Notes
Introduction
1. Computer O/S and Utility S/W
2. Home Directories and JSOC Code CVS Tree
3. Postgres DRMS/SUMS Database
4. Data Capture Archive Tapes
5. Pipeline Archive Tapes
TBD - Sample chart of configuration information.
Introduction    (Top)
This is a backup plan in the event of loss of any or all of the JSOC SDP processing center at Stanford. There are these main sections of the backup plan: 1. Computer O/S and utility S/W 2. Home directories and JSOC code CVS tree 3. Postgres DRMS/SUMS Database 4. Data Capture Archive Tapes 5. Pipeline Archive Tapes
1. Computer O/S and Utility S/W    (Top)
It is assumed that we will have an heterogeneous mix of computers, O/Ss, O/S versions, and utility program versions (icc, perl, perl/tk, etc). We will attempt to keep common O/Ss and utilities at a common version, but some discrepancy is bound to arise. We will have configuration control of the O/S and utility versions, knowing which is where when. The CM will contain full h/w and s/w configuration info. The configuration information will allow restore of the O/S or utility from the original distribution, the location of which is also part of the configuration info. Also there will be a disk image of the system disk, that will allow for a quick, full restore of O/S and utilities. The procedure for such restores is also part of the CM info. There will be an off-site storage center at the Lockheed Martin Solar and Astrophysics Laboratory (LMSAL) in Palo Alto California. Newly generated hard media products will be physically transported to the off-site center twice a week. Electronic backup products will be copied to the off-site center as they become available. There will be both local and off-site storage of the following products: 1. O/S and Utilities Original distributions of hard media or download files. 2. System Disk Images Restorable images of O/S and utilities. 3. Home directory level 0 and level 1 backups 4. Data Capture System telemetry tapes 5. Pipeline Processing System Level 0 tapes (?) 6. Postgres Point In Time Recovery tape 7. Full configuration information JSOC h/w and s/w components and versions, backup info, times and locations and restore procedures as well as motivation for changes.
2. Home Directories and JSOC Code CVS Tree    (Top)
This is our normal nightly backup procedure using the Legato Networker. The JSOC CVS tree is under /home/cvsuser and so is contained in the backup of the home directories. The backup will also be of /usr/local on our web server machines. The backup is an incremental nightly backup with semimonthly level 1 (changes since last level 0) and quarterly level 0 (full) backups. Duplicates will be made of the quarterly level 0, and the semimonthly level 1 tapes and will be sent to off-site storage. A full restore can then be done from the offsite tapes to the last semimonthly period.
3. Postgres DRMS/SUMS Database    (Top)
We will do an On-line backup and point-in-time recovery (PITR) for our Postgres Data Base. A full discussion with sample backup and restore procedures is given in:http://jsoc.stanford.edu/production/postgres_backup_restore.html
4. Data Capture Archive Tapes    (Top)
The Data Capture system archives to tape all the telemetry products received from the Data Distribution System (DDS) at White Sands. All these tapes are off-site storage tapes. The amount is approximately 1.4 TB per day, and will be about 2 tapes per day.
5. Pipeline Archive Tapes    (Top)
Currently it is assumed that there will be no pipeline archive tapes that will be needed in off-site storage. All the pipeline data products can be rebuilt as necessary.
Last modified @ 14:21 Mon Oct 9 2006