The "vhostmng-find" process uses a lot of CPU and never completes

Symptoms

• top command shows high CPU usage by the process vhostmng-find which never completes:

  # top

  

• Application scanning via Plesk > Domains > example.com > Applications > Scan never completes:

  

• Application scanning leads to appearence of the domain-app-scan task in the database which always has started status and never reaches done status:

  # plesk db "select id, type, status, progressStatus, progressValue, finishTime from longtasks where type='domain-app-scan' and status='started' and progressValue='0' \G;"
  *************************** 1. row ***************************
  id: 81
  type: domain-app-scan
  status: started
  progressStatus: queue
  progressValue: 0
  finishTime: 0000-00-00 00:00:00
  
  
  
  
  
  
  

Cause

It is the bug #PPPM-7833 which will be fixed in future updates.

Resolution

Apply the custom fix:

• For Debian 8:

  1. Create a backup of the psa database:

     # plesk db dump psa > psa_backup_pppm_7833.sql

  2. Find all hung domain-app-scan tasks in the psa database:

     # plesk db "select id, type, status, progressStatus, progressValue, finishTime from longtasks where type='domain-app-scan' and status='started' and progressValue='0' \G;"
     *************************** 1. row ***************************
     id: 81
     type: domain-app-scan
     status: started
     progressStatus: queue
     progressValue: 0
     finishTime: 0000-00-00 00:00:00
     
     
     
     
     
     
     

  3. Find the corresponding task parameters by using task id from the previous step:

     # plesk db "select * from longtaskparams where task_id=81;"
     +---------+-------------------+----------------------------------------+
     | task_id |       param       |                   val                  |
     +---------+-------------------+----------------------------------------+
     |    81   |     domainIds     |             a:1:{i:0;i:13;}            |
     |    81   |      referrer     |       s:18:"/smb/app/installed";       |
     |    81   | runTaskUnderLogin |              s:7:"cryptor";            |
     +---------+-------------------+----------------------------------------+
     
     
     
     
     
     
     

  4. Delete found hung domain-app-scan tasks and corresponding task parameters:

     # plesk db "delete from longtasks where id=81;"
     # plesk db "delete from longtaskparams where task_id=81;"
     

     Note: change id and task_id according to the result gotten on the step 2.

  5. Use the top utility to find the pid of all hung vhostmng-find processes and kill them:

     # top

     

     # kill 23736

     Note: change the pid of the process according to the information gotten by execution of the top command.

  6. Download the attached vhostmng-find:

     # wget https://support.plesk.com/hc/en-us/article_attachments/360000080394/vhostmng-find

  7. Make a backup of existing vhostmng-find:

     # mv /usr/local/psa/admin/sbin/vhostmng-find /usr/local/psa/admin/sbin/vhostmng-find.backup.pppm.7833

  8. Move the downloaded vhostmng-find to the required folder:

     # mv vhostmng-find /usr/local/psa/admin/sbin/vhostmng-find

  9. Change permissions, owner and group of vhostmng-find:

     # chmod 750 vhostmng-find
     # chown root:root vhostmng-find
     # ls -la /usr/local/psa/admin/sbin/vhostmng-find
     
     -rwxr-x--- 1 root root 89952 Jan 16 15:54 /usr/local/psa/admin/sbin/vhostmng-find