Hi newtosis,

thanks for providing the details.

So we are talking about the URL monitor in general.
As you can see from the alerting email you attached, the monitor provides a status,
for example "Content match error",
but no HTTP error.

When SiteScope is able to access the page, and the content match failed, there is no HTTP error (as all worked but the content match),
and you receieve an error / status "content match error" adn the overall status -999.
Should the server not exist, it would be "unknown host name",
if it cannot connect (wrong port, web page doesn't respond etc.), it would be "timed out reading" and so on.

This is documented in the SiteScope Monitor Reference manual (I took the one from SiteScope 11.32),
Chapter 103: URL Monitor
..
Status

Each time the URL monitor runs, it returns a reading and a status and writes it in the monitoring log file. It also
writes in the log file the total time it takes to receive the designated document. This status value is also
displayed in the SiteScope Monitor tables and is included as part of alert messages sent by using e-mail.
The status reading shows the most recent result for the monitor. This status value is displayed in the URL
Group table within SiteScope. It is also recorded in the SiteScope log files, email alert messages, and can be
transmitted as a pager alert. The possible status values are:
- OK
- unknown host name
- unable to reach server
- unable to connect to server
- timed out reading
- content match error
- document moved
- unauthorized
- forbidden
- not found
- proxy authentication required
- server error
- not implemented
- server busy

As per the source code there simply is a translation
..
STATUS_MAPPING.put(Integer.toString(400), "bad request");
STATUS_MAPPING.put(Integer.toString(401), "unauthorized");
STATUS_MAPPING.put(Integer.toString(403), "forbidden");
STATUS_MAPPING.put(Integer.toString(404), "not found");
..

which means that SiteScope simply doesn't report the native HTTP error like 404,
but a text for it, in this very case "not found"
while the threshold has to be set on the HTTP status code:

As per the comments on "General Notes/Limitations"
..
When setting thresholds for the URL monitor, the status condition relates only to HTTP status codes
(such as 200, 302, and 404) in the page itself, whereas overall status relates to HTTP status codes in the
page and in components of the page such as images or frames (provided Retrieve images and Retrieve
frames are selected in the monitor settings).
..

Finally there is a table with all other error code reported by the monitor:

Error Messages
The following error codes might be encountered when working with URL monitors.
Error
Code Description and Troubleshooting
(-982) Permission denied.
(-983) SOAP Fault Error.
..
(-989) This message is returned when an error string is found in the HTML content.
Check the HTML content by using the View Source command from a browser.
..
(-999) This message is returned when the match string is not found in the HTML
content. Check the HTML content by using the View Source command from a
browser.
..

To make a long story short, you can use only use one of the following variables to get cloe to the error code.

-- status --
<status>
=> "-996"

-- state --
<state>
=> "timed out reading"

-- overall status --
<overall status>
"timed out reading"

Greetings
Siggi

Hi Adr,

check for an invididual logfile under
C:\Users\BAC_AD~1.UKM\AppData\Temp\HPOvInstaller\HPBsm_9.26
-> Package_msi_HPOprCommon_install.log

You can of course already see what's going on based on the screenshot:
the installer fails to verify whether or not HPOprCommon 9.26.151 (HP OMi Common) exists, or not.

You should have posted the full logfile, as I cannot see whether the same error shows up for other packages as well.

Anyway, to be quite honest, I tend to assume that your box is totally hosed after all the attempts to get rid of the BSM installation.

I would (really careful) recommend to reinstall the OS on the box and start from point zero.

While it might be possible to fix the firs tissue, you most likely face one with the next package, and so on.

Greetings
Siggi

Hi Adr,

looking at the complexity and history of the issue overall I think it's the best to open a support case, so that things can be worked on in a much better way then via the Forum back and forth, including limitations like the inability to add small files :-(

Greetings
Siggi

Hi alejandrosan,

can you please rephrase your question in english?

I (and most likely others here) simly don't understand your querstion.

Thanks
Siggi

Hi Adr,

as already mentioned, the normal (and usually working) way to uninstall BSM to go backwards from the highest patch and uninstall all one by one until you finally uninstalled BSM, for example
- uninstall BSM 9.25 IP2 -> this will get you to BSM 9.25
- uninstall BSM 9.25 -> this will get you to 9.20
- uninstall BSM 9.20 -> BSM is gone

The tool I mentioned is not a cleanup tool in case the uninstallation failed, it has been created mainly to speed up the steps to uninstall BSM prior to 9.26 to be able to install 9.26

In case you have leftovers from a BSM installation, the only way I know is to manually uninstall / delete all leftovers, packages, files, registry entries and so on.
This is why I asked you to open a support case, as it would allow support to check the box and uninstall / fix the issues on the system, while I still think the complete reinstall of the OS would be the fastest solution, if you don't have a lot of other software installed on the box.

Greetings
Siggi

Hi Kevin,

thanks for the update.

> First of all, unfortunattely DPS gets not up at all.
> BUT: After questioning database guys again about the parameter, they told they set:
> SQLNET.ALLOWED_LOGON_VERSION_SERVER=8 instead of
> SQLNET.ALLOWED_LOGON_VERSION=8 . They told us without _SERVER it's obsolete.
I checked that, and in fact with Oracle 12c SQLNET.ALLOWED_LOGON_VERSION has been deprecated.

For my understanding:
can BSM access the database with the new setting
SQLNET.ALLOWED_LOGON_VERSION_SERVER=8
or only with the old setting
SQLNET.ALLOWED_LOGON_VERSION=8

In other words, what's the setting, and do you see any occurences of
Caused by: com.mercury.topaz.cmdb.shared.base.CmdbException: [ErrorCode [-2147483648] undefined error code]
java.sql.SQLException: [mercury][Oracle JDBC Driver][Oracle]ORA-28040: No matching authentication protocol
in the logs when trying to connect to the database?

I will file a service request to update the documentation, if "SQLNET.ALLOWED_LOGON_VERSION" is not accepted anymore.

> At the moment with both configurations just opr-backend is pending "STARTING".
> Do you have an idea what the reason is for that?
There is almost no chance for analysis without the logfiles (<HPBSM>\log\opr-backend_boot.log or <HPBSM>\log\opr-backend\opr-backend*.log)
if you find an obvious error, you can post it here and I will have a look.

> By the way there was an update on another system on Monday and there was the error that mercury_as is in pending state.
> Is there a need to redo a Configuration via wizzard?
I guess I lost track. What BSM version? Where do you see "Pending" state?
The BSM Status page usually shows "Starting" or toggles between "Stopped", "Launching" and "Starting"
Running the Configuration Wizard is only required if something changed in the server configuration or when there was an issue with the certificates and alike, it's not "it will cure all" activity, but usually it doesn't hurt to do so.

Greetings
Siggi

Hi vidz,

I think this question has been asked already, and answered as well:
there is an active enhancement request for this
  QCCR1I118185 API to show status of all monitors
  (https://softwaresupport.hpe.com/km/KM02316738)

As of today the only options available are the one mentioned here, another API call for just returning the status of one or more monitors might be implemented in a later release.

Greetings
Siggi

Hi yangnigon,

the best option is the option which fits your needs the most ;-)

I personally always run BPM as configured by default, so that BPM runs as service under the Local System Account.
I only switch to specific user if the protcol or transaction requires it, same for running BPM as a user process vs. a service.
But this is not in a production environment, but a test environment.

The recommendation makes perfect sense in a production environment
> We recommend that you run scripts for all protocols (with the exception of web protocols) with the BPM
> instance/whole BPM set to run as a specific user.
and is what a lot of customers do, others run multiple instances of BPM, one as system user, one as specific user, and so on.
In the end it mainly depends on your requirements.

Greetings
Siggi

Hi Kevin,

this issue is getting too complex to be handled in the forum, can you please open a support case?

Thanks
Siggi

Hi Adr,

I was faced with a quite different issue in the past, where the BSM 9.26 installation was fine, but ovcert and ovcs didn't work, requiring the (re-)installation of the two components

HPOvSecCS-11.14.014-Win5.2_64-release.msi
HPSharedComp-11.14.065-Win5.2_64-release.msi

(see my post /t5/Application-Performance/BSM-Support-Tip-BSM-after-the-upgrade-to-9-26-issues-with/td-p/6898406)

So with this I can imagine that the uninstallation of BSM 9.26 has the related issue:
all is being uninstalled except of these two components.
You can uninstall the two remaining components by unpacking the BSM 9.26 installation fileset
for Windows this is BAC_00890.zip, there is a folder packages which contains (besides many others)

HPOvSecCS-11.14.014-Win5.2_64-release.msi
HPSharedComp-11.14.065-Win5.2_64-release.msi

The packages can be installed by a simple double-click on the .msi file.
Depending on the installation status of the package you can uninstall it by selecting "Remove".

Maybe this does the trick already?

Why does it happen that the two packages are not installed, or not installed correctly?
I have no idea, as it happens to some installations, but works perfectly fine on others,
I actually experienced both scenarios when installing BSM 9.26.

Greetings
Siggi

Hi Praveen,

if this is new installation and not an update from an earler 9.2x version, I would say that it is a bug.
If it's an updated installation, simply try to delete and recreate the alert, perhaps something broke during the upgrade.
From the stack trace it seems that we fail when trying to add the information "what do we group on" to the alert message.
BTW, what did you select in area
 Data Grouping
   Group data by specific criteria
      Business Transaction Flow
      Transaction
      Location
      Script
      Mobile Device 

I did a quick test on my 9.26 IP2 sxystem, with using Group By Location, the email is sent as expected.
Same when grouping by BTF or Transaction.
Do you use any additional alert variable in your alert definition?

