The timeout for reboot is 10 minutes.
I seems to be reasonable compromise between the two needs:
1) Detecting that compute node couldn't be restarted.
2) Allowing compute node to handle reasonable amount of updates.
There is completely no guarantee about reboot time, so everything we pick is a compromise.
Is there any specific timeout that you suggest to set?
The timeout for reboot is 10 minutes.
I seems to be reasonable compromise between the two needs:
1) Detecting that compute node couldn't be restarted.
2) Allowing compute node to handle reasonable amount of updates.
There is completely no guarantee about reboot time, so everything we pick is a compromise.
Is there any specific timeout that you suggest to set?