Update from 5.2.7 to 5.2.8 issues

[jacek]

Hi,

noticed 2 issues after updating from 5.2.7 to 5.2.8 today (all SNMP related, the 2nd one not solved):

1. "ERROR: running table : Message size exceeded maxMsgSize".
Fixed this by following this thread -> https://support.nagios.com/forum/viewto ... 320#p22328
But why did this change? I didn't change anything in our Win servers?

2. SNMP services check fails at SQL services called 'SQL Server (MSSQLSERVER)', etc.
The XI check looks like that:

Code: Select all

check_xi_service_snmp_win_service! -C our_community --v2c -r -n 'SQL Server (MSSQLSERVER)'

The Core ARG check part looks like that:

Code: Select all

 -C our_community --v2c -r -n 'SQL Server (MSSQLSERVER)'

I've compared the check with our test server (5.2.7) and all checks are missing the "-r" option.
Removing the "-r" option does NOT do the trick.

Using the "-r" gives this error:

Code: Select all

"SQL Server (MSSQLSERVER)" not active : CRITICAL

Using it without "-r" gives this error:

Code: Select all

No services matching "SQL Server (MSSQLSERVER)" found : CRITICAL

It looks to me like the brackets are messing something up with regexp.
If I run the "Test Command" from core I see that the brackets are escaped with "/" on both the 5.2.7 and 5.2.8.

When I run the command from SSH only this one works giving the OK state (and picks the right service, looking at the verbose output):

Code: Select all

/usr/local/nagios/libexec/check_snmp_win.pl -H FQDN.com -C our_community --v2c -n 'SQL Server \(MSSQLSERVER\)' -v

[Box293]

1. "ERROR: running table : Message size exceeded maxMsgSize".
Fixed this by following this thread -> https://support.nagios.com/forum/viewto ... 320#p22328
But why did this change? I didn't change anything in our Win servers?

The plugin on the Nagios server may have been replaced during the upgrade. You may have modified it in the past.

2. SNMP services check fails at SQL services called 'SQL Server (MSSQLSERVER)', etc.
If I run the "Test Command" from core I see that the brackets are escaped with "/" on both the 5.2.7 and 5.2.8.
When I run the command from SSH only this one works giving the OK state (and picks the right service, looking at the verbose output):

Code: Select all

/usr/local/nagios/libexec/check_snmp_win.pl -H FQDN.com -C our_community --v2c -n 'SQL Server \(MSSQLSERVER\)' -v

Due to some issues with how PHP escapes characters the "Test Check Command" does not work in these situations and should be ignored.

So for all further testing of this service you need to:

Make the changes to the service
Save the Service
Apply Configuration
Go back to the home screen and find the Service
When viewing the Service Status Details page click the Schedule a forced immediate check link

Just to re-iterate, for all further testing for this service DO NOT use the "Test Check Command" button, follow the steps above.

Seeing as you have correctly tested it at the command line in an ssh session, proceed to update the service in CCM to reflect the command in SSH and then follow the testing steps above.

[jacek]

Hi,

understood, I managed to solve this problem, but this is a work around, which I don't like, since the Windows SNMP Wizard creates services that don't work.

In the XI service re-configuration command needs to use the "\" bracket escape, otherwise it won't work and this is the only configuration that seems to be working:

Code: Select all

check_xi_service_snmp_win_service! -C our_community --v2c -n 'SQL Server \(MSSQLSERVER\)'

The wizard (tested) generates such an output (which doesn't work):

Code: Select all

check_xi_service_snmp_win_service! -C our_community --v2c -n 'SQL Server (MSSQLSERVER)'

Is there a way to fix this? Revert some component to the 5.2.7 version?
Is there a known reason why this stopped working after the update??
Should I revert to the 5.2.7? Is this possible when I'm having the DB on a separate machine? (to get the data synced)

P.S. with the "-r" switch this doesn't work also, I assume that the switch is used to find the exact service, and not mismatch it with an similar named one?
The "-r" switch was not used in the 5.2.7 version.

[lmiltchev]

It seems like that these issues are caused by downgrading the "check_snmp_win.pl" plugin from ver. 1.2 to ver. 0.6. I already posted an internal bug report (TASK ID 8647).

[Box293]

It seems like that these issues are caused by downgrading the "check_snmp_win.pl" plugin from ver. 1.2 to ver. 0.6. I already posted an internal bug report (TASK ID 8647).

Ahhh ... then to go back to the latest version of the plugin do the following:

Download the wizard from here:
https://exchange.nagios.org/directory/A ... rd/details
Upload it via Admin > System Extensions > Manage Config Wizards

This will replace the existing plugins with the newer ones.

[jacek]

@Box293, this didn't help either, but I managed to solve the problem.

The wizard You sent is an old, 1.4.1 version which doesn't even scan processes and services, etc.
My check_snmp_win.pl was 1.2.
But the check_snmp_win.pl in the 1.4.1 package is in fact the 1.1 version (1.2 is the newest, but there is only one option added, no noticeable differences in the "-r" switch region).

This didn't do the trick.
I decided to go back to the newest wizard -> updated it using the "Check for updates" and "Install update" buttons on the XI interface, which brought my wizard back to the 1.4.7 version.

Now everything seems to be working, which is weird.
"-r" switch is working again, and there is no need to use the bracket escape backslashes, this command works now:

Code: Select all

check_xi_service_snmp_win_service! -C our_community --v2c -r -n 'SQL Server (MSSQLSERVER)'

So for me it looks like:

• some files have been replaced during the 5.2.7->5.2.8 update (with faulty/old ones),
• installing of the 1.4.1 version of wizard didn't replace them,
• along with reinstalling of the newest wizard (1.4.7) they have been replaced.

I see also that on my 5.2.7 machine the wizard version is 1.4.4, but I don't remember when I've last updated the wizard.

[mcapra]

Am I right in understanding you were able to solve this issue? Is it ok if we close this thread and mark as resolved?

[jacek]

I have resolved my problem, but I don't know why this happened.
I provided all info in my previous post, so You maybe would know why this happened.

[lmiltchev]

It is hard to say what happened after the fact. When I was testing the upgrade from 5.2.7 to 5.2.8, my "check_snmp_win.pl" plugin got downgraded from ver. 1.2 to ver. 0.6. Your plugin's version was 1.2 though... so I suppose we can rule out the plugin's version as the main cause of the issue.

In regards to the "Windows SNMP" wizard, mine was ver. 1.4.6, not 1.4.4 (in Nagios XI 5.2.7) but it is possible that I upgraded it sometime in the past. BTW, it is not guaranteed that components/wizards will get upgraded on Nagios XI version upgrade. You need to check for newer versions manually. For example, when I click on the "Check for Updates" button under the "Manage Configuration Wizards" page, I can see that version 1.4.7 of the "Windows SNMP" wizard is available for download (even in Nagios XI 5.2.7).

example01.PNG

[jacek]

Yes, I noticed that some components don't get updated along with the "global" update, so I'm used to check if there is something new.
P.S.
As we are discussing there is one thing that I'm trying to find - is there a place in which I can see the changelog of plugins/wizards and the main XI system?
Sometimes when I update components or the whole system I would like to know what I might expect, etc.