Server management

Server configuration

Once storage space has been created a names file is then constructed for the server. Below is an example names file for a server. This particular names files contains entries for two databases one for calibration and one for a geometric database. (The next section of the tutorial will demonstrate the creation of these database files.)

An example names file for calib and geom dbs files

:nick.config
            :list.ge ca
            :log./hepdb/expdb/logs
            :bad./hepdb/expdb/bad
            :todo./hepdb/expdb/queue
            :queue./hepdb/expdb/queue
            :save./hepdb/expdb/save
            :logl.3
            :wakeup.60

:nick.ge
            :file.<CDEXPDB>.GEO.dbs
            :desc.Geometric database (residing on VM)
            :servers.caliblist1

:nick.ca
            :file./hepdb/expdb/CAL.dbs
            :desc.Calibration Database
            :servers.caliblist2

:nick.caliblist1
            :list.ecal1 ecal2

:nick.caliblist2
            :list.ecal2

:nick.ecal1
            :userid.cdl3
            :node.hepdb
            :localq./hepdb/l3/todo

:nick.ecal2
            :userid.cdl3
            :node.vxl3on
            :queue.disk$db:[cdl3.todo]
            :transport.tcpip
            :localq./hepdb/l3/tovxl3on

To understand the way the server uses the information in the names file one needs to examine the individual tags within it. A brief description of these tags starting with those in the configuration block follows however a more detailed account of this file can be found elsewhere in this text. Note that additional tags exist for use with the program CDMOVE. This tags are only required if you wish to distribute database updates between multiple nodes. See page for more details.

CONFIG

Indicates the start of the server configuration details.

LIST

A list of two character database prefixes. These prefixes act as a pointer within the names file giving specific details about each database file.

LOG

The directory where the server logs are written

BAD

The directory where the server places bad updates

QUEUE

The directory where new updates are placed by HEPDB clients

TODO

The directory which the server scans for new updates. If this is the same as the QUEUE directory then the server operates as the database master. In other cases it will operate in slave mode.

SAVE

The directory where the server saves updates after processing

LOGL

The log level for the server

WAKEUP

The wakeup interval in seconds for the server

SERVERS

The list of remote servers. Each database may have a different list of remote servers. The information on this tag should include all of the remote nodes listed for any of the individual databases, as described below. This tag is processed only by the CDMOVE server.

The next type of block in the names file are the database files blocks, identified by the string :nick.xx where xx is the two character database prefix discussed above.

:nick.xx

The two character database prefix, e.g. ca.

FILE

The full path name of the database file. For VM/CMS systems, the syntax is <user.address>filename.filetype

DESC

Gives a brief description of the database contents. It provides a way of documenting the names file.

SERVERS

The name specified in this tag is a pointer to a list of remote servers. In our example file above the description of our calib.dbs file refers to a serverlist caliblist1 which itself then points to details of the two remote servers ecal1 and ecal2. As described above, there may be a different list of remote servers for each database.

The remaining entries in our example names file are server descriptions. These are identified by an entry :nick.servername and store details of remote servers.

server

The nickname of the servers, e.g. ecal1.

USERID

Userid under which the server runs on the remote node

NODE

Node on which the server runs

QUEUE

Input queue on the remote node

TRANSPORT

Method by which updates are transmitted

LOCALQ

The local directory where updates are written pending transmission to the remote node. This may, in fact, be the same as QUEUE, e.g. when the directory is accessible via NFS or AFS.

Starting the server

Finally we must start the server up. According to the platform you are using the command to do this will vary slightly . The main forms of the command follow:

• UNIX and VMS The server can be run in the background or in batch mode.
• VM On VM the server can be started in one of two ways. You could autolog the machine CDexperiment (eg: CDEXPDB), or alternatively you could log on as for example CDEXPDB, type HDBSTART and then exit the session via the command #CP DISC.