Process Number: 212	Revision Date: 7/22/2003 5:33:53 PM
Revision #: 9	Back | Access Password:

Datatel System Administration

Office Responsible:

DCS - District - Computing Services

Process Type:

internal

Position Title:	Datatel System Administrator		Author:	bnoble
Cross Trained Staff:		Manager:	rkeith	Vice Pres.:

Timeline (Deadlines or Time-Constraints) :

As needed

What must be done before this process is started:

Which processes are waiting for this process:

Source documents or communication with information needed for this process:

What is the end-result of this process, or the hand-off

What steps must be taken independent of the computer system (Manual Process)

Which steps must be taken on a computer (Electronic Process):

Note: These are independent administration tasks and not sequential steps. 1. Maintain the Unix/UniData/Datatel cron. This is a process for scheduling the execution of regularly occuring processes. Login as "datatel" and enter the command "crontab -e" at the Unix prompt. This will start the VI editor and load the crontab text file. When you type ":wq" to save and exit, all changes will be activated. See the existing entries for examples and/or Unix documentation for instructions. 2. Monitor the weekly data cleanup process. The cleanup process happens weekly as part of the cron process, and reports are e-mailed to users specified in each cron file, currently bnoble. 3. Monitor the weekly UniData file analysis (WUFA) process. This process happens weekly as part of the cron process, and reports are e-mailed to users specified in each cron file, currently bnoble. The process includes automatic resizing of static files as needed. Damaged files and dynamic files that show a large level 2 overflow must be fixed individually. They are listed at the beginning of the report. See the UniDataAdmin.pdf file in the Document Library for more information on fixing these files. 4. Monitor the file system disk space usage and availability. A report is generated weekly as part of the cron process and e-mailed to users specified in each cron file, currently bnoble. If any of the file systems are nearing full capacity, some action needs to be taken before they become full. 5. Refresh the test and development accounts as necessary. Generally the test account is refreshed monthly immediately before patches are loaded, and the development account is refreshed whenever DCS decides and agrees that it should be done. See process 210. 6. Load patches into the live and test/development accounts as necessary. Generally this is done monthly, with the test/development account update happening a week before the live account update. See process 211. 7. Monitor UniData license usage. The program S02.UD.USE shows how many licenses are currently in use for each user and the total, as well as some tracking data for the length of the current run of the program. The program S02.UD.RESERVE is used to reserve UniData licenses for specified users. This program currently sends an e-mail to bnoble each time any user unsuccessfully requests a UniData license. This should be updated to send e-mail to anyone who monitors license usage. As a special point of interest, occasionally some of the TREG lines need to be shut down in order to free up licenses for other uses, depending on the priority of the activities. If TREG lines are shut down, be sure to check whether licenses are reserved for accounts "tregc" and "tregv" and reduce the number accordingly, or the licenses will not be freed for other uses. 8. If the ODBC drivers appear to be down (ATI filer, SARS use them) check for 3 critical files on SB2 that sometimes get deleted for no apparent reason. Check by running the unix "odbc.status" script. If they are missing, they should be copied from /datatel/live/collive to /datatel/live/collive/odbc.saves.

Related Documents to Process (e.g. Datatel documentation, Government code books, etc.)

Notes (Cautions, suggestions for improvement, etc.)