Donald_Francis

I have always gone with syslog and generally screened them to just critical/warning/emergency messages to cut the chatter. I think traps can be somewhat problematic where syslog is tried and true.

There is a great article stickied to the forum on how to setup netflows on a 6500 which is great. Unlike a router it is a multistep process due to the distributed nature of the 6500. In essence you have to turn on flows for layer 2 and layer 3 traffic.

Performance metrics that DBA's care about like fragmentation, mem usage by the actual DB processes etc. I am not an authority on it either. Here is an example of what I mean. This is what we use. I do not think there is an issue posting this since Solarwinds does not have a competitive product yet.…

Awesome idea thanks for the tip!

That is the biggest reason I use it in that way There are others but that is the largest.

Looks like the issue is that when a file is in use Windows will misreport the size. So since this file is in use most of the time I never get a proper sizing. This is what support told me was the issue. So I had to abandon monitoring of this file.

How about a way to import large amounts of subnets. I have over 1200 subnets. Sure I can use another Solarwinds tool plus 25 steps to import just the ones I want. Which is a horrid solution by the way. But even then you have to manually assign eash subnet to a gruop 1 by 1. This is not feasible. I am waiting on this…

How about being able to EASILY and QUICKLY import a large amount of subnets from a seed router without having to click a billion times to import and assign to groups.

I would like to add this. For a small install it is perfectly fine. But if you have mutiple modules (especially netflow) and you have a large install you would find its not the best way. You would end up with SQL and SolarWinds fighting for resources and starving each other out.

No actually it does not work like that. A numerical value is created from the options used to balance against (depends on platform capability and configuration) and with that value it will fire everything down one pipe for that communication. And as you may be thinking if the load balancing info is being used over and over…

I think a physical server can make that much difference when you are talking about a server that will use as many resources as a busy SQL box will. I think you can run everything else on VM without an issue so long as you give it enough resources. Just as an example here is my layout. Server 1 - Primary Box and poller…

Without a doubt that would be useful as well. The drops would be the reactive part of that and the bandwidth would be the proactive portion for resource allocation and planning.

I agree totally with DirtySouth. A physical seperate SQL server for just SolarWinds is a HUGE help and probably the best help to your enviroment. The other stuff is dependent on the counts, you have 500 devices but how many elements? 500 devices could be ALOT of elements depending on what they are.

Yes I realize there are solutions out there that will work. My point was SolarWinds could increase their business if they had their own and give a better value to the customer having the box customized for Orion.

Yes I can get them from lots of places but thats not my point.

External shows as down.

AAAAAAAAAAAAHHHHHHHHHHHHHH So I can put this "catch all" rule below my alerting rules and anything that drops through them that I want to drop it will? So in this way I CAN alert on something on a certain level if it matches but if it does not I can drop it ?

I have noticed this on our VPN devices as well. We have VPN concentrators that have upwards of about 500 people at any one time on them we they ALWAYS have high numbers of discards. There are no ACLs running here since they do VPN only, no firewall function. So my thought was always it was packets that the device expected…

Web on a VM works great of course, a poller so long as you give it enough resources is also fine (but I choose not to do it except for 1) and SQL I would really never put on a VM.

Ok So I have done this procedure and it actually leaves out the part where you get your subnets into IPAM. And the way to do that us import them from a file you create with Subnet list tool. So while I am happy there is a way to accomplish this, the utter hidiousness of the procedure is astounding. Especially considering…

Thanks for that info !

No it would not especially since that is a manual process that I would have to put on thousands of ports. I am talking about the function being automated. And I think unplugged just stops the repoting of up/down mainly. Which I still want very badly for whole nodes.

The problem I was having was that Apps would go into an unknown status. My alerts were setup so that anything but up would trip an alert. I found out from support there is a known issue where that was going on. I am awaiting a hot fix now.

That's right Crane kicking, Karate Kid style! To be more specific with my complaint. The process is well OK, until you get to the point where you have to select devices to import. With no sorting or searching capability I am left to having to click on about 2000 cisco devices among probably about something like 15,000…

Ah gotcha. I would like to stress however this saved me several gigs of space.

The whole thought behind my idea is that is it canned and natively supported.

I just read this, I have the same problem. Could it just be made so we can manually pick a router as well? I know I also have an issue where there are multiple routers on site and UDT often picks the wrong one. IE it will select the voice gateway which is not routing anything instead of the L3 switch or the actual site…

OK awesome, this makes me very happy to be wrong. I am seriously digging UDT and I have multiple departments whom are interested and I was starting to wonder if we would even be able to deploy it because of what I mentioned. But again thanks for clearing that up, now I just have to find some money somewhere :P

I would first go to the switch and make sure it is seeing it the same way SW is. If they agree on the misses then it would be best to do something. The whole thing can snowball because when a pool is exhausted the next pool in line gets used. Also keep in mind that cranking up your buffers eats memory. On a newer switch…

Sure, I could also use a clunky free product or a second NPM install but I do not think I should have too.