savell

That's fantastic thanks sedmo. Nice to come into work to find someone esle has done the work for you! Dave.

The other thing you can do is purchase the additional web server, and install it on your database server. Even with our database server and primary poller on the same local network, we found this helped our web response times greatly (4000+ nodes). Sav. 

We are currently running the Additional Web Server at 8.5.1 (everything was upgraded to this level). The only version issue I noticed was on the Hot-Bcakup server - which reported it was still at 8.1. This was sorted out when we forced a takeover (it was merely reporting the level the last time it was utilised). Didn't…

Cool, thanks Craig. That might get rid of our current open issue I suspect - will implement this suggestion tonight. Dave.

This has been a long time in the fixing. I regard this as a bug rather than a feature request. It's very close to unusable when you have 4000+ nodes. I don't see that we are being unreasonable in wanting this fixed sooner rather than later! Dave.

We just upgraded to 10.4.2, and the same full scan/sec trigger is now firing. Ours is doing >500 per second - something is not right (30,000 elements across 4 pollers). The database server seems to be handling it - but its a very large multi-core machine running SSD drives. The node details screen is very slow to display -…

Thanks Mark, We also have memory growth issue on dllhost (dllhost grows to >400meg in 4 days on our additional web server). I haven't logged an issue, but will apply this suggestion and let you know the result next week. Dave.

I use SQL Server Reporting Services to produce our graphical reports. I have posted an article on how to do this in the content sharing zone "> Granted it's not trivial to get started, but once you get the first report done any others are easy to produce. Dave. 

Glad to hear your upgrade went OK. I don't think you waited long enough for the Universal Device Poller to load - ours takes anywhere from 13-15 minutes with 15K elements - which is a very similar size to yourself (and 174meg of memory once loaded) - development is working on this for the next service pack. Node details do…

Still receiving the same error sorry Mark. According to the contest rules, I'm not eligible anyway - the rules state residents of the US only :-( Dave. 

Yes, RAID1+0. This will give you less database space, but better performance. We all work within budgets - Windows 64bit and RAID1+0 should at least make best use of the available hardware! Will you also be using a 64bit version of SQL Server? Dave.

I don't quite agree with the "you would not notice comment". One of our outstanding issues is the I/O performance of the C: (i.e. system drive) drive on our web servers - as a result of the cached maps that end up in this drive (and the continual folder and file scanning that occurs). It seems to be a significant cause for…

The 3 digit number is a vlan on the device. This type of polling is generated by the topology polling process (it will poll each vlan on the device). Topology polling was enabled on supported devices by default in this release (although you can list resources on the device and unselect it if you don't want to collect the…

Item #4 would be very useful for us. Our first database performance priority is Network alarms - Netflow data collection which impacts upon the database performance has always hindered our deployment. We would be more than happy to drop some of the low level detail to increase our Netflow collection (and improve…

We are running 10.0. From the questions I asked, it would appear that map load times didn't receive any attention in 10.1. From extensive dev discussions early this year, it would appear than most of the performance limitations concerning map load times is due to a third party component that is in use. Each map is scanned…

Interesting discussion this one. Whilst we are a service provider, we also perform build/run functions for customers who prefer to run private MPLS Network (the customer space in which I work). These are generally high (10Gig) bandwidth Core Networks with discreet modules connected via paired PE routers. Apart from the…

Funny this should come up again. We raised a similar case yesterday with the Events web page. We can see some events listed in the Events table, but not displayed on the web console. Not sure how long this has been occurring for (I don't think it's something we would have normally noticed). Using 9.5 SP5. Dave. Edit: From…

Jed, What type of device and code (IOS) level is it? Dave.

Geolem - I finally managed to get some time to look at the SQL for this. The InterfaceTraffic_Daily table contains some nicely rolled up data that can be used (retention is 365 days by default). As long as you don't want progress data for the current month, its quite a simple job to rollup the daily stats for each device.…

Support sorted me out quickly with a KB link http://support.solarwinds.com/kbase/ProblemDetail.cfm?ID=1121 All happy! Dave.

Your graph looks like mine charly.. 16 CPU cores running almost flat out - what can it be doing? I was interested to see you you also had a data gap just before the CPU increase. In my case I installed the NCM Integration module, and this gap was caused when the services were stopped/recreated and then started. Did you…

We just run a single RAID10 for both sys/apps and SQL (and then create a c: partition for sys/apps and d: for the sql data). We do this for a couple of reasons.. * Our sys/app boot partition is normally 20Gig, so we might as well use the rest of the drive for the data space. * The more read/write heads (i.e. drives) the…

Anyone else see a memory issue with the VoIP v2 service? Ours grows to over 500Meg within 8 hours. It won't terminate from the Orion control panel - I need to kill the process. I placed a support ticket today, but was interested if it was just us. Dave. 

I might be missing the issue here, but why not use 'equal' in your tigger condition rather than 'not equal'. For our ciscoEnvMonSupplyStateTable alarm, we have the following defined: This way, the alert can ignore Dormant or NotPresent power supply status returns. Dave.

Nice to find these old threads. I just had the same problem - a quick search and it was fixed in under 60 seconds (thanks Andy). Dave.

For the reasons you have outlined, we have over time moved away from both snmp traps and syslog. Instead, we tend to rely more on active probes (i.e. Orion provided for interfaces, up/down etc and custom pollers for array status, temp, power status, fan status etc). We do still have some alerts defined for both syslogs and…

Added here Chandru Dave.

Is there any update on this at the current time? (i.e even a temporay resolution). Dave.

 Thanks Mark - I had a few issues last night and the support team were excellent.

As a temporary measure to deal with the worst performance hit (i.e. the policycacheresults table), I created a non-unique, non-clustered index on the table using the CacheID column. CREATE NONCLUSTERED INDEX [tmp_idx_cacheid] ON [dbo].[PolicyCacheResults] ( [CacheID] ASC )WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE =…