LMS 4.1 changing database + log location

[Eddie]

Goodday all,

Recently I installed LMS 4.1 accidentally on the c-drive, which, as a result, fills itself with a growing database and associated logfiles.

How can I move both items to another drive safely? I allready managed to move configs and downloaded software.

Thanks in advance!

[Vinod Arya]

Once the intallation is done there is no way to change install directory. Only for the long term if you think other drive have more space and will be beneficial, you can take a Backup of LMS and uninstall the existing install and re-install on other drive you want and Restore the backup.

It will have your long term worries fixed of continously thinking and managing disk Space and will not have any loss on you existing data from LMS.

If this is not in Radar for now, you can set the logrot for any log you think is increasing rapidly. Check he document on Maintaining Log Files and Configuring Log file Rotation.

There are some more of important things to keep in mind which may increase disk usage and we can keep a check :

> Configure Job Purge for LMS jobs

> Keep Syslog maintained and size under control.

> Keep configuration backup limited and

> Setting the Purge policy for Audit Trail.

Except for the Purge Policies, Changing Archive default directory and managing syslog their isn't much to keep a check on the Size of LMS. There is no way we can remove the DB to some other directory.

-Thanks

Vinod

[Eddie]

Hello Vinod,

thank you for your detailed reply. I decided to make a backup using the perl-backup script. After that I tried to remove LMS, but for some reason all Windows-related links to LMS were removed as if uninstalled, except for the fact that no folders or files were removed, and all services were running as if nothing happened - even after a restart. So I stopped all services, renamed the CSCOpx-folder to CSCOpx_old, and removed all relevant registry-entries. After that I was able to reinstall a clean version of LMS to the D-drive. That went fine. After restoring the data by using the perl-restore script (which took about 90 minutes with skyhigh cpu util.) I restarted the server. Again cpu was rather high, mostly 100% by several instances of causer. Now when I activate the Cisco Prime application, IE shows me a Forbidden-page (You don't have permission to access / on this server - Additionally, a 403 Forbidden error was encountered while trying to use an ErrorDocument to handle the request. All services are running.

Can you tell me what could have gone wrong here?

Thanks again!

[Vinod Arya]

This is strange and It is difficult to t/shoot after what you have done. As the previous version of ciscoworks was removed or not or if the essential registry entries were removed properly or not is still a question.

There is a script which can help to uninstall LMS,which you could have got if you would have posted before performing uninstall by deleting the registry directly and renaming folder.

You may want to see NMSROOT/MDC/tomcat/logs/stdout.log on why GUI is not coming up. Please try this once :

1. Stop the daemon manager (Net Stop crmdmgtd)

2. Regenerate the SSL certificate (Dont forget to start you shell/command prompt session as Administrator):

     *Remove server.*  files under NMSROOT\MDC\Apache\conf\ssl (take backup of all 4 files before deleting).

     *NMSROOT\bin\perl NMSROOT\MDC\Apache\ConfigSSL.pl -disable

     *NMSROOT\bin\perl NMSROOT\MDC\Apache\ConfigSSL.pl -enable

     *Fill all the details it request for.

     *Ensure that the following files are created under NMSROOT\MDC\Apache\conf\ssl :

server.crt

server.csr

server.key

server.pk8

3. Start the daemon Manager (Net Start crmdmgtd)

We may fix this this now, but this is not a trusted server for long run i guess.

-Thanks

Vinod

[Eddie]

Hello Vinod,

It appeared to be a matter of patience. After a while I was able to logon with the local admin account, and found all devices and settings as before. A little later I was able to logon with my own account using tacacs+. Now the only two things that are left are the roles (You do not have the roles required to access this portlet) and the 98% memory usage (8GB) with the 99% cpu usage.

Thanks!

[Vinod Arya]

Please check the Tacacs+ user you have logged in with has been configured in LMS and has the right set of permissions?

Tacacs+ users can only ger Authentication, but Authorization still needs to be configured locally. A high CPU is usually expected as soon as the server is up for some time.

To check which Ciscoworks Process is taking high CPU check the PID of that process in Task manager and try to match via pdshow.

Example:

Say the  PID of the highest CPU taking process is:  21468, check the PID process name as following:

C:\>pdshow -brief | find " 21468"

  ANIServer             Running with busy flag set                21468

-Thanks

[Eddie]

Hello Vinod,

I created similar users as in the ACS-server, and assigned the appropriate roles to them. This solved the role-assignment issue.

For some reason, the "ANIDbEngine process may be down"  when I open 'Discrepancies' on my dashboard.It says: Unable to connect to Data base. I checked the service, which seems to be running, and verified that the regional settings are set to English (US). Also the Collection Summary states 'Sorry, an error occurred'. Could there be something I forgot?

Thanks again!

[Vinod Arya]

ANIDbEngine stores topology and end hosts data for Network Topology, Layer 2 Services and User Tracking, hence anything related may not work if it is down.

What does the process says? Please post output of pdshow from server command prompt. I would like to see if there is any other related service is down too.

-Thanks

Vinod