Conflicts:
sdk/cli/bin/crunch-job
ban_node_by_slot($job_slot_index);
}
elsif ($line =~ /srun: error: (Unable to create job step|.*: Communication connection failure)/) {
- $jobstep[$job]->{tempfail} = 1;
- ban_node_by_slot($jobstep[$job]->{slotindex});
+ $jobstep[$jobstepidx]->{tempfail} = 1;
+ ban_node_by_slot($jobstep[$jobstepidx]->{slotindex});
}
- elsif ($line =~ /arvados\.errors\.Keep/) {
+ elsif ($line =~ /\bKeep(Read|Write|Request)Error:/) {
- $jobstep[$job]->{tempfail} = 1;
+ $jobstep[$jobstepidx]->{tempfail} = 1;
}
}
}