Ad Widget

**RohrbaGe** · 21-08-2017, 14:22

Kos,

the remote agent did not stopped this weekend, so I think you solved the issues.

I have setup now on the AS400 a subsystem according your documentation.
This is the better way, if you have this possibily. But in most companies that is not wanted, so running the remote agent is the only way.

Are you planning also for your agent a possibilty to start a script / CL/RPG
like on the standard zabbix agent?
The monitoring of spools/disks/cpu/services works already perfect,
but I would like to check now thinks like:

Is the remote printer started?
Did the backup run successfull?
How many records are in the application table?
This is very specific and can be done with CL scripts and giving back to Zabbix the returncode.

That would make your good solution more perfect and maybe you investigated it already.
Or maybe you can easy adapt somehow the zabbix_sender, so that we can
call in a CL the zabbix-sender and transport the values to the zabbix server.

Regards

Gerald

**Kos** · 31-08-2017, 10:26

Gerald, thank you for your update.

I'm glad that the remote mode of agent does, finally, work.

Of course, it could be great to have something like a startup scripts for this platforms. Unfortunatrely, I could not provide any certain schedules about that. The same is true for an additional functionality also

If you have some sripts (CL, QSH or something other), you can use the zabbix_sender utility from the standard Zabbix agent for an AIX platform (they are binary-compatible).

**sancho** · 16-09-2017, 00:16

Hi Kos,
Is there any way to monitor only the qsysopr messages that are waiting to be answered?

Thank you for your work

**Kos** · 06-03-2018, 15:08

sancho, I'm very sorry - I missed your question.
At the current time - no, it's impossible.

The version 0.7.4 was uploaded.

Bug fixes:
The "system.localtime" (or "system.localtime[utc]") returned incorrect value (unixtime that is in future), fixed. The "system.localtime[local]" metric worked correctly.

New features:
1) new as400.cpu.capacity metric returns a float value: amount (in number of physical processors) of current processing capacity of the partition.
2) new metrics as400.systemPool.state[pool,mode] and as400.systemPool.discovery for discovering and processing of some properties of system pools. Note: I'm not sure that number of Faults is returned correctly, the testing is needed. At the moment I just use the appropriate API calls without any post-processing.

**sancho** · 06-03-2018, 18:06

Thanks for the answer Kos,
Today we will install the new version of the agent.

As always, congratulations for your work.

**Kos** · 07-03-2018, 13:56

Thank you, sancho.

Unfortunately, by mistake I yesterday uploaded the wrong version (the same as was before, 0.7.3). I've re-uploaded the correct version (0.7.4) now. If somebody had time to download this version yesterday/today, please download it once more now. Sorry for inconvenience.

**sancho** · 07-03-2018, 15:57

Thanks to you

**hab** · 10-04-2018, 16:30

Hi Kos. I registered to thank you for your effort - very cool

I'm now testing this and it seems to be working. I'm now monitoring hostservers, ASP and CPU usage, subsystem status and jobs in MSGW. Do you think it is possible to return the jobname+user+jobnumber of the job(s) in status MSGW?
I'm currently testing on a partition running OS 7.3. Let me know if I can somehow be of help to do some testing etc.
PS. I sent you an email to the address in your readme before registering here, hope you got it :-).

**Kos** · 26-04-2018, 12:39

Version 0.7.5 was uploaded.
Small bug was found and fixed. The program writes some version information to the log files during startup. All previous versions wrote incorrect version of "jt400.jar" library: the version of library used during compilation instead of really used library version. Currently it obtains this information at run-time.

hab, thank you for e-mail. Sorry for some delay, I'll reply soon.

**RohrbaGe** · 10-05-2018, 14:34

Kos,
your older version is running since 1 year without any issues in our production environment very reliable.
It helps already a lot to identify issues before user notice them.

Still I have some ideas to improve the monitoring.
Do you have any experience with zabbix_sender running von the AS400?
Idea is to run some scripts by schedule or special processes and get the values with zabbix_sender to the zabbix server.
I copied already the AIX binary to the AS400.
Withs strqsh I started the shell , changed to the folder and tried to start zabbix_sender.

But the pocess ends immediateley, without result.

$
> zabbix_sender --v
qsh: 001-0078 Prozess beendet durch Signal 5.
$

Maybe you have some information, how to deal with the zabbix_sender on the AS400.
I would like to start zabbix_sender by CL scripts.

Gerald

**Kos** · 10-05-2018, 17:23

Hello Gerald,

yes, the zabbix_sender utility does work on AS/400 in our environment about 4 years already.
The most probable reason of your problems is compatibility between versions of AIX binary and AS/400 system.
I remember that there was a compatibility matrix; however, I havn't the exact link for it now :-(

Update: probably, links are like that and that.

**Kos** · 22-11-2018, 09:43

We had upgraded our Zabbix Server to version 4.0 recently, and once more bug has been revealed: wrong processing of requests using the new Zabbix server-agent protocol in the passive mode.
This bug has been fixed, version 0.7.6 is available.

