(32)Broken pipe: Error writing to vlogger

Discussion in 'ISPConfig 3 Priority Support' started by francoisPE, May 26, 2022.

Page 2 of 2
  1. francoisPE

    francoisPE Active Member HowtoForge Supporter

    Now, I understood what I did.
    To manage above roundcube bug, and to restore roundcube db, I run 'mysql ... < mysql.sql'
    The problem was that I copy paste wrongly the command forgetting to filter roundcube only database.
    I pushed in mariadb all databases including dbispconfig !
    Luckly, I almost did nothing on that server between creating mysql.sql and run command 'mysql ... < mysql.sql'

    Now, I dbispconfig table syslog, server 3, I have
    upload_2022-6-4_15-6-19.png
    For server 3, which is the one where jobs are blocked, datalog_id is 0 for all lines...
    I had the idea to identify few line and change datalog_id to a value above servers.updated
    But, tstamp is difficult to understand...
    How can I translate that to a date ?
    May be the simpliest would be to clean that table and restart from that ?
    What do you think ?
     
    francoisPE, Jun 4, 2022
    #21
  2. francoisPE

    francoisPE Active Member HowtoForge Supporter

    Would you please advise...
    I'd like avoid increase the problem...
     
    francoisPE, Jun 6, 2022
    #22
  3. francoisPE

    francoisPE Active Member HowtoForge Supporter

    A question
    In 'syslog' table, if i remove row what happens ? Is the related job killed and that's it ? Should I do something else to kill it fully ?
    There are so many jobs in that table... i don't know what they are doing here... it seems to be duplicates.
     
    francoisPE, Jun 8, 2022
    #23
  4. till

    till Super Moderator Staff Member ISPConfig Developer

    Do not delete any records from sys_datalog manually, the system cleans records automatically that are not needed, and a few thousand up to ten thousand records are normal, so do not remove them. And that a record is there is not an indication that it's a pending record, so do not mess with the database or you corrupt your system. same for all other tables, the system takes care to remove unneeded records, do not remove any of them manually.
     
    till, Jun 8, 2022
    #24
  5. francoisPE

    francoisPE Active Member HowtoForge Supporter

    Thank you very much for your help.
    For a reason I ignore, there is no more job in the queue...

    I would like to check "Custom php.ini settings" for each website.
    Where can I find these settings ? I try 'grep -r 'a_part_of_my_settings' /etc/*'
    but I didn't find anything...
     
    francoisPE, Jun 8, 2022
    #25
  6. till

    till Super Moderator Staff Member ISPConfig Developer

    I hope you did not empty it by using mysql truncate command, as no jobs will be processed on other nodes then because the id counter gets reset by truncating the table. You would have alter the values in the updated column of the server accordingly then. And adding new nodes might cause issues in future if you deleted any records with server_id = 0 as the new nodes have to pull the existing server_id 0 records to become in sync with the whole setup, by deleting them, new nodes can't be synced and jobs might fail on them due to missing data that needs to be spread to all nodes.

    In the web_domain database table and the get inserted into the vhost file.
     
    till, Jun 9, 2022
    #26
  7. francoisPE

    francoisPE Active Member HowtoForge Supporter

    I DID NOT empty
    I did nothing except let it running for a while !
    I check in dbispconfig, and, in /etc/php/pool.d/... file and found that this is in line.
    So topic is resolved.
    Thank you for the robstness of your solution :):):)
     
    francoisPE, Jun 9, 2022
    #27
    till likes this.
Page 2 of 2

Share This Page