Greetings
Siggi

Hi JameSPN,

I found some known issues in this area, like
QCCR1I118291 Ldap configuration wizard leads to Content Encoding Error

TECHNICAL PROBLEM DESCRIPTION:
LDAP configuration did not accept null settings

TECHNICAL SOLUTION DESCRIPTION:
Now LDAP settings use empty string instead of null value

Did you change anything in the configuration and afterwards the GUI fails?

Do you go to BSM -> Admin -> Platform Administration -> Users and Permissions and then Authentication Management?

The GUi is supposed to come up and show various settings.

Reproduce the issue and have a look at jboss-as-all.log to see what error is causing this.
If you manage to find a field name it fails on, it might be possible to update that very field via JMX,
and then the GUI should come up again.

Greetings
Siggi

Hi Adr,

you didn't tell us yet what you wanna do with the MIB and the SNMP traps.

The "thing" you send the SNMP trap has to receieve the SNMP trap.
In the documentation it is assumed that this receiver is a piece of software which offers a SNMP management console,
that's where you need to configure what it should do in case of an incomping SNMP trap from BSM.
Typical products are NNMi or NetCool and many others.

BSM comes with a pre-defined MIB file (CIAlerts.mib) for CI Status alerts.

You are talking about EBAlerts.mib, which is a similar file, but used with EUM alerts,
see
BSM Application Administration Guide, Chapter 79: End User Management Alerts Administration

> I have ready EBAlerts.mib file, can i use this?
BSM comes with a predefined one, as explained above.
As long as you follow the varbinds as listed in the section
Alerts MIB Varbinds
I think that you can use yours as well.

BSM uses the OID as listed in the varbinds, for example
  1.3.6.1.4.1.5233.4.2 for the name of the alert
no matter what you do to the EBAlerts.mib file, it will always use
  1.3.6.1.4.1.5233.4.2 for the name of the alert
the file itself is only useful for consumers of the trap, like NNMi.
You need to import / compile / load it it into the SNMP Management Console to enable it to translate the OID to the real names behind it.

Greetings
Siggi

Hi  yangnigon,

==> According to the web site, KB3167679 can apply to Windows Server 2012 R2 Datacenter as well if you look at the detail on the site.
Okay, I sitll only can find this information
"Security Update for Windows 7 for x64-based Systems (KB3167679)"

but that's fine with me, although I also found

==> KB3167679 breaks password changing on Win7 joined in NT style
but also
==> Changing Active Directory user passwords in c#/asp.net after MS patch KB3167679
(http://stackoverflow.com/questions/38997727/changing-active-directory-user-passwords-in-c-asp-net-after-ms-patch-kb3167679)

and for the other patch:

==> MS16-101 (KB3177108) Breaks password reset via AD
https://social.technet.microsoft.com/Forums/office/en-US/a3e9a9a2-b976-462b-b247-da6d72df119b/ms16101-kb3177108-breaks-password-reset-via-ad?forum=winservergen

or

==> Cannot contact domain controller
http://forum.netwrix.com/index.php?showtopic=2665
..
I had this exact problem this week, turns out windows update broke it. After having a look through the KB articles,
I'm pretty sure it's this one that borked it:
3177108
Specifically, this changes some stuff with the way AD behaves with password changes and resets.......yay!
..

In any case, you are using AD for LDAP, so I wonder whether a MS patch has been applied there as well?
I  checked our databases for any reported issues with one of the two MS patches, but didn't find anything.

==> BSM admin account works without any issues, but user accounts do not work. (the user accounts are hooked with LDAP).
error is
"This may be due to incorrect login name/password combination. Enter your login name and password again. ..."

Using the error message and parts of the logfile snippets,  I checked our DBs again and found a couple of issues with BSM 9.25 IP1, but most of them showed up after the upgrade from for example 9.24, not during a MS patch installation, but you might wanna try this out:

if not done during the IP patch installation, you might need to re-enter the LW-SSO Configuration using Step 16 "Update the LW-SSO Configuration" of the BSM 9.25 Patch
Installation Guide. Don't remove anything from the text, espcially no linefeeds, this will create a malformed version of this in the DB.
The best is to use Firefox to paste the InitString into the newValue field of the setGlobalSettingValue() JMX.
(An updated version of the manual now states
Note: Make sure that the contents of the lwssofmconf.xml file is formatted as a single line.)

