System Configuration Collector Server FAQ

For which version of scc-srv is this FAQ?
Release: 1.18.87
Copyright (C) 2014-2015 QNH.

What to do when I find a bug?
Check the SVN repository at sourceforge to find out whether the bug already has been fixed in the latest code. When that is not the case, please refer to the support page of the website.

Why does the rpm install fail with message: rpm is for a different operating system?
The rpm is produced on a Linux system and therefore rpm 'complains' on systems with other operating systems. Use the --ignoreos option to install the rpm on a system with a different OS.

Why does the rpm install fail with message: missing dependency /bin/sh?
The install scripts of the rpms require /bin/sh and on systems where this in a symbolic link created in the postinstall of an rpm, it is not provided and the install fails. Use the --nodeps option to install the rpm in this case. Please use the --requires option to verify the requirements of the scc rpms.

How are snapshots compared?
Parts of the snapshots of two different systems are compared to find the differences between these systems. All data whose classification start with 'fix' is extracted from the two snapshots into two temporary files. These two files are compared by means of diff and the output is processed to indicate to which snapshot/system each reported difference belongs.

Why is the picture of the logo missing on the main page?
Run scc-update with the -d option to indicate the path of the logo under the document root.

Why does the web interface not work?
When you read the documentation in scc-srv(5) and still cannot find the cause, use scc-debug(1) to collect specific configuration data of the server. Examine this data to find the cause of the problem. When you are unable to solve the problem and ask for assistance, please send this data along with the request.

Why does scc-realm --delete not delete a system?
Make sure you correctly specified the name. The program will not report absent files for hosts whose names are misspelled.

Why does the mailed SCC-data not arrive at the server?
First, run scc-update with the -m option to indicate the ownership of the destination directory of the email. When this does not correct the problem, look into the mail log to find other problem area's.

Why does a new top-level class not appear in the list for comparing?
To make scc-update faster, this list is not updated on every run of scc-update. Use the -f option to force scc-update to generate the lists based upon all systems in the realm.

Why does the -f option of scc-summary not update the summary files?
Make sure that the cache of your browser is not interfering.

Why is the style-sheet not working for the tables in the snapshot and logbook?
Most likely the scc-data has not been added via scc-transfer(1). Use the -u option of scc-transfer(1) to import the data and then use scc-update(1) to process the html-files properly.

How can I import scc-data into the website?
Usually scc-data is either pushed by clients or pulled by scc-pull(1). Another (server) method is the -u option of scc-transfer(1) that requires the snapshot, logbook and their html-files to pack them for processing by scc-update(1). These files should reside on the scc-srv system and are packed and copied to the transfer directory of scc-srv, where a subsequent scc-update will add the files to the website.

How do I get data from systems with identical hostnames?
SCC client uses hostname to name the files and this is also the system name in scc-srv. When 2 systems use the same hostname, the files of the last system overwrite in scc-srv the files of the first system. To avoid this, you should set the SCC_HOST_ALIAS variable in scc-localize of scc on one of these systems. Refer to scc(5) for details of this file.

How do I "randomize" running scc on many clients?
When the -d or --delay option with scc or scc-pull is used, the software waits for a random number of seconds before collecting configuration data. The argument of this option specifies the maximum number of seconds to wait.

Why am I unable to find data for "foo*matic"?
The input field for the search is handed over to grep, so you have to use regular expression syntax. So, use "foo.*matic" to search for a wild-card.

Why is the searched data by default limited to 10 lines?
Searching for non-specific data in a large list of systems, will flood your browser. To avoid this, the number of lines is limited to 10 when multiple systems are searched and 1000 when a single system is searched. Uncheck the 'Limit search results per hosts' checkbox to remove this limit.

When do you plan to move to a mysql backend to store the data?
There is little use for a database backend as all data is already classified by the client parts of SCC. As grep and diff are sufficiently optimized, little performance boost is to be expected from a database backend.

How can I handle a huge list of systems?
You can use realms to group systems. Access to data and several configuration items can be altered per realm. Use scc-realm to create and maintain realms. It is also possible to assign the SCC-data from new systems automatically to certain realms by means of the scc-auto-realm.conf configuration file. Refer to the manual pages for more information.

Why does scc-transfer report about a saved snapshot and log file for a host?
The new log file must contain the date and time of the current snapshot on the server. When this is not the case, the new log file may be the result of a run on a system with a restore of an old backup or the scc data arrives in the wrong order (when the mail connection was down for some time). To avoid loss of data, the snapshot and logbook on the server are renamed before the newly arrived data is put in place. Inspection of the renamed and newly arrived data cannot be done via the web-interface, it has to be done via the command-line.

At what frequency should I run scc-update?
The frequency and time of running scc-update should match the frequency and time of running SCC on the clients. When the clients send their data during the entire day, you can run scc-update on an hourly base.

How can I transfer SCC data from clients to server?
Basically there are three setups: push, pull and two-step. Refer to manual page scc-srv(5) for more details.

Why does scc-pull with -u option fail after upgrading the scc client?
This happens when the client is installed via RPM and the client did not yet run with the -u option. In that case the permissions of the transfer directory on the client are reset and have insufficient permissions to grant a non-root user acces from scc-srv. Run the client with the -u option before running scc-pull with the -u option to correct this.

Is it possible to preserve scc-data from obsolete systems?
Use the -A option from scc-realm(1) to archive systems from a realm that are no longer in use. You can combine this with the -d option to remove these systems to avoid them clobbering the web-interface.

How do I remove scc-srv from a system?
Start by removing the integration with Apache by means of "/usr/local/bin/scc-setup -r". Then remove the scc-srv software and configuration via the delete/remove option from the package manager or (in case of source install) issue command "rm -rf /usr/local /etc/usr/local". Finally remove the scc-srv data by means of "rm -rf /usr/local/share/scc-srv". Do not forget to reconfigure scc clients sending their data to scc-srv.