jzola_hyperic
Enthusiast
Enthusiast

no metrics data on new agent for few hours. Cache, performance...?

Hi!

I upgrade Hyperic server from 4.3 to 4.5.1.
Everythings seems ok, but yesterday i upgraded one of agent and agent not collecting metrics.
I succesfully added new autodiscovery items.

On agent i see these files are 0 byte:
data/measurement_spool
data/measurement_schedule

I reset collection intervals but nothing...
And few hours later measurement_schedule now 158720 bytes and on the server shown some metrics data.
Today all metrics work.

Why need to wait hours?


Maybe a cache problem?
I have Platforms (82) | Servers (383) | Services (3107).
I atteched my health page.
Somebody known which cache settings is wrong?

thanks,
Zoltan
0 Kudos
9 Replies
admin
Immortal
Immortal

Now that the agent seems to be collecting metrics, are they refreshing at a shorter interval?
0 Kudos
jzola_hyperic
Enthusiast
Enthusiast

Now working, but some other agent Availability is wrong, but other metrics are ok.
before Server 4.5.1 i havent seen that
0 Kudos
jzola_hyperic
Enthusiast
Enthusiast

I added two more agent.
Now more than 24h metrics still not collecting.

I think not ok something.

If i add a new service etc to existing agent still not collecting..
I have now 700 services that is down...


What do you think? can i go back Hyperic server 4.4? but i want use currently database.
0 Kudos
jzola_hyperic
Enthusiast
Enthusiast

My problem solved, partly...

The problem is some platform confused(File server mount, network etc.) services
Because of this bug(or by "design"):
http://communities.vmware.com/thread/355192?tstart=0

I enter hyperic sql and renamed confused IP s in eam_ip, eam_aiq_ip tables.
I delete orphaned services on platforms.

And after I restart agents, everything works ok, new agent new service metrics.

I known you dont fix this issue because only me report a problem.
0 Kudos
admin
Immortal
Immortal

jzola,

It could be that multiple plug ins had detected the same service, is that what you mean by "platform confused(File server mount, network etc.)"? That will sometimes happen. Otherwise it may be a bug.

--jeremy
0 Kudos
jzola_hyperic
Enthusiast
Enthusiast

No, my problem is that some agent has same internal IP address. like 192.168.0.1
0 Kudos
jzola_hyperic
Enthusiast
Enthusiast

Still we has a lots of problem.

I dont know what else we can do. Nobody has cache problems??? Our hyperic health page has lots of red numbers. We tried to set memory, cache metrics but not working.

Only solution start over?
93 platfrom need to register to new fresh database hyperic 4.6
1 day - 1 platform: four month migration? 😞 😞 😞
0 Kudos
admin
Immortal
Immortal

Hi,

Have you looked at http://support.hyperic.com/display/EVO/Scaling+and+Tuning+Hyperic+Performance which covers the topic of tuning caches?

Starting over, especially with 4.6, wouldn't be ideal as you'll see more issues with 4.6 and also the same issues as before since it's not likely a database issue.
0 Kudos
admin
Immortal
Immortal

Also the "red" on the cache page is a very crude calculation and doesn't have much real value so you can ignore it.
0 Kudos