So we've recently moved away from this expensive software called Kaseya which was providing us monitoring for our servers.

I'm currently trying to configure and test OpManager which uses SNMP and WMI for obtaining hardware status. I'm trying to get my PC at home to act as the monitoring server and it's working OK with SNMP but WMI authentication always fails for me. Anyone know what I'm doing wrong or a better solution for monitoring CPU, memory, hard disk and WAN?

Argent

http://www.argent.com/products/ac.php

last edited by d[o_0]b at 11:00:50 30/Oct/09

Nagios, Cacti and my own random scripts to grep through logs looking for anything nasty.

Nagios, Cacti and my own random scripts to grep through logs looking for anything nasty.

this, rrdtool rox my sox

also; http://munin.projects.linpro.no/

example of munin running here; http://munin.ping.uio.no/

I prefer 19" LCDs ...


... I crack me up

http://www.argent.com/products/ac.php

Hey Argent Console looks pretty good. Is there a trial download of it? I'm confused by all their different products.

Argent's pretty pricey although if you are going to so CISCO switches etc all together it is nice to use. You can also do cool maps of your data centre and then when you click on the rack expands into the rack layout and then shows you were the server that had the alert is positioned.

Other good thing about Argent is probably it's escalation -eg- you can make it have a certain alert say less than 10% disk space then it sends and email to a group and flags it on the maps and in the system, the person in that group replies to the email within a certain time period it flags the problem as being looked at and by who. If not one actions the job within this time period it will escalate it to another person and so on and so on.

We are using IBM Director and a few other things to do most our monitoring at my current place of employment.

Wait for users to complain and then think about having a look.

Why did you move away from kaseya?, I knows it costs money, but it has heaps better support than something you can hack together yourself, and surely the time you spend costs money too.

OpManager is a great tool

Re WMI, what account level permissions are you using to access the server? You also need to do the authentication string as domain\userid

what shad said

Speaking of server monitoring; is there some simple way to monitor directories and files? I have the need to monitor a bunch of files in folders, sounds pretty simple but if there are any files in said folders for more than two hours an email fires off to alert somone.

If someone would point me in the right direction that would be swell.

P.S this would be a pretty easy thing to code with vbscript or even a batchfile?

cJay just VBScript it (although I would powershell is), something like umm

Dim oFs, oFolder, oFile

Dim Folder, FolderSpec, FileCollection



FolderSpec = "C:\Temp\Something"

Set oFs = CreateObject("Scripting.FileSystemObject")



If oFs.FolderExists(FolderSpec) Then

   Set oFolder = oFs.GetFolder(FolderSpec)

   Set FileCollection =  oFolder.SubFolders

   On Error Resume Next 'For lazy debugging

   For Each subfolders In FileCollection

      For Each oFile In subfolders.Files

         If oFile.DateLastModified < (Date() - 1) Then 'Mod this for time

            oFile.Delete(True)

         End If

      Next

   Next

End If

include some logging so you can see what has been deleted and then just setup a scheduled task to run it using cscript every 5mins or what ever.

last edited by Clubby at 12:03:50 30/Oct/09

cacti fan, and use nagios too.

don't bother with microsoft's SCOM, it creates more problems than it helps with.

Kaseya is just too expensive.

Re WMI, what account level permissions are you using to access the server? You also need to do the authentication string as domain\userid

I'm trying to login with an administrator username and password but it says authentication failed. Even putting the domain first tells me that authentication for WMI failed.

Wait for users to complain and then think about having a look

Your ideas are intriguing to me and I wish to subscribe to your newsletter

Thanks Clubby, I'm not good with VB but I'll give it a crack. I can also use emailer.exe to fire off something to a user perhaps?

Actually, I would also need to exclude some subfolders if I understand the code correctly, is that easy to do?

last edited by cJay at 12:24:43 30/Oct/09

Nagios, Cacti here

cJay: Just google "send email vbscript" you don't need an app or anything it is easy if you have an SMTP mail server somewhere you can send to.

Alize`: you need to elevate the privileges probably since the default is something that doesn't have access to certain areas. (Edit: sorry to clear this up a bit maybe check out impersonation level for WMI scripting)

Loving Powershell for doing all this sort of stuff at the moment.

last edited by Clubby at 12:38:27 30/Oct/09

GFI - http://www.gfi.com/pages/prodinfo.htm
PRTG, - http://www.paessler.com/prtg/?ref=PRTG5

Those are the two we use.

GFI is great as you can configure it for monitoring different areas on each server, say a ping test or a free space check, and have it configured to email an alert to you. Plus GFI also has a built in web page that can give you errors and alerts in realtime.

PRTG is used to monitor network, network traffic, LAN and WAN. Also can send alerts.

Jim, svn is broken

Nagios!, been using it for over 2 years now.

Also recently started using a nagios firefox plugin someone on here directed me to, works like a charm.

We use a Nagios setup to monitor over 100,000 devices. Does pretty well, but we also have a team to monitor and manage just the Nagios system.

Sometimes we also use the psychic admin which somehow knows when there is an issue even before monitoring software can pick it up.

Monitoring software is expensive.. we just hire some pleb uni grad to sit and stare at it. Much cheaper.

and just are reliable I'm sure :P .. specially at 2 in the morning.

It really depends on the complexity etc you are after .. if you have simple monitoring on only a few servers something like IPSentry would probably do for a cheap price.

Wait for users to complain and then think about having a look.

Screaming User Datagram Protocol never fails.
I'm sure SUDP is described in an RFC somewhere.

Sometimes we also use the psychic admin which somehow knows when there is an issue even before monitoring software can pick it up.

We had a guy like that once, nicknamed the Server Whisperer. He'd sometimes trail off in mid sentence as if he could smell dropped packets in the wind. Other blokes said he just enjoyed putting his vibrating pager in his pocket for long periods of time.

^^ we also had a guy like that, we called him the f***wit noob who wasn't allowed in the data center

nagios + cacti in the past 4 places I worked at, 2 of them introduced by me others already had it

We had a guy like that once, nicknamed the Server Whisperer. He'd sometimes trail off in mid sentence as if he could smell dropped packets in the wind. Other blokes said he just enjoyed putting his vibrating pager in his pocket for long periods of time.

huelelelele

We use Zabbix to monitor all of our servers and network gear. On the servers we not only track uptime, but also standard stats like disk/net/cpu/ram utilisation, and some crazy stuff like how many processes are running on each processor and how much ram those processes are using.

We also use Splunk to gather all of our syslog and other logs in a nice searchable format.

We use a Nagios setup to monitor over 100,000 devices. Does pretty well, but we also have a team to monitor and manage just the Nagios system.

do you have a manager to manage the team of managers of the nagios system that manages your management information systems?

yeah I was going to say, bugger having 100k systems with nagios
I used to do support for HP on their NAS platform (network automation system) and that was scalable way beyond 100k but s***, it was designed to from the ground up to do so

nagios & 100k host entries can't be fun

depends how you organise it, and how many nagios hosts you split the load between
I can't see anything wrong with it