Topic
  • 11 replies
  • Latest Post - ‏2012-04-24T14:10:49Z by SystemAdmin
SystemAdmin
SystemAdmin
13279 Posts

Pinned topic Director 6.3 fails to discover agent

‏2012-04-20T15:11:02Z |
I saw an old APAR that sounded similar to this, but I'm not sure I ever deleted this server from Director...

I know the Agent exists. I know it is running. I know the port is open. But when I run discovery by name, IP, etc it does not find the server. I've used several IPs, renamed the server, etc. Director is simply ignoring it. I have 58 identical blade servers with identical OSes that are NOT cloned, and they all worked fine - except for this one.

So how do I discover a server (or just add it manually) when I know the agent and server are there and fine but Director just fails to acknowledge it?
Updated on 2012-04-24T14:10:49Z at 2012-04-24T14:10:49Z by SystemAdmin
  • bassemir
    bassemir
    39 Posts

    Re: Director 6.3 fails to discover agent

    ‏2012-04-20T15:34:54Z  
    Without know the history behind this blade here is what I would do.

    On the blade look at the endpoint.properties file and see what ISD server is listed in that file. The ISD system should be listed multiple times in that file, at least that has been my experience. This file is found at:

    /opt/ibm/director/agent/runtime/agent/config

    Here is what the tail of mine looks like:

    AgentManagerQuery.Host=129.40.73.10
    agent.version=1.4.2.2
    status.lastContactAttemptTime=1334942550228
    ScheduleSynchronizer.LastNewToken=-1
    TrustedCertificateQuery.Host=129.40.73.10
    fips.compliant=false
    PatchService.Host=129.40.73.10
    slp.enable=true
    scheduler.lastexecution=22249047
    AgentManagerQuery.URI=/AgentMgr/AgentManagerQuery
    root:bugs:/opt/ibm/director/agent/runtime/agent/config
    -->

    The AgenManagerQuery.Host is my 6.3 ISD

    Maybe the agent is registered to another ISD??

    I might try to "unmanage" that agent and then try to discover it with your ISD (when it discovers it, the agent turns from unmanaged to managed.

    The configure.sh command can be used to do this.

    Just a thought....

    Rich
  • SystemAdmin
    SystemAdmin
    13279 Posts

    Re: Director 6.3 fails to discover agent

    ‏2012-04-20T16:18:20Z  
    Thanks for the reply - I did look at that as it was one of the 'shots in the dark' I found when googling. All of the .hosts are pointed at localhost on the client already. I've also tried reinstalling the agent several times (I didn't mention that in the original post)
  • bassemir
    bassemir
    39 Posts

    Re: Director 6.3 fails to discover agent

    ‏2012-04-20T19:55:07Z  
    Thanks for the reply - I did look at that as it was one of the 'shots in the dark' I found when googling. All of the .hosts are pointed at localhost on the client already. I've also tried reinstalling the agent several times (I didn't mention that in the original post)
    Does the endpoint.properties file have a keyword like unmanagedAgent=true?

    If it does, that tells me it has not seen any request from a ISD server which fits your scenario if your ISD can not discover it. I am sure you have pinged the two systems already so you know they can talk.

    From the client I have had success registering with an ISD server by running this command on the client.

    /opt/ibm/director/agent/runtime/agent/toolkit/bin/configure.sh -amhost <your ISD IP> -passwd <your agent manager password> -force

    That worked for me.

    Good luck.

    Rich
  • SystemAdmin
    SystemAdmin
    13279 Posts

    Re: Director 6.3 fails to discover agent

    ‏2012-04-23T16:47:15Z  
    • bassemir
    • ‏2012-04-20T19:55:07Z
    Does the endpoint.properties file have a keyword like unmanagedAgent=true?

    If it does, that tells me it has not seen any request from a ISD server which fits your scenario if your ISD can not discover it. I am sure you have pinged the two systems already so you know they can talk.

    From the client I have had success registering with an ISD server by running this command on the client.

    /opt/ibm/director/agent/runtime/agent/toolkit/bin/configure.sh -amhost <your ISD IP> -passwd <your agent manager password> -force

    That worked for me.

    Good luck.

    Rich
    Ok first the good news. I ran the command you suggested and it says it successfully configured and installed the agent etc.

    But it still doesn't show up in IDS and when I run discovery on it it still comes back with nothing!
  • SystemAdmin
    SystemAdmin
    13279 Posts

    Re: Director 6.3 fails to discover agent

    ‏2012-04-23T16:50:53Z  
    Ok first the good news. I ran the command you suggested and it says it successfully configured and installed the agent etc.

    But it still doesn't show up in IDS and when I run discovery on it it still comes back with nothing!
    Has this server ever had maintenance on it (i.e. planar board replaced)? Tho rare, I've seen issues with duplicate UUIDs. When ISD discovers them, it thinks they are all the same server. Check the UUID of this blade (look in the AMM) and then search the ISD Resource Explorer to see if that UUID already exists anywhere.

