Preliminary information
The backup scripts are stored in /u3/scripts;
The log files for these scripts are stored in /u3/scripts/logs;
- The script u2_backup.sh is responsible to create a backup of /u2/dcs in /u3 and launch nfs_rsync.sh script to rsync data to nascoda1. The u2_backup.sh script is launched by crontab on host dafnenfs every day (see Table 2 for details);
- The script nfs_rsync.sh is responsible to rsync the data from the host where it is launched to nascoda1. It will send an email to Paolo Ciuffetti at his end to inform him on the backup status. It is launched by crontab on host danteweb and vlmonitor01 every day (see Table 2 for details);
- The host dafnenfs share /u3/scripts; danteweb and vlmonitor01 mount that folder via NFS on boot (folder present in /etc/fstab);
- The host dafnenfs has a custom configuration for logrotate present in /etc/logrotate.d/nfs_rsync to rotate log files generated by these scripts. It will rotate all log files every day, keeping them compressed in /u3/scripts/logs/old folder for 5 days.
Data
Control system data was saved in a Sun Storagetek 2550 Disk Array (LAT room) connected with two fiber optic paths to beatrix and virgil, the old production cluster. Now it's all stored and exported to us from the IT department of LNF; These are the folders in which data is organized:
| Host | Folder | Backup Folder (nascoda1/2) | Description |
|---|---|---|---|
| dafnenfs | /etc | server_data/dafnenfs/etc | Configuration files |
| dafnenfs | /data/diskless | diskless | Filesystem for VMIC diskless clients |
| dafnenfs | /data/export/exec/Solaris_9_sparc.all/usr | force_usr | User Filesystem for Force diskless clients |
| dafnenfs | /data/export/root | force_root | Root Filesystem for Force diskless clients |
| dafnenfs | /data/export/swap | force_swap | SWAP for Force diskless clients |
| dafnenfs | /data/u1 | u1 | |
| dafnenfs | /data/u2 | u2 | Control system, home folders (dafne, vmic e VMs) and other files |
| dafnenfs | /data/u3 | u3 | LVTemp, /data/u2 backup and other sensible files |
| dafnenfs | /siddharta | siddharta | Data written from Siddharta experiment and from DAFNE |
| dafnenfs | /data/backup_linac | - | vllinacctr001 VM backup (not backed up) |
| danteweb | /etc | server_data/danteweb/etc | Configuration files |
| danteweb | /opt | server_data/danteweb/opt | Service scripts and data, including a crontab snapshot |
| danteweb | /var/lib/mysql | server_data/danteweb/mysql | Database files (backup disabled; to be re-enabled soon) |
| danteweb | /var/www | server_data/danteweb/var/www | Web files (backup disabled; to be re-enabled soon) |
| vlmonitor01 | /etc | server_data/vlmonitor01/etc | Configuration files |
| vlmonitor01 | /opt | server_data/vlmonitor01/opt | Custom scripts (with Nagios custom plugins) |
| vlmonitor01 | /var/spool/nagios | server_data/vlmonitor01/var/spool/nagios | Nagios home (with public/private keys) |
Table 1 - data folders
Backup planning
Backup crontab are planned as Table 2 reports:
| Server | Backup frequency | Time | Description |
|---|---|---|---|
| dafnenfs | daily | 01:30 | Create a backup of u2/dcs in u3 and then launch rsync script |
| danteweb | daily | 01:10 | Launch rsync script |
| vlmonitor01 | daily | 01:05 | Launch rsync script |
Table 2 - Backup planning
Snapshot and replica
Backups are planned on NAS nascoda1.lnf.infn.it (LAT room) and rsynced on NAS nascoda2.lnf.infn.it (Control Lab, Building 2, Room 117) with following planning (including single volume snapshots):
| Source NAS | Destination NAS | Operation | Frequency | Time | Description |
|---|---|---|---|---|---|
| nascoda1 | local | Data volume snapshot | Daily | 00:05 | Snapshot kept for 1 month |
| nascoda1 | nascoda2 | NAS system volume snapshot and replica | Daily | 00:55 | Local NAS system volume snapshot and replica on nascoda2 |
| nascoda1 | local | server_data volume snapshot | Weekly (Sunday) | 01:05 | Snapshot kept for 1 month |
| nascoda1 | nascoda2 | Data and configuration volume sync | Daily | 04:00 | rsync from nascoda1 to nascoda2 of all volumes |
| nascoda2 | local | Data volume snapshot | Daily | 00:05 | Snapshot kept fpr 1 month |
| nascoda2 | nascoda1 | NAS system volume snapshot and replica | Daily | 00:55 | Local NAS system volume snapshot and replica on nascoda1 |
| nascoda2 | local | server_data volume snapshot | Weekly (Sunday) | 01:05 | Snapshot kept fpr 1 month |
Table 3 - Snapshot and replica planning
Monthly manual backup
Manual backup made every month (.tar.gz) of all data folder and saved in shared folder "DCS Backups" on nascoda1.lnf.infn.it
Recover data from backup
Read this documentation page.