I observe strange behaviour of the reported inventories from some 2.8 Windows Agents.
Some works fine, some not... There is a growing difference between the known machines in dashbord and the realy known machine in the machine list...
I had found out, that some agents get ein html 200 and started to collect inventory data, but when they send it there will result in an html 500 error...
Example:
172.17.1.134 - - [12/Nov/2020:16:32:28 +0100] "POST /ocsinventory HTTP/1.1" 200 235 "-" "OCS-NG_WINDOWS_AGENT_v2.7.0.1"
172.17.1.134 - - [12/Nov/2020:16:34:08 +0100] "POST /ocsinventory HTTP/1.1" 500 527 "-" "OCS-NG_WINDOWS_AGENT_v2.7.0.1"
10.242.2.15 - - [12/Nov/2020:16:34:52 +0100] "POST /ocsinventory HTTP/1.1" 200 235 "-" "OCS-NG_WINDOWS_AGENT_v2.8.0.0"
10.242.2.15 - - [12/Nov/2020:16:36:15 +0100] "POST /ocsinventory HTTP/1.1" 500 527 "-" "OCS-NG_WINDOWS_AGENT_v2.8.0.0"
Agent-Log: https://dpaste.org/hYWq
Everytime when in access_log a 500 appears, then this apears in error_log:
DBD::mysql::db do failed: Table 'ocs.softwares' doesn't exist at /usr/local/share/perl5/5.32/Apache/Ocsinventory/Server/Inventory/Cache.pm line 77.
DBD::mysql::db do failed: Table 'ocs.softwares' doesn't exist at /usr/local/share/perl5/5.32/Apache/Ocsinventory/Server/Inventory/Cache.pm line 79.
DBD::mysql::st execute failed: Table 'ocs.softwares' doesn't exist at /usr/local/share/perl5/5.32/Apache/Ocsinventory/Server/Inventory/Update.pm line 108.
Database works well, the password(s) in config files set well, most other clients/agents can send inventory w/o problem...
How can I debug that and find out where the Data is lost???
OCS Reports from Repo for Fedora 33:
