Donald_Francis

Actually this is still hosed in 10.1 as far as I can tell. External still shows under "Down Nodes".

Oh also, I am not sure why this is but the DB files get slammed while the log files sit idle in comparison.

I personally am not sure of the max number of pollers but I do know that you can squeeze about 10,000 elements onto a poller assuming its beefy. But all my boxes and my 2 pollers are blade servers. Here is my layout Poller 1- 4400 elements 2300 APM elements Poller 2 - 7000 elements SQL Box - dual quad core, 10gb mem, 8…

VM + SQL = bad Thats like virtualizing something virtual. There are components of SQL that are virtual and putting it on a VM from what I understand is a bad idea.

Just to be clear, my thought here with the power thing is that only major brand large UPS's would be supported. I personally do not see any reason to to waste resources going beyond that. That would cover the majority of people with a NOC of any decent size. 

My SQL box is Dual a dual quad-core blade with 10gb of RAM and 2 RAID 10 arrays on a SAN (DB files and temp DB) and 1 RAID 5 on a SAN for the logs. The pollers are all dual quad-core blades with 4gb of ram and generally RAID 1 local drives. My pollers seem to top out at about 9000 elemenets when they hit that level they…

I have actually tried to unmanage and remanage them and that seemed to partially worked. They came back but were missing a few interfaces. I do not see them even if I go into settings.

OK here is what I have so far. Listed here for anyone else that needs it or wants to confirm or deny anything. PowerNet-MIB1.3.6.1.4.1.318.1.1.1.2.2.1.0upsAdvBatteryCapacity.0100Percentage Battery CapacityPowerNet-MIB1.3.6.1.4.1.318.1.1.1.2.2.2.0upsAdvBatteryTemperature.022Temp in…

Same here, I run a lot of APM. I always set the pollers as to what the tuner recommends.

If you guys end up doing this please do not drop regular login support as we use the current account setups to hand out access t ogroups of people. We currently have a dashboard per business line and it is a very handy thing to create 1 account and just hand out credentials to a group to give them access.

Per the admin guide it is fully supported.

Fully supported.

You might also want to purchase an additional netflow license so you can put netflow on a box of its own so you do not swamp your Orion box.

Get a real business phone like a BlackBerry and you will have no problems. There is a reason Iphones are not prevalent on employer networks and this is one of them.

I agree. Their Teir 1 is horrid but their process is sound. However if you can get past teir 1 especially to backbone TAC you are golden.

Brandon I am glad there is SOME way to do it. But honestly do you really think that's the best implementation? Just because I CAN drive to New York (I am in Louisiana) to get a true New York style cheesecake does not mean it is the best solution for me to do so. For someone with 1000+ nodes or more and maybe have 25-50 or…

SolarWinds does have some really talented engineers who know the products well. But they are not usually the first people you talk too. You might try letting them know this is a chronic issue and ask for a next level engineer. Everytime I have gotten to the level 2 people I have always come away happy.

All great points and I agree. Many of the modules are that way, it seems to either to not have taken larger customers into consideration at all or been tailored to smaller customers. Now if that is the core audience I understand, but if not I think it is bad design and poor testing as in not enough. But to be honest they…

Good come back :) And you are very welcome!

And truth be told in practice a poller will use both NICs and send polls from both duplicating polling so you need to disable a NIC or have it in a team.

So it is not as easy as monitoring all the sub channels for up / down and calculating usage percentage based on the amount of time the channels were in use?\ I am not being sassy here just asking. I just kinda thought it was something untried at this point.

Just out of experience I would say having that large of an install with your SQL DB and main box coesixting would be almost a disaster. The busiest peice by far and the thing that always seems to be the bottleneck for performance is SQL. Just shooting off the hip, but moving that to another box I would think would do you…

We had the issue as well in 2.5, we are on 3 and we still get it just not quite as bad.

This may sound stupid but I have to ask it. Have you actually looked and seen that a particular process (hopefully just the browser) has eaten up all the RAM? If it is a true leak and not over utilization you may not but I was just curious.

Are these split tunneled? Would just monitoring the outside interface not do it for you?

I noticed the same thing the day CBQOS was added. As far as I have noticed if it has QOS on the interface and you have auto add turned on it will add it as a netflow source. In fact I had to turn of auto add for that reason. It kept adding all my chassis switches as sources due to the use of QOS.

I have had this issue more than once and each time I got with support and each time it was errors in the program here or there. My suggestion would be to get with support and let them dig into your system, after this was done for me it solved issue for awhile.

Well what it looks like it is great at is gathering information about gear. I was floored by the amount of data it pulled back. We are now buying the 3000 license and will pull in all of our network gear into it to run inventory reports. This saves us a ton of time from manual inventory.

After years of going back and forth on the subject I have come to the point to where I will make 1 large RAID 10 array and just use different partitions for OS and DATA or whatever else and let everything including the page file make use of the spindles and the speed.

I was just contacted by Cisco. They are not sure the bug has been fixed. I am actually testing the code release for them now.