set the email protocol to SMTP. The values you enter here must match a real mailbox that exists on your hosting account. Create that mailbox first inside cPanel under Email Accounts (or DirectAdmin under E-Mail Accounts), for example crm@yourdomain.com, and set a strong password you can paste into RiseCRM.
Use the local server as the SMTP host so authentication and relaying stay inside your account. A reliable configuration on LiteSpeed shared hosting looks like this:
SMTP Host: mail.yourdomain.com
SMTP Port: 465
Security: SSL/TLS
SMTP User: crm@yourdomain.com
SMTP Password: (the mailbox password)
SMTP Charset: utf-8Port 465 with SSL is the most consistent on LiteSpeed-backed mail. If the connection stalls, try port 587 with STARTTLS instead. Avoid port 25, which is commonly filtered on shared infrastructure. The SMTP User must be the full email address including the domain — using only crm is the single most common cause of authentication failures, because the mail server treats the bare username as belonging to the default system domain.
Next, set the sender identity. In the same panel, define the From email and From name. The From address should match the mailbox domain you authenticated with. If you authenticate as crm@yourdomain.com but send From noreply@anotherdomain.com, the mismatch breaks SPF and DMARC alignment and invites rejection. Keep the authenticating mailbox and the visible sender on the same domain.
RiseCRM includes a test button on the email settings page. Send a test message to an external address you control, such as a Gmail inbox, and inspect the result. If the test fails, the application surfaces the SMTP error. A "could not authenticate" message means wrong credentials or username format. A "connection timed out" message means the host or port is wrong or blocked. A message that sends but never arrives points to DNS and reputation rather than transport, which is covered in the companion article on SPF and DKIM. If you need a broader reference on the same transport concepts across applications, the LaraClassified email and SMTP delivery guide walks through the same auth-and-queue pattern.
Driving the queue with a proper cron job
RiseCRM's scheduled features — recurring invoices, reminder emails, auto-generated notifications, and delayed sends — depend entirely on a cron job hitting the application's cron endpoint. Without it, queued mail never leaves. Set this up in cPanel under Advanced → Cron Jobs, or in DirectAdmin under Advanced Features → Cron Jobs.
RiseCRM exposes a web-accessible cron controller. The most portable method on shared hosting is to call it with the account's PHP CLI binary. First confirm which PHP version your RiseCRM install uses in MultiPHP Manager (cPanel) or PHP Selector, then build the command against that same version. A typical entry run every five minutes looks like this:
*/5 * * * * /usr/local/bin/php /home/username/public_html/index.php cron >/dev/null 2>&1Replace username and the path with your real document root. If RiseCRM lives in a subfolder such as crm/, adjust the path to /home/username/public_html/crm/index.php. Where CLI calls are unreliable, use a curl-based call to the public cron URL instead:
*/5 * * * * curl -s "https://yourdomain.com/index.php/cron" >/dev/null 2>&1RiseCRM protects the cron endpoint with a key you can view and regenerate inside Settings → Cron job settings. The dashboard shows the exact command for your installation along with the last run timestamp. Copy that command verbatim rather than guessing the path. After adding the job, wait one interval and refresh the cron settings page — the "last run" time should update. If it never changes, the path or PHP binary is wrong, and the account's error_log in the install directory will usually record the failure.
One practical limit on shared hosting is the PHP execution ceiling. Large queue batches can exceed the default max_execution_time. You can raise it for the account by adding a .user.ini file in the RiseCRM document root:
max_execution_time = 120
memory_limit = 256MThese values take effect for the PHP handler without any server-level change. Keep them modest; a 120-second window is enough for the cron process to drain a reasonable batch without tripping CloudLinux resource limits.
Bounces, logging, and keeping delivery healthy
Once transport and the queue both work, attention shifts to what happens after a message leaves. RiseCRM does not manage bounce inboxes automatically, so set the From and Reply-To addresses to a real, monitored mailbox. Bounce notifications and client replies then land somewhere a human reviews instead of disappearing into a non-existent noreply account that fills its quota and silently rejects mail.
When diagnosing a stubborn delivery problem, three logs matter. The RiseCRM application writes PHP-level errors to an error_log file in its own directory, viewable through File Manager — enable "Show Hidden Files" to see it. SMTP handshake failures and queue exceptions appear here. Second, cPanel's Track Delivery tool (under Email) shows the server's view of every outbound message: whether it was accepted, deferred, or rejected, and the remote server's response text. This is the fastest way to separate an application problem from a recipient-side rejection. Third, the RiseCRM email log inside the AdminCP records which messages the application believes it sent.
If Track Delivery shows messages accepted locally but rejected by the recipient, the problem is reputation and authentication rather than configuration inside RiseCRM — publish SPF and DKIM records and align your sender domain. If messages never reach Track Delivery at all, the application never handed them off, which returns you to the SMTP settings or a stalled cron. Avoid sending large client blasts from a brand-new domain in one burst; spread them over the cron intervals so your sending IP and domain build a steady reputation. Keep the authenticating mailbox well under its disk quota, because a full mailbox can cause the mail server to refuse new authenticated sessions, which RiseCRM reports only as a vague SMTP connection error.