    ----
    Craig Elliott
    IBM Advanced Technical Skills
  • SystemAdmin
    SystemAdmin
    13279 Posts

    Re: Director 6.3 fails to discover agent

    ‏2012-04-23T17:32:55Z  
    Has this server ever had maintenance on it (i.e. planar board replaced)? Tho rare, I've seen issues with duplicate UUIDs. When ISD discovers them, it thinks they are all the same server. Check the UUID of this blade (look in the AMM) and then search the ISD Resource Explorer to see if that UUID already exists anywhere.

    ----
    Craig Elliott
    IBM Advanced Technical Skills
    I've got the UUID but I can't see how to add the UUID column to the Resource Explorer. I can add GUID but it's not the same thing. Best way to search UUIDs? Also different format. IDS does 8-4-4-4-12 and the hardware AMM is 4 4 4 4 4 4 4 4 so i'll have to tweak that before searching
  • bassemir
    bassemir
    39 Posts

    Re: Director 6.3 fails to discover agent

    ‏2012-04-23T18:49:13Z  
    Ok first the good news. I ran the command you suggested and it says it successfully configured and installed the agent etc.

    But it still doesn't show up in IDS and when I run discovery on it it still comes back with nothing!
    By any chance are there multiple System Directors in your environment?

    Can you post the contents of the endpoint.properties here? Is the management server, the ISD server listed in the endpoint.properties file multiple times?

    this is hard to debug via forum posts, all I can offer is what I would do next....

    Next I would look to see what agents are registered with the Agent manager (which is the same system that I run ISD on).

    /opt/ibm/director/lwi/runtime/agentmanager/toolkit/bin/RetrieveAgents.sh -toolkitPassword <your password>

    If you have lots of agents, grep for the IP address you are trying to discover.

    Are we dealing with IPV6?

    Rich
  • SystemAdmin
    SystemAdmin
    13279 Posts

    Re: Director 6.3 fails to discover agent

    ‏2012-04-23T19:48:02Z  
    • bassemir
    • ‏2012-04-23T18:49:13Z
    By any chance are there multiple System Directors in your environment?

    Can you post the contents of the endpoint.properties here? Is the management server, the ISD server listed in the endpoint.properties file multiple times?

    this is hard to debug via forum posts, all I can offer is what I would do next....

    Next I would look to see what agents are registered with the Agent manager (which is the same system that I run ISD on).

    /opt/ibm/director/lwi/runtime/agentmanager/toolkit/bin/RetrieveAgents.sh -toolkitPassword <your password>

    If you have lots of agents, grep for the IP address you are trying to discover.

    Are we dealing with IPV6?

    Rich
    IDS is running on only one server, Windows 2008 R2. All agents are Common Agents only, and running on Windows 2008 R2 as well.

    Hey - it shows up when I run that command! It reports 57 agents in there but I only show 56 in resource manager. I notice a lot of them the agent name does not match the HostName, but then a lot of them do - any reason? Should they match?

    Agent Name: localhost
    HostName: GOAWHRD2-NEW
    IP address: 125.254.154.32, 192.168.3.32
    Agent Port: 9510
    Install Directory: file:///C:/Program Files (x86)/IBM/Director/agent/runtime/agent
    ManagedElement ID: 2aa8f1fa591d3b18b23c2734d8704861
    Operating System ID: 917ce1a0f5d311e0ada200145ed03f26
  • bassemir
    bassemir
    39 Posts

    Re: Director 6.3 fails to discover agent

    ‏2012-04-23T20:28:07Z  
    IDS is running on only one server, Windows 2008 R2. All agents are Common Agents only, and running on Windows 2008 R2 as well.

    Hey - it shows up when I run that command! It reports 57 agents in there but I only show 56 in resource manager. I notice a lot of them the agent name does not match the HostName, but then a lot of them do - any reason? Should they match?

    Agent Name: localhost
    HostName: GOAWHRD2-NEW
    IP address: 125.254.154.32, 192.168.3.32
    Agent Port: 9510
    Install Directory: file:///C:/Program Files (x86)/IBM/Director/agent/runtime/agent
    ManagedElement ID: 2aa8f1fa591d3b18b23c2734d8704861
    Operating System ID: 917ce1a0f5d311e0ada200145ed03f26
    My Agent name and Host name don't match. I don't think that is a problem.

    Which IP address were you using to do your discovery? 125.254.154.32 or 192.168.3.32?

    Which network is your ISD server on?

    Rich
  • SystemAdmin
    SystemAdmin
    13279 Posts

    Re: Director 6.3 fails to discover agent

    ‏2012-04-23T21:04:48Z  
    • bassemir
    • ‏2012-04-23T20:28:07Z
    My Agent name and Host name don't match. I don't think that is a problem.

    Which IP address were you using to do your discovery? 125.254.154.32 or 192.168.3.32?

    Which network is your ISD server on?

    Rich
    Both servers are on both networks, and I've tried discovery on both.
  • SystemAdmin
    SystemAdmin
    13279 Posts

    Re: Director 6.3 fails to discover agent

    ‏2012-04-24T14:10:49Z  
    Since it appears in the command dump results, that must mean it is known in the database. I saw an APAR on an older version that said when a server has been deleted but not properly removed by the app from the database, something like this can happen. The APAR said it was fixed in a newer release, but you had to contact support to resolve it before that release. Contacting support is beyond painful and we're definitely not paying for support on this buggy product.