==> from QCCR1I110909 Unable to log in with LDAP after upgrade

The parameter ldap.usersFilter was empty, the workaround is to set the value to

<setting nameResource="settings.ldap.usersFilter.name" name="ldap.usersFilter" displayInUI="false" refreshRate="Immediate" descResource="settings.ldap.usersFilter.desc"
sectionResource="sections.ldap"><value type="string">(&(uid=*)(objectclass=inetOrgPerson))</value></setting>

This needs to be done in two files,
It needs to be done in two files:
<HPBSM>\conf\settingsLDAPIntegration.xml
<HPBSM>\conf\LdapGlobalSettings.xml

==> from QCCR1I110909 Unable to log in with LDAP after upgrade
as well

In the BSM Management DB table setting_parameters there are two entries for the LDAP configuration,
a good one and one with NULL entries:
SELECT * FROM SETTING_PARAMETERS where SP_CONTEXT like 'ldap%'

After removing the entry with the NULL values and restarting BSM, LDAP users could login fine

if this all doesn't help or doesn't apply, enable DEBUG logging in the files
(create abackup copy first)
<HPBSM>\conf\core\Tools\log4j\EJB\topaz.properties
change
loginloglevel=INFO
to
loginloglevel=DEBUG

<HPBSM>\conf\core\Tools\log4j\EJB\authentication_wizard.properties
change
loginloglevel=INFO
to
loginloglevel=DEBUG

C:\HPBSM\conf\core\Tools\log4j\EJB\usersync.properties
change
loginloglevel=INFO
to
loginloglevel=DEBUG

wait two minutes, then try again to login
afterwards check the logs
<HPBSM>\log\EJBContainer\login.log
<HPBSM>\log\auth_wizard.log
<HPBSM>\log\usersync.log

As you cannot send in the logs, and cannot copy&paste them here, you might need to open a Support case if the steps from above do not help, and/or if you don't find an obvious error in the logs.

Greetings
Siggi

Hi Adr,

thanks for the clarification.
As far as I know the .mib file is not used actively within BSM, all the information (OID and alike) is hard-coded.
You say that you cannot open the dialog in the EUM alert in order to enter the IP of the destination (Netcool).
What happens when you click in Definitions Details on "specified host address(es)?
I cannot see an error ...

It should open a dialog like this one:


Greetings
Siggi

Hi niravu,

the path you select shouldn't matter, as you tell SiteScope what classpath to use for the files,
and then select the check box
"Load additional classpath first"

It might be good practice to place them into <SiteScope root directory>\WEB-INF\lib\ subdirectory,
as we do with the database libraries when using the database monitors,
or into <SiteScope root directory>\templates.applications\lib

Greetings
Siggi

Aaah, we finally reached the point of the problem ;-)

As this is BSM 9.26, give it a try with IE11, FF or Chrome.

Greetings
Siggi

Hi Adr,

I had a look at the code but couldn't find any DEBUG logging statements, so I actually cannot find out what logfile to look at,
but I found one interesting thing I forgot about initially:
..
boolean contentEnabled = authentication.isSNMPPermitted() && !storyPanel.isViewOnly();
..

Does your user have SNMP permission assigned, as described in the manual:
..
1. Set up the appropriate administrative privileges
You can set the appropriate administrative privileges to create a command that can be attached to
an alert scheme and run when the alert it is attached to is triggered.
To set the appropriate administrative privileges:
a. Select Admin > Platform > Users and Permissions > User Management.
b. Select the appropriate user in the left column, and click the Permissions tab.
c. Select the Platform context, and under Business Service Management, click Send SNMP
trap.
d. Click the Operations tab, and select the Change option.
..

did you try to reset the Infrastructure Settings
Default SNMP target address
and
Legacy SNMP target address
back to the default and try again?

Did you already try to setup another alert and try it there?

If this all doesn't help, check out the logs for any related looking errors (like unable to read setting, or a null pointer exception) or alike.

Greetings
Siggi

Hi AK2,

you need to use the BSMC 9.2x (9.23 is the latest version) to work with BSM 9.2x.
The 10 version of BSMC only works with OMi 10.x and above.

Your only option is to get hold of the BSM 9.20 bits, which also contain BSMC 9.20:
  Software_HP_BSM_Connector_9.20_TB998-15023.iso
and then apply the patch for your OS:

  BSM9IA_00011 - windows 2003-x32 OV OMI9.2 BSM Connector 9.23 Windows x32 
  BSM9IA_00012 - windows 2003-x64 OV OMI9.2 BSM Connector 9.23 Windows x64
  BAC_00838 - intelLinux RedHat5 OV BAC9.23 BSM Connector 9.23 RHEL5 x64
  BSM9IA_00010 - intelLinux RedHat5 OV OMI9.2 BSM Connector 9.23 RHEL5 x64

to upgrade BSMC to 9.23

I do have the .iso image here, do you have a FTP server I can use to upload the file (it's 2.3 GB in size)?

Greetings
Siggi

Hi James,

thanks for your reply and the good news.

Greetings
Siggi