RMA: Connection error

[Ross]

Hello everyone

I am new to this forum.

Been using Hostmon for a while now.

Some reason started getting Connection Error and this thing floods my mailbox like crazy.

some test are good and they all worked fine but for some reason it started this email, i get alert bad and then alert good.

can someone help me with this please.

[KS-Soft Europe]

Actually this is main purpose of HostMonitor - send alerts when something goes wrong.
Could you please give some details:
- What HostMonitor version do you use?
- Is it Started as Application or Service?
- RMA version?
- What kind of RMA? RMA for Windows? Active RMA for Windows? RMA for Linux?
- What is Status and Reply field of these tests when problem occurs ?
- Is there other test methods for the same server that work without triggering alerts ?

[KS-Soft]

1) When you have unreliable connection between HostMonitor and RMA, its good idea to check where exactly is the problem. E.g. you may start with Trace utility or Trace test method. Check timeout specified for this agent, etc. If timeout too short, increase it. If some router does not work properly, replace it...

2) If you cannot fix network problem, you may change HostMontor settings. E.g. you may tell HostMonitor do not start actions for "Unknown" test status.
- unmark "Treat Unknown as Bad" test option
- mark "Action depends on Bad one" option for "good" actions assigned to the test items.
Also you may use "Repeat test" action using "advanced mode" action and expression like ('%Status%'=='Unknown') and (%Recurrences%==1)
For more information please check the manual or visit our web site at
http://www.ks-soft.net/hostmon.eng/mframe.htm#actions.htm#advancedaction
http://www.ks-soft.net/hostmon.eng/mframe.htm#actions.htm#actRepeat

3) In any case when you have many tests that depends on some common application (e.g. RMA) or single network connection, its good idea to setup Master test.
E.g. setup 1 test to check connection to the agent (Ping localhost or Ping rma itself) and make other RMA tests dependant on this test item.
For more information please check the manual or visit our web site at
http://www.ks-soft.net/hostmon.eng/rma-win/index.htm#Settings - check "How to use" section near end of the page.
http://www.ks-soft.net/hostmon.eng/mframe.htm#tests.htm#Master

Regards
Alex

[Ross]

Hi Alex

thank you for the reply.

The situation is that we use Hostmon to get alerts from our client servers.
Everything was working great, we would get alert bad when things such as server restarts happen or anything we basically configured hosmon to do.
Alert Good would come on once server would come on.

We migrated our hostmon server to a datacenter and it was working no different for around 2 weeks.

suddenly started getting crazy amount of emails which tells us that servers are down, services are stoped so on.
none of that is correct since everything is running fine on client end.
we cannot trust Hostmon untill that is fixed.

I upgraded hostmon on 1 client to the latest version and on our server to the latest version.

last night still got around 300 emails, alert bad then alert good.
Connection between our data center and client is definitely reliable, people work remotely on server and never get drop outs.

Our test are setup like you are describing, we have Master Agent that pings and all test depend on that, so if its not working other test tell us Waiting for Master.

All agents are passive and all run on windows environment.

We have setups such as 5 bad or 2 good and it sends emails.

Biggest problem is, for example one of my client has like 30 tests, they show disk space, services running and certain other things.

When i look at the tests they are pretty much all green so everything is working fine.

after certain amount of time, some tests start to show "RMA:Connection error".
if i refresh the test that has RMA error on it, it goes back to being green.
Sometimes it tells me that Connection was forcibly closed by other side.

this is why we keep getting emails, all of our emails have RMA Connection error.

Here is an example

Reply : RMA: Connection error,
Status : Unknown ( 2 times in a row ),
Time of Event (from log) : %NTEventTime%

------------------------------------
HostMonitor (test changed status)

Test : XXXXX
Folder : >\XXXXXXX\Servers\
Status : Unknown ( 2 times in a row )

Date : 2012-08-29 9:32:33 AM
Time of Event (from log) : %NTEventTime%

Method: ping (timeout - 2000 ms)
Reply : RMA: Connection error


Last status: Unknown
Total tests: 370768
Alive ratio : 93.68 %
Dead ratio: 0.04 %


Regards

Ross

[KS-Soft]

[Ross]

Hi Alex

thank again for a quick reply.

RMA Connection Error is the problem we started to experience, in reality nothing changed.

Looking at all the tests i can see them working fine but around 50 times a day i get this email sent to us saying the RMA error.
When i look at the test it says the same and if i manually refresh it, it goes back to normal and then the same RMA error comes up a bit later.

The only time we would get RMA Connection Error is when ISP or Server is really down and we wouldn't get Alert Good until the agent is back online.

it never happened before and nothing at our clients have changed.
If anything our connection is 100Megabit up and down and thats a lot faster than we previously had when everything worked fine.

The 1 client i mentioned was done for testing purposes, i tried to isolate problem such as using different version of the package and recreated some tests but i still keep getting RMA ERROR.

this problem literally happens to every one of our clients.

i am guessing you are correct about connection issue that Server cannot contact RMA agent and it puts out this error.

but how can i fix it, there is no issues with our connection, its all reliable and if i go to RMA Manager i can easily GET INFO on any of the agents which tells me that they are Operational
Regards

Ross

[KS-Soft]

[KS-Soft]

PS
If you cannot fix network, then in addition to other solutions (see my previous mails), there are 2 more options:
- add Backup RMA
- use Active RMA instead of Passive RMA.

Quote from the manual
================
Several procedures help to monitor networks over unreliable connections:

- Active RMA may store test results when network connection unexpectedly brakes off, it will try to reconnect to HostMonitor and send test results upon connection;

- HostMonitor uses more flexible schedule for the tests that should be performed by Active RMA, e.g. if test has to be performed immediately but Active RMA is not connected while it was connected several minutes ago, HostMonitor may wait for another connection up to 4 min before assigning Unknown status to the tests. Note: if you select the test item and click Refresh button, HostMonitor will not wait for connection, it will set Unknown status right away;

- For each active agent you may setup Backup Active RMA. Using this feature, HostMonitor is able to balance load between agents and use the backup agent when the primary one cannot establish communication channel.
================

Regards
Alex

[Ross]

Hi Alex

We have W2k8 r2 on our end and either w2k or 2k8 on client side.
All patched with latest updates.

We use TMG and clients use windows plus hardware firewall, nothing fancy thou.

All ports are opend up as required.

We have around 3000 tests for various clients.

There is definetly no network issues due to us actually hosting a virtual server for client in our data center so there is no need for internet in this case to report back and we are still getting connection error using same virtual switch between us.

We been using host mon for 5 years or so and only now having this rma connection problem.

I triple checked all agents and their depends and cannot find anything wrong.

For example i got 100 tests for 1 client, most tests show correct result such as ping, disk space and other stuff.

Thoes same tests for no reason display connection error and than next time it refreshes it shows correct result.
This keeps happening for all of the clients.

I did try a few of the things you mentioned but it only stops notification, it does not fix false result for the test.

Any other ideas?

Thanks.

[KS-Soft]

[pato]

Hi, first post here.

I know that this is an old thread but this issue drive me crazy until I figure out where was the problem.

Nobody mention to change the f*king PORT for communication of RMA. I have this problem in a APP server, using port 1055 for default. I think that the 1055 port was in use by another app (AV ok, fw disable, etc). So, I change the port to 1056 and everything works perfect.



bye.

Pato.-

[KS-Soft]

When HostMonitor cannot open port for listening, it records error in the log.
Also, there is Auditing Tool (menu View->Auditing Tool), its good to use it from time to time even if everything looks fine.

Regards
Alex

[pato]