[Spread-users] Five minute timer?

[Mark Eliot]

I've got a Spread network of about a dozen computers.  Occasionally  
(more often than I'd like), when one is rebooted or its software is  
restarted, a particular computer (not the one restarted) will lose  
its connection with the rest of the network.  According to the Spread  
log, this computer thinks that it is part of its own network.  The  
rest of the computers appear to stay together in the common network.   
The curious thing is that *exactly* 5 minutes after the particular  
computer partitions itself, it rejoins the common network.  I've seen  
this behavior repeatedly.

So, questions for the group:

1.  Is there something magic about 5 minutes?
2.  Any ideas on how I can prevent, or at least minimize the time,  
this one computer is isolated?

Other info:  The isolated computer is a Mac running OS X 10.4.6 and  
Spread 3.17.3.  It has two IP nets.  Public net is has the Spread  
network; private doesn't.

Here's the partitioning event:

[Fri 09 Feb 2007 16:57:03] G_handle_trans_memb: Received trans memb  
id of: {proc_id: -2146303162 time: 1171069023}
[Fri 09 Feb 2007 16:57:03] Memb_regular
Membership id is ( -2146303162, 1171069024)
[Fri 09 Feb 2007 16:57:03] --------------------
[Fri 09 Feb 2007 16:57:03] Configuration at ams-server is:
[Fri 09 Feb 2007 16:57:03] Num Segments 1
[Fri 09 Feb 2007 16:57:03] 1 225.0.1.1 3333
[Fri 09 Feb 2007 16:57:03] ams-server 128.18.3.70
[Fri 09 Feb 2007 16:57:03] ====================
[Fri 09 Feb 2007 16:57:03] G_handle_reg_memb: with (128.18.3.70,  
1171069024) id
[Fri 09 Feb 2007 16:57:03] G_handle_reg_memb in GTRANS
[Fri 09 Feb 2007 16:57:03] G_handle_reg_memb: skipping state transfer  
for group AlarmMonitorIf.
[Fri 09 Feb 2007 16:57:03] G_handle_reg_memb: skipping state transfer  
for group InventoryIf.
[Fri 09 Feb 2007 16:57:03] G_handle_reg_memb: skipping state transfer  
for group LogMonitorIf.
[Fri 09 Feb 2007 16:57:03] G_handle_reg_memb: skipping state transfer  
for group NodeIf.
[Fri 09 Feb 2007 16:57:03] G_handle_reg_memb: skipping state transfer  
for group NodeMonitorIf.

Here's when the "ams-server" computer rejoins:

[Fri 09 Feb 2007 17:02:03] Send_join: State is 4
[Fri 09 Feb 2007 17:02:03] Memb_handle_message: handling join message  
from -2146303161, State is 4
[Fri 09 Feb 2007 17:02:04] Send_join: State is 4
[Fri 09 Feb 2007 17:02:04] Memb_handle_message: handling join message  
from -2146303161, State is 4
[Fri 09 Feb 2007 17:02:05] Memb_handle_message: handling join message  
from -2146303161, State is 4
[Fri 09 Feb 2007 17:02:05] Send_join: State is 4
[Fri 09 Feb 2007 17:02:06] Memb_handle_message: handling join message  
from -2146303161, State is 4
[Fri 09 Feb 2007 17:02:06] Send_join: State is 4
[Fri 09 Feb 2007 17:02:07] Memb_handle_message: handling join message  
from -2146303161, State is 4
[Fri 09 Feb 2007 17:02:07] Send_join: State is 4
[Fri 09 Feb 2007 17:02:08] Memb_handle_message: handling join message  
from -2146303161, State is 4
[Fri 09 Feb 2007 17:02:08] Memb_handle_token: handling form2 token
[Fri 09 Feb 2007 17:02:08] Handle_form2 in FORM
[Fri 09 Feb 2007 17:02:08] Memb_transitional
[Fri 09 Feb 2007 17:02:08] G_handle_trans_memb:
[Fri 09 Feb 2007 17:02:08] G_handle_trans_memb in GOP
[Fri 09 Feb 2007 17:02:08] G_handle_trans_memb: Received trans memb  
id of: {proc_id: -2146303162 time: 1171069328}
[Fri 09 Feb 2007 17:02:08] Memb_regular
Membership id is ( -2146303162, 1171069329)
[Fri 09 Feb 2007 17:02:08] --------------------
[Fri 09 Feb 2007 17:02:08] Configuration at ams-server is:
[Fri 09 Feb 2007 17:02:08] Num Segments 1
[Fri 09 Feb 2007 17:02:08] 10 225.0.1.1 3333
[Fri 09 Feb 2007 17:02:08] ams-server 128.18.3.70
[Fri 09 Feb 2007 17:02:08] scs-server 128.18.3.71
[Fri 09 Feb 2007 17:02:08] sns-server 128.18.3.72
[Fri 09 Feb 2007 17:02:08] trs-server 128.18.3.74
[Fri 09 Feb 2007 17:02:08] sds-server 128.18.3.75
[Fri 09 Feb 2007 17:02:08] srs-server 128.18.3.77
[Fri 09 Feb 2007 17:02:08] mvs-server 128.18.3.78
[Fri 09 Feb 2007 17:02:08] rms-server 128.18.3.81
[Fri 09 Feb 2007 17:02:08] sim-server 128.18.3.82
[Fri 09 Feb 2007 17:02:08] sim2-server 128.18.3.84
[Fri 09 Feb 2007 17:02:08] ====================

Thanks,
-M
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.spread.org/pipermail/spread-users/attachments/20070212/d78792ab/attachment.html