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

Solarwinds stops monitoring

I have Solarwinds set up on a 2000 server. It seems every other day or so, the Network Performace Monitor stops graphing info from my nodes. If I close the program and restart it, this corrects it for a day or so.

Anyone else have this problem? More importantly, is there a fix?

Thank you,

Frank

Find more posts tagged with

Accepted answers

All comments

Network_Guru

Sounds like you are having database issues. Perhaps running out of DB file space?
When nightly maintenance runs it clears some room for another day or two of polling stats.
Check the debug.log file in the directory where you have installed Orion.

-=Cheers=-
NG

Impulse

If you are running the MSDE version of SQL that is included with Solarwinds, the max database size is 2gig. If your db size goes above that then you won't have an stats collected.

simon

I also experience a similar problem. I am running SQL server on the same box as Orion, but every now and then the polling service stops. All I have to do is restart it and it carries on again. It is very irritating to lose data whilst the poller is not working.

ftsang

I get the same issue. I have solarwinds set up on a win 2k3 server, an additional polling engine installed on another win 2k3 server and SQL server installed on a separate server. This issue where the polling engine freezes happens intermittantly, I'd say about once every 4-5 weeks. Restarting the polling engine does resolve ths issue.

Everytime I goto support they tell me they haven't heard of this issue before and to upgrade to the latest version of Orion. I'm on version 7.4 currently and have had this problem since runing version 7.1

rwbuckn

I had this problem for many month through several releases. I worked with Solarwinds support on several occasions only to be led in the complete opposite direction from what I hope finally fixed my problem. Support told me to ditch the "Polls per second" tuning tool and suggested some settings that were more than double what the tool said. They even remote connected to my environment for awhile, collected data and made adjustments. Thier final answer was to add a second polling engine. That didn't set well with my budget so I kept plugging away. Finally I went back to the polls per second figures that the tool recomended. I'm happy to say that my poller hasn't hung in probably more than 2 months straight (was every other day for the last year). I'm polling 4460 nodes, 7788 interfaces, and 12 volumes. Max polls is set at 30 and Max stat polls at 52. Polling completion is at 99.46%. I'm running Orion on a dual proc 3ghz Xeon and Ent SQL server on another dual Xeon.

pancamo

Do you get missed polls?

Seems that if our Polling completion is below 99.0% we start missing polls. I have been increasing the max polls per sec, but I'm now thinking that I should be lowering it...

I have also noticed that if the NetPerfMon Service thread is near 100% CPU, we miss polls.

I wish SW would monitor itself and tune the polling parameters dynamically!

dp

quote:Originally posted by rwbuckn

I had this problem for many month through several releases. I worked with Solarwinds support on several occasions only to be led in the complete opposite direction from what I hope finally fixed my problem. Support told me to ditch the "Polls per second" tuning tool and suggested some settings that were more than double what the tool said. They even remote connected to my environment for awhile, collected data and made adjustments. Thier final answer was to add a second polling engine. That didn't set well with my budget so I kept plugging away. Finally I went back to the polls per second figures that the tool recomended. I'm happy to say that my poller hasn't hung in probably more than 2 months straight (was every other day for the last year). I'm polling 4460 nodes, 7788 interfaces, and 12 volumes. Max polls is set at 30 and Max stat polls at 52. Polling completion is at 99.46%. I'm running Orion on a dual proc 3ghz Xeon and Ent SQL server on another dual Xeon.

rwbuckn

I'm not getting as near as many missed polls as I did before "lowering" my polls per second settings. I do on occassion have nodes on the higher end of the table (higer nodeID, last ones added to Orion database) go to warning state for a cycle or two. My worst problems with missed polls was with switches on the same subnet as the poller. This was resolved with a change that Solarwinds added to a later release. If I remember right it added the number of retries before a timeout occured. You have to manually add a key to the registery. Can't remember the details now but do a search and you should find it.
Yes my stats didn't look good until I lowered the polling way down from what Solarwinds support told me. Now they are more in line with what the polling tuner app suggests.

gannanf

Thanks everyone! all this info has been very helpfull. I really appreciate it.

Mithrilhall

Sounds very similar to a problem I just had.

Every few months it would look like Orion stopped polling. I would have to stop and restart the service or server to get it to work properly again. Last week I had the same problem so I rebooted the server. It would poll during the Baseline calculation and then it would stop working a little after it finished. After stopping and restarting the service/server a few more times Orion would look like it stopped polling during the Baseline calculation.

That was happening with 7.5 so I upgraded to 7.7 hoping it would fix the problem, which it didn't. After working with Solarwinds support and putting a packet sniffer on the server I notice that ICMP packets would stop but SNMP packets did not. I worked with support on this for over a week on and off. They ened up sending me the following file ICMP_v50.ocx. After renaming the old ocx file , registering the new one, and rebooting my computer my ICMPs have been working.

rwbuckn

quote:Originally posted by Mithrilhall

Sounds very similar to a problem I just had.

Every few months it would look like Orion stopped polling. I would have to stop and restart the service or server to get it to work properly again. Last week I had the same problem so I rebooted the server. It would poll during the Baseline calculation and then it would stop working a little after it finished. After stopping and restarting the service/server a few more times Orion would look like it stopped polling during the Baseline calculation.

That was happening with 7.5 so I upgraded to 7.7 hoping it would fix the problem, which it didn't. After working with Solarwinds support and putting a packet sniffer on the server I notice that ICMP packets would stop but SNMP packets did not. I worked with support on this for over a week on and off. They ened up sending me the following file ICMP_v50.ocx. After renaming the old ocx file , registering the new one, and rebooting my computer my ICMPs have been working.

That's interesting. I wonder what they changed and why there is nothing posted in the knowledge base about this new file? Did they close the problem out with you or are you still "testing" this fix? This has been my problem in the past put support always plays the "you need more polling engines" wild card!

Mithrilhall

When I first talked to support they told me that I was the only person that was having this problem. After about a week of working with them I came in on a Tuesday after Jury duty and noticed one of their networking or programming guys left me a message to call him.

When I talked to him he said he was going to send me a small file that should fix the problem. I asked him what made him think that this was going to fix the issue I was having and he told me that he tried it on another server the day before and it worked. I then asked him if I was the only one with this issue and he told me they had about 2 or 3 people with this problem.

Some of my specs below:

Server: Dual Processor Xeons (2.8Ghz)
4Gb RAM

We're also running SQL on the same server as Orion.

Network Node Elements: 283
Interaces Elements: 7109
Volume Elements: 57

If you're running version 7.7 of Orion open the "Orion System Manager" and go to File-->Polling Status.
I noticed once my "Packet Queues-->ICMP Outstanding" hit around the 900 mark the "Status Pollers-->ICMP Status Polling Index" would go to "0 out of 7449". When everything is running normal my "Status Pollers-->ICMP Polling Index" (after Baseline Calculation has run) is "7449 out of 7449" and my "Packet Queues-->ICMP Outstanding" is relatively low.