RichardLetts ✭✭✭✭✭

1. No -- personally we use the HTTP post from a solar winds Alert to our service now instance to create alerts (*). 2. Not sure there is a question there. If it breaks, we/you get to keep all the pieces. I did the bit that does the HTTP post, and one of our application integration engineers did the JavaScript backend…

what makes you think that is what NPM thinks? a) are you monitoring it using SNMP? => what is the sysObjectID => what is the Vendor => what is the MachineType b) on https://your-server/Orion/Admin/ViewsByDeviceType.aspx => what is the View you have associated with the Checkpoint Machine Type? c) on…

a) set up sqlserver replication between the two using the same interface use for failover replication [see Setting Up Database Mirroring] b) replicate the database files, configure the slave instance to use those database files, and figure out a way to only have one instance of SQLserver running... [I could figure out a…

look at creating a custom sql alert on a node... your 'where' clause will now look something like: trigger: where NodeID in (select nodeid from FROM [dbo].APM_SqlBbServerErrorLog_Detail] WHERE [Message] LIKE '%untrusted%' ) reset : where NodeID NOT in (select nodeid from FROM [dbo].APM_SqlBbServerErrorLog_Detail] WHERE…

Which graph? what is it measuring? You can put more than 5.5.Gbps into a FWSM, doesn't mean you're going to get more than 5.5Gbps out. The 5.5Gbps figure quoted is for 1500-byte frames -- are you running jumbo frames? (its throughput is limited by the number of PPS, not the total bandwidth) Richard

No... not exactly. ( at NPM>=12.x) you still have to trigger an alert, but you can set the reset action so the alert clears automatically: Make sure your trigger condition only triggers ONCE when you want it to, or you will be spamming someone's mailbox. You can also add email actions from the Trap and Syslog viewers on…

(3) uninstall all of the solarwinds software off the old primary server, make sure the ProgramData and Program Files directories are empty (4) install the additional polling engine package pointing at the new primary poller and database. If there is much of the old primary server software lying around the licening will…

you will need to be running discovery on the nodes, since interfaces only contains things you are monitoring... SELECT TOP 1000 di.*FROM discoveredinterfaces di LEFT OUTER JOIN interfaces i ON di.discoverednodeid = i.nodeid AND di.interfaceindex = i.interfaceindexWHERE i.interfaceindex IS NULL AND di.interfacetype = 6…

Yes FGA: Only ask questions with yes/no answers if you want "yes" or "no" as the answer. What type of device? does it have two power supplies? (if it only has one the you will be out of luck), does it have a UPS? or do you want an alert from a UPS?

I don't know about on cisco, but on Juniper you have to be able to create a node with the polling ip address and a different community string, and then poll the additional VRF using that. In my environment we only do this for the top couple of VRFs since you can only do this by adding an additional polling engine. Please…

You should just be able to re-install it. I believe the installers for the components are all located at : C:\ProgramData\Solarwinds\Installers I think if you run the main NPM installer it'd upgrade/replace your orioncore anyway... when I ran VM servers I'd always take a snapshot before letting support folks mess with a…

You are probably trying to set custom properties at the time of node creation -- this is nolonger supported. what does the create node call look like?

are you doing a 'getnext' or a 'get' in the UnDP ?

as a caution, all nodes have their sysUptime wrap after ~497 days (it's a 32-bit centisecond counter)

I've reproduced this in my environment. 1. install Growl for Windows on your Orion server (or other server); this will act as the rendezvous point for the client growl installs set this up to allow clients to subscribe to alerts (pick a password they will use) 2. install Growl for Windows on your clients, and configure…

ok, this looks like growl for windows: Growl for Windows running on the desktop how do you get the alert manager to talk to it?

here is how I do it... #!/usr/bin/perluse Getopt::Long;use SOAP::Transport::HTTP;use File::Spec;use Data::Dumper;use Memoize;use XML::Simple;use SW::Orion;use SW::InformationService;use Net::SNMP;use strict;my $hostname = 'xxxxxxxxxxxxx'; # fill in a hostnamemy $username = 'yyyyyyyyyyyyy';my $password = 'zzzzzzzzzzzzz';my…

NPM monitors the sysUptime which is the length of time the SNMp agent has been running -- download my poller and alert for monitoring the hrSystemUptime which is what uptime(1) returns. hrSystemUptime.UnDP hrSystemReboot.AlertDefinition and vote-up this idea: http://thwack.solarwinds.com/ideas/2698

Imagine you did not have NPM, how would you solve the problem? Now the question becomes how do you use NPM to automate the solution to the problem. The reason I phrase it like this is that every environment is different -- e.g. I have a shopping street with every other business offering free WiFi running along the edge of…

Yes it's common to have to manually select the switch ports; it's a pain. One way you might be able to improve things is to put something useful like 'uplink' in the port description on the switch, so you can then search for switch ports a that don't have that in the description and remove them from monitoring afterwards.…

"It's worse than that, he's dead Jim" I've tried to use network sonar discovery to get ports and L2 discovery added and have opened up (I think) four support tickets on things that are broken.

see the last post from my on this thread: SNMP Trap Alerting questions

Network Configuration Manager (NCM) is the module you want -- it grabs the configurations and also will maintain an inventory of component serial numbers (which I assume you would then be able to link to your financial asset database)

see my post:http://thwack.solarwinds.com/message/186426 this has a more fully worked perl example including adding the device pollers -- note my comment "From an API perspective I'd really like to invoke an 'addPollersToNode' to add the correct pollers to a node based on the sysObjectID rather than my less than perfect…

Yes. though I don't quite know why everyone is so keen on using CNAME's -- they cause additional lookups and don't [quite] mean what people think they do. why not simply have: solarwindsdb.example.com IN A new.server.IP.address we migrated from SQLserver 2008 to a 2016 cluster -- the DBA moved the database, and I updated…

groups have to be uniquely named -- it's the primary key on the table so if you have a group called Location 1 it can be a member of several groups... I name my groups: group gp Subgroup gp sgp subsubgroup kind of sucks , but's is wot I got.

UDT needs the sub-interfaces of access ports, not the physical ports -- the client MAC address and VLAN information is associated with the subinterface only. i.e. ge-0/0/0.0

How about some basic troubleshooting here. like are you receving thousands of traps from a rogue device? what version of the product?

The base interface is the actual bits-n-bytes transferred through the interface The .0 will eliminate many of the layer2 management traffic, e.g. vrrp, lldp, bpdu, incorrect vlanid,... etc you need the '.0' to work with UDT, and that is also the unit that shows up with the LLDP collection, so I'd say go for the '.0' if…

I make use of LLDP to identify new interfaces connecting monitored nodes. We use that to update our CMDB, and flag them for monitoring [manually]; when I go to NCM7.3 and one database for both NPM and NCM I can automate this. These are then added using an API call (as unpluggable so they won't generate an alert) when the…