Solaris

Hi,

I have my scheduled backups running on a Solaris Box using Veritas at 06:00 hrs daily. Now these backups are to be executed using Autosys Scheduler in a job workflow and hence the scheduled backups needed to be converted to scripts which could be executed by the Autosys Scheduler. On account of certain bugs ,some option do not work when converting Veritas Scheduled backups to shell scripts. These limitations have forced us to continue with the scheduled backups from veritas .

In order that the dependent activities on the Scheduled Backup such as an EOD batch run are run only after succesful completion of the backup by measn of the Autosys Scheduler we need to create a script which checks and ensures that the current backup is completed successfully. This is done by checking a log file dumped by Veritas for the successful completion of the backup. This log file is created everyday at 08:00 hrs.

Problem would be in a scenario whereby I have refired the scheduled backup successfully again during the day say at around 09:00 hrs for some reason. If the backup scheduled at 06:00 hrs the next morning fails but the check backup script which checks for only success status of the backup executes at 06:15 hrs ,it will find the entry of the successful backup fired at 09:00hrs of the previous day on the file and treat it as a successful backup.

Question here would be how would the script be able to check for the latest entry of the backup and determine the status based on the last entry. Have tried but am not getting any solutions. Would be great if anyone had any suggestions to counter this problem.

Regds,
Jobby

I did not understand that rambling text, but based on the title on this thread, maybe this is what you need...

Code:

$ cat file
one  something
two  something
three  something
last line
$ sed -n '/something/h;${;g;p;}'  < file
three  something
$

Hi,

Guess had too much coffee while posting the issue and hence the rambling text. Same works fine. Can I have a breakup of the command if it's not too much trouble. Just wanted some clarity... Thanx any ways for the update and Will try to keep the cofee under control...

Regds,
Jobby

if you can put the output into a file and then do tail -1 on the file you will get the last entry.

MizzGail, the last line in the file may not have the pattern being sought. Look at my sample. You would display "last line" rather than "three something".

That sed command:

Code:

sed -n        means do not print automatically...require an explicit p command
/something/h  means if a line matches something move it hold area
;             just separates commands
$;{...;}      $ addresses last line  and the braces let me run several cmds instead of just one
g;p           get the hold area and then print