ODBC advanced logging, halting for 10 minutes due to errors

[blowery]

We log all test results to a SQL Server table where we mine and report from.

However, we are I/O bottlenecked on our monitor server where HostMon runs from. Until we get a H/W upgrade, we have done some things to alleviate the problem. While finding and testing this issue which we know we have to fix, we noticed that HM would deliberately stop logging to an ODBC database if a threshold of ODBC errors was encountered. Thus, we missed recording test results. We do not want to record to a log file; we need the info available to other downstream systems such as our incident ticketing system, etc., so it needs to be in a database.

I'd like to have a feature added whereby HM will do the following:

1. Allow an admin to define a secondary ODBC logging database, similar to the secondary SMTP server allowed.

2. On an ODBC error wait a user-defined-amount of time, then retry the operation. This should be repeated a user-defined number of times before switching to the alternative ODBC conenction. That connection should be tried a user-defiend number of times, then finally fail if the secondary exceeds the limit. No complete or hardcoded delay or halt in ODBC logging should occur unless both primary and secondary fail past their thresholds.

3. Modify the syslog to capture the exact verbose ODBC error that is occurring, for diagnostic purposes.

[KS-Soft]

Allow an admin to define a secondary ODBC logging database, similar to the secondary SMTP server allowed.

Ok

On an ODBC error wait a user-defined-amount of time, then retry the operation. This should be repeated a user-defined number of times before switching to the alternative ODBC conenction.

Often HostMonitor must perform 30 or more tests per second. Such "retry" operations can make a real problem. Its better to switch to backup server right away.

Regards
Alex