Due to the circumstance I have no access anymore to the host in question, I'll report a problem occured out of the blue around last week's update of CURRENT with poudriere and swapspace. Problem: under heavy load, the host dies - no ssh connection possible anymore, all jails are in the state "dead". The box in question is running CURRENT, most recent, last update yesterday morning (28th of June, around 1400 UTC). Revision numbers are added as soon I have access to the box again. The host has 16 GB phsyical RAM and 64 GB configured swap - which the kernel complains about to increase swapzone or something similar. The host runs poudriere with both CURRENT and 12-STABLE jails (both recent versions). In the past 18 months we pushed the box to the limits with poudriere allwoing 4 poudriere jobs with each 4 threads - never had any problem except slowing down the system, but always responsive anyhow and never crashing or loosing network connection. The first time the box died this way was 28th, after the last update of both host and jails has been performed 26th June, ~ 1400 UTC. Jails running 12-stable are the first poudriere jobs running and that is the state were the first crash/hung occured yesterday. Is this a known problem? Kind regards, oliverReceived on Mon Jun 29 2020 - 05:03:45 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:41:24 UTC