**dknoxville** · 30-11-2018, 14:29

Hi, I managed to install the plugin to monitor the AS400 (https://share.zabbix.com/operating-s...s-400-platform) as a subsystem on the iseries but the job ends and I get this in the logs:
24:20181126:144417.345 agent #-1 stopped [ZabbixAgent config]
27:20181126:144417.370 Procstat.updateJobinfoList() ended
27:20181126:144417.371 Exception in Procstat.updateJobinfoList():
java.util.NoSuchElementException
at com.ibm.as400.access.JobEnumeration.nextElement(JobEnumeration.java:79)
at as400.perfstat.Procstat.updateJobinfoList(Procstat.java:216)
at as400.thread.CollectorThrea...(CollectorThread.java:30)
27:20181126:144417.372 agent #1 stopped [collector]
26:20181126:144417.847 agent #0 stopped [Cache Controller thread]
28:20181126:144417.848 Thread[listener #1] interrupted in run(): java.lang.InterruptedException
28:20181126:144417.849 agent #2 stopped[listener #1]
29:20181126:144417.850 Thread[listener #2] interrupted in run(): java.lang.InterruptedException
29:20181126:144417.851 agent #3 stopped[listener #2]
30:20181126:144417.852 Thread[listener #3] interrupted in run(): java.lang.InterruptedException
30:20181126:144417.853 agent #4 stopped[listener #3]
1:20181126:144417.854 PassiveCheck.init(): finished
1:20181126:144417.854 Zabbix Agent stopped. v0.7.6

Any ideas ?

Thanks

**Kos** · 30-11-2018, 16:18

Hi!

These messages seem a bit strange for me.
First of all, you have a DebugLevel more that 3, as your log-file has additional information (for example, about thread "ZabbixAgent config" start/stop).
It's good.
However, I don't see some other messages that here should be. For example, all other threads are started just after the "ZabbixAgent config" thread stopped; however, I see messages about "ZabbixAgent config" stopping as well as all other threads ("collector", "Cache Controller thread" and "listener") stopping, but I don't see any messages about these threads starting.
As well, I don't see the very first lines of log file (the current versions of Java, libraries, and system info should be printed).

As I understand, this job ends immediately after start, does not it?
Does it fail to start every time or sometimes starts successfully but then fails after some time?
Are you using the active mode ("ServerActive=" parameter in config file)? I don't see any mentionings about "active checks" threads in the log.
Could you attach, please, (or send me directly by e-mail mentioned in documentation) the fresh full log (with DebugLevel=4); if it contains some confidential information (like host name, IP-addresses) - you can replace them by "***".
--
Constantin Oshmyan

**Kos** · 30-11-2018, 18:34

dknoxville, thank you for your input. I'll check this situation. It is still very unusual for my understanding: I see that some threads, that should be started later, logged their messages earlier than event after which they had been started.
For comparing - part of normal start with similar config file in my test environment:

Code:

    10:20181130:174128.719 As400Metric.process() is OK for system.uname: 'IBM OS/400 BANKTEST V7R1M0, Oracle Corporation Java HotSpot(TM) 64-Bit Server VM (v1.8.0_181)'
    10:20181130:174128.719 end of ZabbixAgent.process()
    10:20181130:174128.719  Agent hostname: 'banktest', System info: IBM OS/400 BANKTEST V7R1M0, Oracle Corporation Java HotSpot(TM) 64-Bit Server VM (v1.8.0_181)
    10:20181130:174128.782 agent #-1 stopped [ZabbixAgent config]
    13:20181130:174128.782 agent #0 started [Cache Controller thread]
    14:20181130:174128.782 agent #1 started [collector]
     1:20181130:174128.797 PassiveCheck.init(): starting
    14:20181130:174128.797 Procstat.updateJobinfoList() started
    15:20181130:174128.797 agent #2 started[listener #1]
    16:20181130:174128.797 agent #3 started[listener #2]

First: "Agent hostname: '...', System info:..."
Second: "agent #-1 stopped [ZabbixAgent config]"
Then: "agent #0 started [Cache Controller thread]"
Later: "agent #1 started [collector]"
Finally: "PassiveCheck.init(): starting"
The key moment is that "Procstat.updateJobinfoList() started" is far later than "agent #-1 stopped [ZabbixAgent config]".
In your log the message "Procstat.updateJobinfoList() started" is far earlier (4 seconds!) than "agent #-1 stopped [ZabbixAgent config]".

Unfortunately, I'm unable to reproduce this case: I see that you have a newer versions of your system (OS, Java, JTOpen library) that I don't have.
However, I'll try to do all my best to resolve it, if possible.

Just as a troubleshooting step, could you try to use a remote mode (with network access to AS/400 from a different host), please? Is it successful or causes to the same problem?

Anybody else has this agent emulator working with the same configuration: on System i V7R3M0, with Java IBM Corporation IBM J9 VM (v1.8.0) and JTOpen 9.6?
Are you successful or do you have the same problem?

Ad Widget

AS/400 Monitoring solutions

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment

Comment