5 of 5 linux/ppc64 builders are missing

[dmitshur]

From https://farmer.golang.org/#health:

/cc @andybons @golang/osp-team

[laboger]

I am trying to contact OSU for information on this.

[ceseo]

I see the system is responding to ping:

PING 140.211.169.164 (140.211.169.164): 56 data bytes
64 bytes from 140.211.169.164: icmp_seq=0 ttl=48 time=245.184 ms
64 bytes from 140.211.169.164: icmp_seq=1 ttl=48 time=256.996 ms
64 bytes from 140.211.169.164: icmp_seq=2 ttl=48 time=302.820 ms

Maybe it's just a matter of logging in and restarting the builder?

[laboger]

osu ticket #31316 was created for this. They are saying the machine is up and can be ssh'ed to. There were some issues yesterday that may have caused some instances to be rebooted. As Carlos noted the ppc64 builder might just need to be restarted. I don't have a key to get on those machines.

[ramereth]

@bradfitz please let me know if I can add anyone's key to the machine so that we can get it going again.

[cagedmantis]

@ramereth Thanks for your work on this. We are able to log in.

[ramereth]

The service seems to be running however it's erroring out with the following:

Oct 02 14:30:25 go-be-xenial-3 rundockerbuildlet[2975]: 2020/10/02 14:30:25 Creating go-be-%d01 ...
Oct 02 14:30:25 go-be-xenial-3 rundockerbuildlet[2975]: 2020/10/02 14:30:25 Error creating go-be-%d01: exit status 125,
Oct 02 14:30:25 go-be-xenial-3 rundockerbuildlet[2975]: See 'docker run --help'.
Oct 02 14:30:25 go-be-xenial-3 rundockerbuildlet[2975]: 2020/10/02 14:30:25 Creating go-be-%d02 ...
Oct 02 14:30:25 go-be-xenial-3 rundockerbuildlet[2975]: 2020/10/02 14:30:25 Error creating go-be-%d02: exit status 125,
Oct 02 14:30:25 go-be-xenial-3 rundockerbuildlet[2975]: See 'docker run --help'.
Oct 02 14:30:25 go-be-xenial-3 rundockerbuildlet[2975]: 2020/10/02 14:30:25 Creating go-be-%d03 ...
Oct 02 14:30:25 go-be-xenial-3 rundockerbuildlet[2975]: 2020/10/02 14:30:25 Error creating go-be-%d03: exit status 125,
Oct 02 14:30:25 go-be-xenial-3 rundockerbuildlet[2975]: See 'docker run --help'.
Oct 02 14:30:25 go-be-xenial-3 rundockerbuildlet[2975]: 2020/10/02 14:30:25 Creating go-be-%d04 ...
Oct 02 14:30:25 go-be-xenial-3 rundockerbuildlet[2975]: 2020/10/02 14:30:25 Error creating go-be-%d04: exit status 125,
Oct 02 14:30:25 go-be-xenial-3 rundockerbuildlet[2975]: See 'docker run --help'.
Oct 02 14:30:26 go-be-xenial-3 rundockerbuildlet[2975]: 2020/10/02 14:30:26 Creating go-be-%d05 ...
Oct 02 14:30:26 go-be-xenial-3 rundockerbuildlet[2975]: 2020/10/02 14:30:26 Error creating go-be-%d05: exit status 125,
Oct 02 14:30:26 go-be-xenial-3 rundockerbuildlet[2975]: See 'docker run --help'.

Here's what the systemd unit for that service looks like:

[Unit]
Description=Run Buildlets in Docker
After=network.target

[Install]
WantedBy=network-online.target

[Service]
Type=simple
# The (-n * -cpu) values must currently be <= number of host cores.
# The host has 10 cores, so the -n=5 (five containers) * -cpu=2 (two CPUs per container) == 10.
# -memory=3.9g doesn't work with crun; TODO: tiborvass is investigating
ExecStart=/usr/local/bin/rundockerbuildlet -basename=go-be-%d -image=golang/builder -n=5 -cpu=2 -memory= --env=host-linux-ppc64-osu
Restart=always
RestartSec=2
StartLimitInterval=0

[cagedmantis]

@ramereth I was able to get it back up and running by changing ExecStart to ExecStart=/usr/local/bin/rundockerbuildlet -basename=ppc64_ -image=golang/builder -n=5 -cpu=2 -memory= --env=host-linux-ppc64-osu. Thanks again for your help.