While checking server uptime using cronjob, I found strange things.
My output from the cronjob shows multiple entries (on cerritos cluster).
Sometime, it seems like my cronjob is executed up to 4 times.
Also, the output of “ps” command is a bit strange.
Normally, the process id should grow and then restart from the smaller value.
But it doesn’t go like that on cerritos, nor on another server.
The PID jumps up and down as if our processes are assigned to arbitrary server for the given job.
In other words, it seems DH is using LoadBalancer of some sorts.
But on cerritos, somehow, they screwed up the setup and cron job might be running on multiple servers.
Well, that’s my guess, so far.
Anyone seeing something like these?
Any idea (or official/unofficial DH response/hint)?
As multiple execution of cron job may cause file corruption and or more serious consequences, DH shouldn’t take something these too lightly, IMO.
On cerritos, CGI output is fully buffered before sent to the browser, while on another server it isn’t.
Is DH using reverse-proxy type LoadBalancer or something similar on Cerritos ?