I am defo sure that my syntax was fine.
I have tried the same syntax to remove the logs manually from the
folder and it worked perfectly.
The cronjob was set from root, so I am assuming it has the right
privileges over the folder in cause.
@ Adrian the cluster runs on Rocky9
There is no error from cron that is the weird bit as well.
I do not have MAILTO set up on cron.
Regards,
Paul
On 02/02/2025 19:57, Peter J. Holzer wrote:
On 2025-02-02 12:12:07 +0000, Paul Brindusa wrote:
I had the exact same query as Junwang proposed.
Assuming that by "query" you mean the crontab entry:
Well, if if was *exactly* the same it's unlikely to work since you
probably don't have a directory literally called "/path/to/logs". If you
made the obvious substitution, it should work provided it runs with
appropriate privileges.
What is the output if you remove the «-exec rm {} \;» bit? What happens
if you reduce the mtime?
Was mega upset that I could not get the cronjobs to work, and from what I
can tell from @Laurenz's response above we have the names of the logs
customised to posgtres-%d-%m-%y.
Earlier you wrote that the pattern was actually
«postgresql-%Y-%m-%d.log». «find ... -name "*.log"» would find that but
of course not «posgtres-%d-%m-%y».
hp
