logs:
03/10 19:42:53.427 DEBUG|          ssh_host:0153| Running (ssh) '/usr/bin/update_engine_client -status 2>&1 | grep CURRENT_OP'
03/10 19:42:53.696 DEBUG|        base_utils:0268| [stdout] CURRENT_OP=UPDATE_STATUS_IDLE
03/10 19:42:53.705 INFO |        servo_host:0510| servo host chromeos4-row9-rack7-host11-servo does not require an update.
03/10 19:42:53.707 DEBUG|          ssh_host:0153| Running (ssh) 'test -f /var/lib/servod/config'
03/10 19:42:53.885 DEBUG|          ssh_host:0153| Running (ssh) 'pgrep servod'
03/10 19:42:54.101 DEBUG|        base_utils:0268| [stdout] 476
03/10 19:42:54.102 DEBUG|        base_utils:0268| [stdout] 547
03/10 19:42:54.102 DEBUG|        base_utils:0268| [stdout] 548
03/10 19:42:54.106 INFO |        servo_host:0381| servod is running, PID=476,547,548
03/10 19:42:55.310 INFO |             servo:0496| Setting usb_mux_oe1 to on
03/10 19:42:55.653 INFO |             servo:0496| Setting prtctl4_pwren to off
03/10 19:42:57.915 DEBUG|             servo:0225| Servo initialized, version is servo_v3
03/10 19:42:57.915 INFO |        servo_host:0540| Sanity checks pass on servo host chromeos4-row9-rack7-host11-servo
03/10 19:42:57.982 DEBUG|          ssh_host:0153| Running (ssh) 'test ! -e /var/log/messages || cp -f /var/log/messages /var/tmp/messages.autotest_start'
03/10 19:42:57.983 INFO |      abstract_ssh:0749| Starting master ssh connection '/usr/bin/ssh -a -x   -N -o ControlMaster=yes -o ControlPath=/tmp/_autotmp_1XvZtNssh-master/socket -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o BatchMode=yes -o ConnectTimeout=30 -o ServerAliveInterval=900 -o ServerAliveCountMax=3 -o ConnectionAttempts=4 -o Protocol=2 -l root -p 22 chromeos4-row9-rack7-host11'
03/10 19:42:57.984 DEBUG|        base_utils:0177| Running '/usr/bin/ssh -a -x   -N -o ControlMaster=yes -o ControlPath=/tmp/_autotmp_1XvZtNssh-master/socket -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o BatchMode=yes -o ConnectTimeout=30 -o ServerAliveInterval=900 -o ServerAliveCountMax=3 -o ConnectionAttempts=4 -o Protocol=2 -l root -p 22 chromeos4-row9-rack7-host11'
03/10 19:43:03.052 INFO |      abstract_ssh:0764| Timed out waiting for master-ssh connection to be established.
03/10 19:45:06.505 ERROR|        base_utils:0268| [stderr] ssh: connect to host chromeos4-row9-rack7-host11 port 22: Connection timed out
03/10 19:45:06.506 INFO |            remote:0074| Failed to copy /var/log/messages at startup: ('ssh timed out', * Command: 
    /usr/bin/ssh -a -x    -o ControlPath=/tmp/_autotmp_1XvZtNssh-
    master/socket -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null
    -o BatchMode=yes -o ConnectTimeout=30 -o ServerAliveInterval=900 -o
    ServerAliveCountMax=3 -o ConnectionAttempts=4 -o Protocol=2 -l root -p 22
    chromeos4-row9-rack7-host11 "export LIBC_FATAL_STDERR_=1; test ! -e
    /var/log/messages || cp -f /var/log/messages
    /var/tmp/messages.autotest_start"
Exit status: 255
Duration: 123.423555851

+@jrbarnette, the ssh timeout could also be caused by the servo update?

Here's the error that triggered the failure:
    03/10 19:45:06.505 ERROR|        base_utils:0268| [stderr] ssh: connect to host chromeos4-row9-rack7-host11 port 22: Connection timed out

That's a problem with the DUT, not the servo.

I think (but I haven't checked in depth) that most servo
failures will be ignored in this context.

Adding more information, I think many of the failures in question
happen because the DUT is offline when provision starts.

I don't know why the DUT is offline; that needs investigation.

For the specific failure on chromeos4-row9-rack7-host11, I checked
the history:  The DUT was offline at the start of provisioning.
The provisioning failure triggered repair, and the servo repaired
the DUT by power cycling it.

The DUT is in service now, and seems to be running tests successfully.

@ richard - is there any more work needed on this? Should we use this as the FR for fixing provisioning to be smarter in the future? 

I've filed  bug 594828  for improvements to Provision task
failure diagnosis.

I don't think that for this failure there's anything else to
be done.