nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Community
- Command Central
- MVP Program
- Monthly Mission
- Blogs
- Groups
- Events
- Media Vault
Products
- Observability
- Network Management
- Application Management
- IT Security
- IT Service Management
- System Management
- Database Management
Content Exchange
- SolarWinds Platform
- Server & Application Monitor
- Database Performance Analyzer
- Server Configuration Monitor
- Network Performance Monitor
- Network Configuration Manager
- SQL Sentry
- Web Help Desk
Free Tools & Trials
Store

Anyone running pollers across WAN to DB server?

BryanBecker

We've been running 2 of our 4 remote pollers across the WAN. Out DB server is in 1 data center, these 2 pollers are in a second data center about a 1000 miles away. We have P2P wan circuits between the 2 data centers and the ping time from the poller to the db server is ~35ms. It seems we are starting to see gaps again in the graphs and before I jump ship on this, and move the polling to the other 2 pollers in the same data center as the db, I wanted to see if anyone else was trying the same thing.

Thanks.

BB

Find more posts tagged with

Accepted answers

All comments

WINNT

Yes, we are running one remote poller across a WAN to the DB server.

Here are our specs:

(1) Orion V8.5.1 SP1 SLX running Web Site/poller/v8.5 Application monitor
(2) Orion V8.5.1 SP1 SLX remote pollers
MS SQL2000 x86 SE
40,000+ elements

All servers are x86 Windows 2003 SP2

BryanBecker

So how many elements do you have on that remote poller? What's the latency between it and the DB server? What are your settings for polls, statistic collection, etc?

Sorry for the questions....just seeing how different your set up is to mine. Thanks.

BB

Network_Guru

I'd be interested in how you have this working as well.

I've tuned my polls per seconds below the recommended values & my polling completion percentage has increased from 98.6% to 99.1% for the remote poller.
I now have it running across a lightly loaded OC12 WAN link (622mbps) to the DB with 36ms latency.
However I still only record SNMP data every 1-2 hours in the CPU, Mem & Utilization graphs.

Thanks for any assistance.

njoylif

apparently the "head geek" is going to do a webinar on this subject. see the top page thwack.com to register.

FYI

BryanBecker

No...I think that webinar is talking about how to use Orion to monitor your WAN links....we are talking about how to get Orion to work with pollers of WAN links.

I tweaked the time the poller issues pings for up/down; rediscovers and stuff but that didn't help. I still have alot of gapping (dots on graphs) on my remote poller. I have to make a decision soon on what to do but my WAN link is very good and I don't see how its any different from what you have.

BB

Network_Guru

Bryan, are you seeing gaps in just SNMP collected data, or ICMP as well?
My ICMP polling charts are fine (polling every 5 minutes).
This proves that it can be done without gaps in the data.

BryanBecker

Yeah....looks like pings are fine even at 5-min granularity. The snmp graphs at 5 minutes is horrible looking.

BTW...what's the specs of your DB server? I'm not sure if this is hurting any but Orion repoerts the physical memory on my server is 97% utilitize (3.8GB RAM). CPU % is about 30-70% utilized. I will say that I went from a avg 10% PCU utilization to 30% once I enabled the NetFlow add-on. My server specs are ProLiant DL385 G1 (2) AMD Opteron /2.40GHz (4GB RAM).

BB

FormerMember

FWIW, we don't recommend this kind of configuration. A few people get it to work, but the latency inevitably causes the kind of issues you are seeing.

We do appreciate the problem, and we are working on a solution that would allow to place an appropriately-sized Orion on either size of the WAN link with the ability to roll up the data from both Orions into a single console. We should have a beta that you can try out in Q1.

BryanBecker

FWIW, we don't recommend this kind of configuration. A few people get it to work, but the latency inevitably causes the kind of issues you are seeing.

We do appreciate the problem, and we are working on a solution that would allow to place an appropriately-sized Orion on either size of the WAN link with the ability to roll up the data from both Orions into a single console. We should have a beta that you can try out in Q1.

Denny...I think what your proposing would be great...I look forward to looking at it.

In the meantime I do have some questions/comments. I'm having a hard time understanding this as a latency issue. If its a latency issue with getting stuff to the DB then I'd expect the ICMP graphs to have gaps too...this is not the case. The 5 minute graphs are good and the raw data has no misses. It's just SNMP. I don't think it's a load issue on my poller either....my 2 remote pollers (over the WAN) have ~1000 & ~900 elements respectively...no where near an upper limit. From what I can see I'm missing about every 4th or 5th SNMP poll. I'm going to put a sniffer on the poller today to look at what its doing. From what I see I don't see this as a latency issue..no offense. 35 ms is very good on the WAN side and since my ping graphs are updating well I'd expect the SNMP ones to do so too.

Thoughts?

BB

BryanBecker

I'd be interested in how you have this working as well.

I've tuned my polls per seconds below the recommended values & my polling completion percentage has increased from 98.6% to 99.1% for the remote poller.
I now have it running across a lightly loaded OC12 WAN link (622mbps) to the DB with 36ms latency.
However I still only record SNMP data every 1-2 hours in the CPU, Mem & Utilization graphs.
Thanks for any assistance.

NG...you say you tuned BELOW the recommended value? Was that the one provided by the polls per sec program? I'm getting ~96% completion rate on my 2 remote pollers (ones across the WAN).

BB