Hi Swathi, Many thanks for the detailed review. Please see my comments below, I will update the docs to reflect your comments/requests.
Regards Neil Swathi.Devulapalli at Sun.COM wrote: > Hi Neil, > > Please find my comments below: > > 1. OHAC is supported on S11. Any specific reason for mentioning > Solaris 10 nodes in the first paragraph? You're right, I'll update the doc to read "Nevada". > 2. Other agents documents will have a one/two-liner description of > what the application is. It would be better if we can state in a line > what Informix Dynamic server is. Agreed, I'll update the doc. > 3. Validate function: What exactly does it validate? Mention any > specific properties in brief? The validate function will validate the run time environment for IDS. I'll update the doc to reflect this. > 4. Will start function call validate before starting the IDS resource? Yes, I'll update the doc. > 5. Could not understand the clean-up logic in start function. > If a non-clean environment is discovered, clean up steps will only be > performed if, > ? They are recognized DBA steps for clean up as identified by IDS. > ? Any non-recognized DBA steps for clean up as identified by IDS will > only be performed if > the administrator requests such action when registering the OHAC agent > for IDS. > Here I am trying to explain that if a clean up is required and the clean up is a recognized DBA step as identified by IDS, then it will be done. If on the other hand we have to deviate to clean up then I'm simply suggesting that the administrator has to choose this when registering the IDS agent. For example, imagine there is a clean up task that should only be performed if requested to do so by IDS support. If that clean up was thought of as a non-recognized DBA step, i.e. not something that should be done ordinarily, then if it is required the administrator would need to choose that option when registering the IDS agent. Essentially I'm not making any non-recognized DBA steps the "default". > 6. Probe function: A short description of what the probe function > exactly contain in its probe logic. Agreed, I'll update the doc. > 7. Who would choose Option 1/2? Is it user? I think this functionality > is already present in the existing gds-based agents. I'm not sure if you are referring to the probe here, although I will assume so. In this regard I have a short probe and a detailed probe. A short probe is only run while the gds "start_command" is still running, whereas a detailed probe is run after "start_command" has completed. The reason to separate probing like this, is simply to avoid unnecessary probe error messages while the resource is still being started, i.e. while the start command is still being issued. Once the start command has been issued, it may still take a number of seconds/minutes for an application to fully start up and in this scenario a more detailed probe is run. I'm simply trying to avoid probing the application before it has been started. > 8. Registration of IDS resource: What needs to be edited in the > ids_config file and why? The ids_config file contains all the runtime values for the IDS resource. In this file the administrator will specify values for specific IDS variables, i.e. ONCONFIG, INFORMIXDIR etc. I'll update the doc to reflect this. > 9. Caller sequence Validate method: AFAIK, As of today, we do not have > a mechanism to determine if the update is on a system_property > or an agent specific property. So, agent's validate is called on all > the properties. You mentioned in Page#4 > of "Caller sequence" that separate validations will be performed on > different updates. > Agreed, I'll remove that statement for the doc. > Thanks, > Swathi > > Neil Garthwaite wrote: > >> Hi Everyone, >> Just a gentle reminder to say that the HA-Informix project page >> http://opensolaris.org/os/project/ha-informix has two documents that >> outline the design and caller sequence between the various components >> that this agent will use. >> The docs were posted on 7th August 2007, I'd appreciate any comments >> you have by the 18th October 2007. I stated in the project proposal >> that my aim is to "deliver this agent back to Solaris Cluster as a >> fully supported Solaris Cluster agent." >> >> Your help with reviewing these docs is much appreciated and will be >> helpful in getting this project into Solaris Cluster. >> >> Regards >> Neil >> -- >> >> This message posted from opensolaris.org >> >> _______________________________________________ >> ha-clusters-discuss mailing list >> ha-clusters-discuss at opensolaris.org >> http://mail.opensolaris.org/mailman/listinfo/ha-clusters-discuss >> >> >> > >
