- Community Home
- >
- Servers and Operating Systems
- >
- HPE ProLiant
- >
- Server Management - Systems Insight Manager
- >
- That dreaded discovery task
Categories
Company
Local Language
Forums
Discussions
Forums
- Data Protection and Retention
- Entry Storage Systems
- Legacy
- Midrange and Enterprise Storage
- Storage Networking
- HPE Nimble Storage
Discussions
Forums
Discussions
Discussions
Discussions
Forums
Discussions
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
- BladeSystem Infrastructure and Application Solutions
- Appliance Servers
- Alpha Servers
- BackOffice Products
- Internet Products
- HPE 9000 and HPE e3000 Servers
- Networking
- Netservers
- Secure OS Software for Linux
- Server Management (Insight Manager 7)
- Windows Server 2003
- Operating System - Tru64 Unix
- ProLiant Deployment and Provisioning
- Linux-Based Community / Regional
- Microsoft System Center Integration
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Community
Resources
Forums
Blogs
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
07-02-2006 08:57 PM
07-02-2006 08:57 PM
That dreaded discovery task
For the majority of our servers this works excellent, the servers exist in SIM and should they for some reason not exist, have been deleted etc, they will be rediscovered as soon as they send a trap (for example at reboot). Good for keeping all servers in SIM, but not so good on our "high-availability" solutions, namely clusters and load balancers.
We use 3 kinds of HA-solutions, MSCS, Veritas Cluster Server and NLB, and a similar problem happens to all of these.
We are not exactly sure whan it happens or what triggers it, if it's the discovery, a cluster failover or both. But the scenario goes like this
1. Starting without any of the servers in SIM, we can perform a discovery of them, this will add the servernames to SIM. In the case of MSCS the clusternames will also be discovered. So we end up with
Server1, IP-1, belongs to ClusterA, IP-A
Server2, IP-2, belongs to ClusterA, IP-A
Server3, IP-3, belongs to ClusterB, IP-B
Server4, IP-4, belongs to ClusterB, IP-B
(IP=IP-address, dedicated + clustered)
This is just fine and how we'd want it to look. However, after a few days instead what we may see could look like this:
Server1, IP-1&IP-2, belongs to ClusterA, IP-A
(Server2 not listed)
Server3 is now ClusterB, IP-B
Server4, IP-4, belongs to ClusterB, IP-B
Some small variations on this exist depending on whether we are talking about NLB, MSCS or Veritas, the last being the worst since it may even list the servername as any IP adress in the cluster, but not necessarily the same IP adress as the one given in the servername.
I have tried setting IP restrictions on the HP agents on all servers, I have tried editing the details in SIM and enabling the switch that says SIM should not be allowed to change this name during discovery, and I have tried excluding the IP adresses of the clusters and I have also tried both with and without cluster agents enabled on the monitored servers, so far with little success as SIM completely ignores me and continues to change the servernames seemingly at random.
All I want is for SIM to list actual physical server names, IP adresses should preferrably be correct but does not need to show all adresses just one working is fine by me, and whether the server is member of a cluster or not I could not care less for at the moment.
Is there any more settings I could try and check to keep the servers named correctly so all reportings doesn't go completely bananas ?
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
07-02-2006 09:30 PM
07-02-2006 09:30 PM
Re: That dreaded discovery task
CLUSTER ISSUES
Cluster identification employs many different agents, including WBEM
and SNMP, to determine if a system is a cluster or a cluster node.
Starting with the SmartStart 6.30 CD, one of the clustering SNMP
agents has not been fully installed even though SmartStart states
that all clustering agents are installed. If there are any cluster
identification issues, the following key must be added to the registry
on each cluster node:
HKEY_LOCAL_MACHINE,"SOFTWARE\Compaq\CompaqCommonClusterAgent\
CurrentVersion", Value: Pathname,%REG_EXPAND_SZ%,
"%SystemRoot%\System32\svrclu.dll"
Restart the SNMP service on each cluster node.
This problem should be corrected on the SmartStart 7.4 CD.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
07-02-2006 10:09 PM
07-02-2006 10:09 PM
Re: That dreaded discovery task
Still I'm looking into testing this setting on some of the clusters just to see whether it will give me anything and if it possibly still is an issue in the 7.51 agents/drivers.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
07-02-2006 10:12 PM
07-02-2006 10:12 PM
Re: That dreaded discovery task
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
07-06-2006 07:57 PM
07-06-2006 07:57 PM
Re: That dreaded discovery task
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=1040805
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
10-12-2006 09:22 AM
10-12-2006 09:22 AM
Re: That dreaded discovery task
Did you get to a solution for this problem, if you did, please share it. Im abaot to bang my head against the wall :(
/BJörn
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
10-12-2006 06:33 PM
10-12-2006 06:33 PM
Re: That dreaded discovery task
Nice to hear others have this issue too, maybe not for you, but at least for me since it confirms for me that this is a SIM issue and not something in our specific setup.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
10-16-2006 07:35 PM
10-16-2006 07:35 PM
Re: That dreaded discovery task
All I got out of that are 20 servers with trust errors (I only tested on 20 machines instead of all ;))
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
11-20-2006 04:12 AM
11-20-2006 04:12 AM
Re: That dreaded discovery task
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
11-20-2006 11:17 PM
11-20-2006 11:17 PM
Re: That dreaded discovery task
Anyway, since SNMP works I am of the opinion it would be more of a priority to be able to trust that your servers are actually being monitored than fixing an additional protocol that is not really needed since SNMP could be used.
I'm seeing this as a major liability issue that you can't trust that a server you added yesterday cannot be found today but resources/id now is assigned to a completely different physical node, i.e. another machine.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
06-13-2008 12:50 PM
06-13-2008 12:50 PM