Weird behavior after performing repair database script

[Berto]

The server is a VM with 10GB of memory, I've bumped it up to 16GB and retrying the repair script again.

[scottwilkerson]

Let us know if you get another Segfault

[Berto]

The script to a long time to run but after increasing the memory it completed successfully:

===============
REPAIR COMPLETE
===============
Stopping ndo2db: done.
Starting ndo2db: done.
Running configuration check...
Stopping nagios:. done.
Starting nagios: done.

=======================
nagios database repair succeeded
nagiosql database repair succeeded
nagiosxi database repair succeeded


So I ran these two request commands and here is the output for those:

echo "select max(logentry_id) from nagios_logentries;"| mysql -pnagiosxi nagios
max(logentry_id)
5432521


echo "describe nagios_logentries;"| mysql -pnagiosxi nagios
Field Type Null Key Default Extra
logentry_id int(11) NO PRI NULL auto_increment
instance_id int(11) NO MUL 0
logentry_time datetime NO MUL 0000-00-00 00:00:00
entry_time datetime NO 0000-00-00 00:00:00
entry_time_usec int(11) NO 0
logentry_type int(11) NO 0
logentry_data varchar(255) NO MUL
realtime_data smallint(6) NO 0
inferred_data_extracted smallint(6) NO 0

[scottwilkerson]

good, are you still seeing lines like this in /var/log/messages

Code: Select all

mysql_error: 'Duplicate entry '5431088' for key 'PRIMARY''

Code: Select all

tail -f /var/log/messages

[Berto]

These are the errors I'm seeing in regards to myslql_errors

Mar 15 10:58:05 lnsvr0370 ndo2db[]: [user.info] ndo2db: mysql_error: 'MySQL server has gone away'
Mar 15 11:00:48 lnsvr0370 ndo2db[]: [user.info] ndo2db: mysql_error: 'MySQL server has gone away'
Mar 15 11:01:01 lnsvr0370 ndo2db[]: [user.info] ndo2db: mysql_error: 'MySQL server has gone away'


I also see after the script from the admin web page (screenshot attached) that there's still lots of red from the XI component status. Plus now I'm getting HTTP 500 errors when I try to navigate around Nagios web interface.

[scottwilkerson]

PM Sent

[scottwilkerson]

Assisted in Remote, there was a backlog of 60,000,000 items in xi_events from over a month ago and so many in xi_meta that mysql couldn't count them in 3 minutes....

removed these and all was working as expected