RichardLetts ✭✭✭✭✭

you missed out the INNER JOIN at the start of your custom SQL (actually re-reading the previous answer I think the meant you could delete that entire line, since the Where clause only mentions the table in the pre-defined part of the query)

The rate-limits are in aggregate, so while you might be limiting your netpath traffic that doesn't mean we are not being hammered by other traffic triggering ICMP messages. We have a lot of reachable public IP space, and when people scan it looking for hosts to attack they trigger the ICMP throttling on the routers, and…

This still does not work quite right at 11.5.2 :-( it's not a problem with the FQDN; it's that a) the configuration wizard doesn't build the web request to https b) at some point the wizard stops the redirection from port 80, but does not restart it If it did either one of these it would probably work, and I'm not sure why…

see look below for an early look at new content hovers!

Support case would be the best way forward because this is internal communications between components. well the Log messages indicate that the InformationServiceV3 isn't reachable on the localhost, but it's running. I'd check its logs to see if there is some error there I'd check the firewall logs in case someone's blocked…

the SWQL doesn't return any data -- it times out. And, now I look at that it's executing on the SQL server I can see why: it's decomposing the query into individual table select's; this is probably never going to work -- a simplistic in memory join of cirrus.interfaces with cirrus.interfaces on MAC address will generate 57…

erm, No Experimental.1057.1.0 is supposed to hold the IP address of the system that originated the trap. It's there so if the trap is forwarded (e.g. by a SNMP proxy agent, NAT gateway) then the source of the UDP packet containing the trap can be ignored and one matches on the content of the varbind. /RjL

Are you sure that is the reason? My code that adds nodes doesn't set Nodes.Description in any way (NPM 9.x through 11.0.1) and they all display correctly. Here is my literal code: my $AddedNode=$swis->Create('Orion.Nodes', {'Caption'=>$ip, 'IPAddress'=>$ip, 'IPAddressguid' => ip2guid($ip), 'DynamicIP'=>0, 'UnManaged'=>0,…

this is somewhat like the hardware EOL/EOS feature (but for software rather than hardware) e.g. here is the version list for Junos: Juniper Networks - JTAC Recommended Junos Software Versions - Knowledge Base Junos Dates & Milestones - EoL - Support - Juniper Networks

please share the case# here so I can add my support to it.

Complain to your equipment vendor; it's the default value reported (by cisco at least) for their tunnel interfaces: IF-MIB::ifDescr.15 = STRING: Tunnel1 IF-MIB::ifType.15 = INTEGER: tunnel(131) IF-MIB::ifMtu.15 = INTEGER: 17916 IF-MIB::ifSpeed.15 = Gauge32: 100000 IF-MIB::ifAdminStatus.15 = INTEGER: up(1)…

box 4 is not really large enough for the FSDB/NTA flow server: NTA Flow Storage Database Requirements Needs 16GB of RAM according to the docs.

Michal, Thank you for confirming that the upper limit on the polling rate is hard-coded into the polling engine. From the article: (3000 (Total Job Weight)/2600 (Maximum polling load for NPM jobs)) x 100 = 115% (polling rate). The denominator in that performance calculation is a fixed constant, The numerator is the sum of…

I don't have access to your data, so it's really hard to figure out what is happening here. What do the trap for a set and reset of an alert look like in the trap viewer? Because you have colored them differently they should have different background colors. it should look something like this: TIME OF TRAP IP ADDRESS…

there is the pollercheckertool that you run on the server that might allow you to bulk-correct this. (look in C:\Program Files (x86)\SolarWinds\Orion)

michalB wrote: There is TrunkMode column in UDT_Port table. Values: Unknown = 0, Trunking = 1, NonTrunking = 2 Ports with two or more VLANs are recognized as trunk ports. Really? That is not a good decision. Edge mode and trunk mode have very different meanings in spanning tree on some switches (e.g. edge-mode ports start…

I didn't see anything in this question/thread about a DHCP server until you mentioned it.

We should call this DanielleH 's or Michael Torok's law: "Any user or customer forum eventually grows large enough to support a job posting board"

on some routers/firewalls the sending of ICMP time-exceeded is rate-limited as a DoS protection mechanism; TTL Expiry Attack Identification and Mitigation - Cisco Rate Limiting > Cisco Router Firewall Security: DoS Protection check if you're triggering that behavoir in the box. /RjL

I am sorry, but you're not positing enough information for us to be able to help you. a) can you post something from wireshark on the server to show the packets are being received? b) can you bump the logging level up on your snmp trap receiver to show the error messages

I am not sure you are right. Have you tested this? [this is a building under construction, and I changed the polling addressed of the nodes so they would appear down while testing this] 1. change the building aggregation switch to make the parent appear down This is what I see for a node that still has connectivity, even…

it's not -- Matthew figured it out here and we got it working just. I'll let him share what he did Comments: 1. Solarwinds programmers need to learn how to spell; there is nothing so bad as reading logfiles with spelling mistakes in them 2. QA needs to check the spelling of log messages; they also need to check the NCM…

at best that is a hack, but that is what we do. we also change the first octet of the IP address to 127 to avoid warnings about duplicate nodes.

Have you configured the solarwinds alert service to interact with the desktop? What is in the windows event log for when the program runs?

please start a new thread for this question -- it's got nothing to do with the original question and it muddies the water for have a 9 year-oldthread come back to life.

if your server is running on vmware, which implies that you have your SW database on another server why not a) build a shiny new 2012 server b) shutdown the old solarwinds vm image... c) install products on the new server pointing at the old database d) move the licenses. I'm not a fan of in-place upgrades for windows;…

gripe: a) if NULLS are not supported in the database then do not allow them to be inserted; I mean you're literally allowing trouble to the stored for the future. b) if you don't want to do that then default them to a sensible value; e.g. if setting an unknown machinetype to '' is the right thing for the program inserting…

Also, in the Database manager you can do something like this: update nodes set in_UDT=1 where nodeid in (select nodeid from udt_job) to set it to true for nodes that have a UDT job running against them.

fx: shrug. you're right. Solarwinds seems to lack a metadata manager who blesses their database structure and names. Notice we have nodeid, interfaceid, but wireless access points are 'ID'. ObjectType has this longish string value, but in the pollers table we have 'N', and 'I'

I think that is a 10.7 or 11 feature.