Thanks for writing this up. Syslog tends to be one of the more forgotten pieces of functionality in PBS. Spending some time on it is a good thing.
Use case 1: You say logging via all daemons and then including objects like jobs and reservations. The objects are not daemons. What did you mean by referring to daemons and objects together like that?
Use case 2: PBSEVENT_DEBUG/PBSEVENT_DEBUG2 are not syslog log levels (they’re pbs log levels). There is a different log levels for syslog. Some examples are LOG_DEBUG or LOG_WARNING.
Use case 4: I don’t see how the user will be setting up syslog in a failover environment. I thinks this is just admin
Use case 6: This seems a little weird. Part of this is use case 5 (turning on both local and syslog together). The other part talks about running tracejob. How do these meet together? From reading it as it is, I think you’re saying you want to be able to run tracejob on either a local log or syslog. If you are truly just reverse engineering the syslog functionality and not adding anything new, then you currently can’t run tracejob on syslog logs. This might be a good enhancement to tracejob, but it can’t do it now.
Use case 7: I’m not sure we need this use case. This sounds like something completely in syslog’s corner. We don’t get into the business of how long logs are kept around