CentOS7下Sendmail服务启动失败求助：模糊报错排查

Hey there, let's break down this Sendmail problem you're facing. From the truncated logs you shared, we can spot two key details:

• Sendmail's daemon actually successfully starts up (the line sendmail[3712]: starting daemon (8.14.7): SMTP+queueing@01:00:00 confirms this)
• Systemd is throwing a warning about not being able to read the PID file /run/sendmail.pid right after launch

This is almost always a timing mismatch between systemd's quick check and Sendmail's PID file creation, or a permission/configuration misalignment. Here's how to debug it step by step:

Step 1: Verify Sendmail is running and check PID file details #

First, confirm if Sendmail is actually active in the background:

ps aux | grep sendmail | grep -v grep

If you see running Sendmail processes, check if the PID file exists and what permissions it has:

ls -l /run/sendmail.pid

The file should be owned by root:root with at least 644 permissions (so systemd can read it). If it's missing or has overly restrictive permissions, that's a clear red flag.

Step 2: Fix systemd service timing or configuration #

Systemd might be checking for the PID file too quickly before Sendmail has a chance to write it. Let's adjust the service file:

1. Open the Sendmail systemd service file (usually located here):

vi /usr/lib/systemd/system/sendmail.service

2. Look for the TimeoutStartSec parameter. If it's missing or set to a low value (like 10), increase it to give Sendmail more time to create the PID file:

TimeoutStartSec=30

3. Also double-check that the PIDFile line correctly points to /run/sendmail.pid (it needs to match the path Sendmail uses).
4. Save the file and reload systemd to apply changes:

systemctl daemon-reload

5. Restart Sendmail and check the status:

systemctl restart sendmail.service
systemctl status sendmail.service

Step 3: Confirm Sendmail's PID file configuration #

Sendmail might be configured to write its PID file to a different location. Check its main config file:

grep -i "pidfile" /etc/mail/sendmail.cf

Look for a line like O PidFile=/run/sendmail.pid. If it points to another path (like /var/run/sendmail.pid), you'll need to:

• Either update Sendmail's config to use /run/sendmail.pid (then restart Sendmail)
• Or update the PIDFile line in the systemd service file to match Sendmail's path (then reload systemd and restart Sendmail)

Step 4: Get the full systemd log #

Your original log was truncated (ends with se...). To see the complete error message, run:

journalctl -u sendmail.service --since "2024-04-12 01:00"

This will show all Sendmail-related logs from around the time of the failure, which might reveal additional hidden issues we haven't spotted yet.

内容的提问来源于stack exchange，提问作者aannjose