Changing hpSIM Notification for 1 Server only

[Dave K_6]

We have a server that needs to reboot itself every night between 1:00AM and 1:15AM.

Sometimes hpSIM sees it go down and sometimes it doesn't. Depending on how quick the server reboots, so it seems. But, we get pages on this stating that the server is unreachable, and then reachable again. Does anyone know how I can change the hpSIM to NOT monitor this for that timeframe, or do I have to remove the server from our list of monitored servers because hpSIM does not allow for this feature?

Thanks for any info,
Dave

[David Claypool · Accepted solution]

Let me see if I understand the requirement:

Monitor all servers 24 hours a day except this 1 server which should not be monitored between 1AM and 1:15AM.

First, my curiousity -- why does this server need to reboot daily?

Back to the problem:

hpSIM supports time filters, but only for the complete system list. You would want to edit the Hardware Status Polling for Servers task and except that server from the list. You can then create a new polling task for that one server and use a time filter so it doesn't monitor during that time period.

[Dave K_6]

We have a web development team that uses an app that is very buggy, and to clear the memory issues and other problems, they reboot this server every night.

We actually have 3 or 4 web servers that do this, but this server is an HP DL380, and we would like to monitor its vitals. But if it can not been done, we will have to remove it from hpSIM, and hope that nothing goes down on it.

I knew the old IM couldn't, was wondering if this new had this feature implemented for single servers. I guess not.
Thanks,

[David Claypool]

So my suggestion won't work for you? It's really designed to do exactly what you're saying.

[Dave K_6]

Follow me on this one...

Setup the Hardware Status to monitor everything, or ANY.

Then I create another Hardware Status to exclude just that one server?

Upon doing this, will the second one I created overwrite the first one? In which, when the server goes down between 1 and 1:15, it will see the rule I setup for this specific server and disregard the first Hardware Status rule that includes everything.

I don't see an option to include every server but 1. So I am thinking, add every server, then make another rule EXCLUDING that one, hoping that it will take precedence over the first.

Does that make sense? ;)

Thanks,
Dave

[David Claypool]

Okay. There is a default task called "SNMP Status Polling for Servers" that is set to poll every 5 minutes. Some people may want it more frequently, others less so, and its scheduling can be changed.

The task is tied to the Server Status Polling List. If you look at this list by using the [Customize] button in the list box at the left, the criteria is

Search for systems

where [system type] [is] [server]
or [system type] [is] [cluster]
or [system type] [is] [management processor]

We're going to leave that list alone, so let's create a new one, call it "Most Servers Polling List" and make it look exactly like the above, and when you do, use the [<
Search for systems

where [system type] [is] [server]
or [system type] [is] [cluster]
or [system type] [is] [management processor]
and [system name] [is not] [NAME_OF_SERVER_THAT_REBOOTS]

Save that list to the System Functions category.

Now go to Logs --> View all scheduled tasks. Like we don't want to touch the default custom list, we're not going to play with the task here, just disable it. Choose the radio button for Server Status Polling Task and click [Edit]. You're not changing the list, so click [Schedule]. In the bottom just to the left of the [
Now we need to create 2 new tasks.

Options --> Status Polling --> Hardware Status Polling

Use the [Change Targets] button and select "All Systems in the list:" and set it to the new list we created earlier, the "Most Servers Polling List." Click [Apply Changes] and then [Schedule]. You want periodically, every 5 minutes.

Now for the second new task:

Options --> Status Polling --> Hardware Status Polling

Use the [Change Targets] button and select "Individual Systems in the list:" and select the server that is NAME_OF_SERVER_THAT_REBOOTS. Click [Apply Changes] and then [Schedule]. You want periodically, every 5 minutes, but this time additionally we want to check the "Use time filter box." Click [Manage Filters] and then [New] to create a new one. Name it something like "Reboot window" and now you have a grid with a bunch of green boxes. By clicking you can un-select the boxes for any hour time period Sun-Sat. I would recommend unchecking the Midnight and 1AM hours just so you don't have to worry about clock accuracy.

You've accomplished your task. If it doesn't work out for you, delete these new tasks and re-enable the default task; we've left it alone to make it easy to fall back.

[Rob Buxton]

For the flexibility you're after you might want to try executing a task rather than a straight e-mail.

I have a requirement where Server Outages are logged to an Automatic Outage System.
To do this I run a task, format the message and then pass it on to the System.
But, as part of that it reads a list of Servers to exclude which I can modify easily.

You could do similar, either exclude the Server from e-mailing or constrain it with time settings.
I use perl as my language of choice and it seems to work okay.

I would like to see Server Lists be available for Event Notifications. If it were there may be ways of using that as a means of accomplishing what you want.