2010-10-08 11:43 AM
I have Operations Manager installed and monitoring volumes for capacity levels based on Full and Nearly Full thresholds. At night when we do our verifications for Exchange and SQL the FlexClones that are created are instantiated at 100% capacity which breaches both thresholds and sets off the alarms. Is there any way to exclude FlexClones from the volume level capacity alerting?
2010-10-11 05:16 PM
You could set the threshold interval on the FlexClone so that the volume-full threshold never gets a chance to be breached. If your FlexClone only exists for 24 hours or less, you could set the threshold interval to 24 hours:
> dfm volume set <object> volFullThresholdInterval=86400 # FlexClone must exceed threshold for 24 hours before threshold is considered breached and an alarm is generated.
This could be accomplished with a script that executes after the creation of the FlexClones.
volFullThresholdInterval The value (as time interval) controls the generation of ``volume full'' and ``volume nearly full'' events. The events are generated only if the event condition persists for specified interval. If the interval is set to ``0 seconds''(default) the event conditions are checked every monitoring cycle. If the interval is not set global value will be used.
It not perfect, but it should accomplish what you're looking for until this RFE is added to Operations Manager.
2010-10-11 05:42 PM
Thanks for the response. It did not dawn on me that there were two elements to an alarm of this nature: the capacity threshold and how long the threshold has to be exceeded for the alarm to occur. Thanks for clearing that up. What is not quite clear to me is where the script would be run from. From the SME host?
Second question: is there a way to automatically pull volumes, clones or otherwise, into a group based on wildcard of say their name? For example a clone created for verification via SME might have the name sdw_exc05_sg3_db_cl_exchsnap__hsfcorpexc05__recent_ss_0. We could then use a wildcard of sdw*cl* to pull them into a group where the thresholds could be set statically as you suggest and let the volumes or clones come in and out of the group dynamically. This would allow us to be as granular as a group without have to be as specific as clone or volume. And, no script...
BTW, I am not afraid of scripting but I would like to keep this in GUI. Trying to keep it as simple as possible at least for now. Thanks again!
2010-10-11 09:26 PM
That's a smart workaround to get rid of the nagging volume alerts.
But this should be set to individual volumes and not at global level as that will apply for all volumes.
2010-10-11 09:45 PM
I don't think we are suggesting setting global values with a script. Reid suggested doing it at the volume level which I agree is a descent workaround. I am suggesting that the product provide a way to include objects into a group by matching a regular expression. We can then set the desired alert settings on the group and have the volumes that become members inherit said settings. Static alert thresholds, dynamic group membership, no spam and no scripting.
Of course the original RFE might work also depending on how it get's implemented.
2010-10-11 09:50 PM
I just cautioned that, global level options should not be set.
BTW, there is another RFE for dynamic groups creation too.Add your case.
2010-10-12 05:28 AM
Ideally, this script would be run from the server that is creating the FlexClone as part of a post-clone process. You want to tell DFM about the volume, and set its threshold interval, before DFM has the chance to automatically discover it and generate an alarm. So the SME server could run the script, and it could communicate with the DFM server to issue the appropriate commands. You could manually add the FlexClone volume to the DFM database (dfm volume add ...), and then immediately set the threshold values and threshold interval on that volume. (dfm volume set ...). I'm not sure yet whether this is something that can be easily done via the Open Management API. Adai likely knows if or how this is possible.
Regarding the dynamic grouping. Until the RFE is available, you could use the same script to add this new FlexClone to an existing (or new) DFM group. Ideally, when you destroy the FlexClone, it will automatically be removed from any/all groups.
2010-10-21 03:35 PM
I had another thought about this request. DFM has an event for "Volume Clone Discovered". You could create an alarm based on this event. The only action of the alarm could be to run a script on the DFM server. A single script could be used to:
1. Determine the name of the new clone
2. Set the volume thresholds for the newly discovered clone.
3. Set the threshold interval for the newly discovered clone.
4. Use if/then/else logic and regular expressions to automatically add the clone to a DFM group.
This would automatically prevent DFM from alarming on the new clone AND automatically add it to a group without requiring a script on the SME server.
Just a thought.
2010-10-22 01:37 PM
Just one step ahead, instead of setting the thresholds, rather delete the flex vol/lun clone from dfm using dfm volume delete <flexvolclone> or dfm lun delete<lunclone> in the script.