Is it possible to repair broken datagroup where class match do not work as expected

Hi Guys,

I have a strange problem in one of our datagroups we had records which had CIDR at the end of address (added by SOAP):

"{'netmask': '255.255.255.255', 'address': '1.1.1.1/32'}" 

but should looks like

"{'netmask': '255.255.255.255', 'address': '1.1.1.1'}" 

I fixed this with python (bigsuds), and recreated all records with tmsh. Datagroup looks just perfect.

Unfortunately "class match" cannot find new records at all, and only find them if I recreate broken ones with CIDR at the end (?!?). Right now we have some standard record which are working and some broken records which also works, but this is not acceptable solution at all. To make it works I have doubled all broken records this way:

{'netmask': '255.255.255.255', 'address': '174.142.113.196'}
{'netmask': '255.255.255.255', 'address': '174.142.113.196/32'}

Actually this broken records are leading to "General database error" in Web Gui if I try to modify datagroup, OR to "IP address not found" or "This is not IP address" from tmsh.

So for me looks like datagroup is somehow broken, but I am out of ideas how to fix it. This is our production system and downtime is not acceptable also is working in clustering mode and I am not sure that if reboot LBs one by one will actually fix the problem.

Any advice is welcome, Thank you in advance.