Opened 3 years ago

Last modified 3 years ago

#3581 new defect

LatentBuildWorker stop while starting race condition

Reported by: tardyp Owned by:
Priority: major Milestone: 0.9.5
Version: master Keywords:
Cc:

Description

If we stop the master while starting a latent worker, we got following in the log.

2016-07-12 14:47:08+0000 [-] Weird: Got request to stop before started. Allowing worker to start cleanly to avoid inconsistent state
2016-07-12 14:47:08+0000 [-] doing housekeeping for master 1 7e989d65be2d:/var/lib/buildbot
2016-07-12 14:47:08+0000 [-] (TCP Port 8010 Closed)
2016-07-12 14:47:08+0000 [-] Stopping factory <buildbot.www.service.RotateLogSite instance at 0x7f11583e3d88>
2016-07-12 14:47:08+0000 [-] (TCP Port 9989 Closed)
2016-07-12 14:47:08+0000 [-] Stopping factory <twisted.spread.pb.PBServerFactory instance at 0x7f11589fbd88>

Obviously, if 9989 port is closed, the worker can never start, so we hang the master until start timeout.

Change History (6)

comment:1 Changed 3 years ago by skelly

Related: ticket:2869.

My hypothesis for the cause at the time was because the startup is in its own thread to poll the instance being started.

comment:2 Changed 3 years ago by tardyp

  • Milestone changed from 0.9.0 to 0.9.1

comment:3 Changed 3 years ago by tardyp

  • Milestone changed from 0.9.1 to 0.9.2

Ticket retargeted after milestone closed

comment:4 Changed 3 years ago by tardyp

  • Milestone changed from 0.9.2 to 0.9.3

Ticket retargeted after milestone closed

comment:5 Changed 3 years ago by tardyp

  • Milestone changed from 0.9.3 to 0.9.4

Ticket retargeted after milestone closed

comment:6 Changed 3 years ago by tardyp

  • Milestone changed from 0.9.4 to 0.9.5

Ticket retargeted after milestone closed

Note: See TracTickets for help on using tickets.