SDP-241
Closed
Update systemd templates to add new wait targets.
Per an email from Giles:
Just in case anyone else happens to encounter a problem like this:
I had a call where a replica would start but not replicate after a reboot of a
Centos 7 machine on which it was running (don't ask about the validity of rebooting
without shutting the Helix Server down first !).
The error thrown for the 'pull' thread was:
'Replica access to P4TARGET server failed. Remote server refused requesst.
Please verify that service user is correctly logged in to remote server, then retry.
TCP connect to si-perforce:1667 failed. Name or service not known'
Using the SDP scripts to stop and restart the replica allowed replication to continue
normally.
It was found that the system was using systemd to control the start of the replica
at bootup; my suspicion was then that the network had not completed its startup
before P4D was started, hence the error.
This was confirmed by adding some additional parameters to the startup 'p4d_1.service'
file to have some control over the order of the startup (full procedure at the end of this email):
Requires=network.target NetworkManager-wait-online.service
After=network.target NetworkManager-wait-online.service
Lo and behold, P4D now waited for the network to come up before it tries to start and
replication takes place properly.
This may be simple stuff for some of you guys, but I hope that may help some people
and save some time in the future; I lay no claim to be anything more than a novice
with systemd so if there are experts out there you may be able to provide refining elements
to this.
See also: SDP-58
Per an email from Giles:
Just in case anyone else happens to encounter a problem like this:
I had a call where a replica would start but not replicate after a reboot of a
Centos 7 machine on which it was running (don't ask about the validity of rebooting
without shutting the Helix Server down first !).
The error thrown for the 'pull' thread was:
'Replica access to P4TARGET server failed. Remote server refused requesst.
Please verify that service user is correctly logged in to remote server, then retry.
TCP connect to si-perforce:1667 failed. Name or service not known'
Using the SDP scripts to stop and restart the replica allowed replication to continue
normally.
It was found that the system was using systemd to control the start of the replica
at bootup; my suspicion was then that the network had not completed its startup
before P4D was started, hence the error.
This was confirmed by adding some additional parameters to the startup 'p4d_1.service'
file to have some control over the order of the startup (full procedure at the end of this email):
Requires=network.target NetworkManager-wait-online.service
After=network.target NetworkManager-wait-online.service
Lo and behold, P4D now waited for the network to come up before it tries to start and
replication takes place properly.
This may be simple stuff for some of you guys, but I hope that may help some people
and save some time in the future; I lay no claim to be anything more than a novice
with systemd so if there are experts out there you may be able to provide refining elements
to this.
See also: SDP-58
- Status
- Closed
- Project
- perforce-software-sdp
- Severity
- C
- Reported By
- giles
- Reported Date
- Modified By
- C. Thomas Tyler
- Modified Date
- Owned By
- tom_tyler
- Component
- core-unix
- Type
- Bug