mysqld_multi启动报错求助：硬盘故障执行fsck后遇启动问题

Hey Marco, sorry to hear you’ve been stuck troubleshooting this mysqld_multi issue for two days—dealing with a hard drive failure followed by fsck is already stressful enough. Let’s get this sorted out.

Why You’re Seeing This Warning #

The error about ledir=... means mysqld_safe can’t locate the actual mysqld binary executable anymore. This is almost certainly a side effect of your fsck operation: it might have altered file permissions, broken symbolic links, or even shifted the path to your MySQL binaries without you noticing.

Step-by-Step Fix #

• Find the real path to your mysqld binary
  Run either of these commands in your terminal to locate where mysqld is installed:

  which mysqld
  # Or if that doesn't work (e.g., if mysqld isn't in your PATH):
  find / -name mysqld -type f
  

  You’ll get an output like /usr/sbin/mysqld or /usr/local/mysql/bin/mysqld—make sure to note the directory part (e.g., /usr/sbin or /usr/local/mysql/bin, not the full path to the binary itself).

• Update your my.cnf configuration
  Open your my.cnf file in a text editor (e.g., nano /etc/my.cnf or vim /etc/mysql/my.cnf), then find each [mysqldN] group (where N is your instance number, like [mysqld1], [mysqld2]). Add the ledir parameter to each of these groups, pointing to the directory you found in the previous step:

  [mysqld1]
  ledir=/usr/sbin
  # ... your other existing config parameters (datadir, port, etc.) ...
  
  [mysqld2]
  ledir=/usr/sbin
  # ... your other existing config parameters ...
  

• Restart mysqld_multi and verify
  Stop and restart your MySQL instances to apply the change:

  mysqld_multi stop
  mysqld_multi --verbose start
  

  Check if the warning is gone, and confirm all instances are running as expected with mysqld_multi report.

Extra Troubleshooting Tips #

• Double-check permissions on the mysqld binary: Ensure the mysql user has execute access with ls -l /path/to/mysqld.
• If you compiled MySQL from source, confirm the installation directory wasn’t moved or corrupted during fsck.
• Verify other path-related parameters in my.cnf (like datadir) are still pointing to valid, accessible directories—fsck might have affected those too.

内容的提问来源于stack exchange，提问作者Marco