[Spread-users] retrans of 1 requested while Aru is 2

[Martin Domig]

Hello

After upgrading to apache 1.3.26 and to spread 3.16.2, spread started to
show an odd behaviour: it crashed. And following murphys laws, it did so
at 3:00 am. 

After turning on some debug information I've got this:

INSERT HOLE 1 IS 2 My_aru is 1, Highest_seq is 3
INSERT HOLE 1 IS 3 My_aru is 1, Highest_seq is 3
Memb_handle_token: handling form2 token
Handle_form2 in FORM
Memb_handle_token: handling form2 token
Handle_form2 in EVS
Answer_retrans: retrans of 1 requested while Aru is 2
Exit caused by Alarm(EXIT)

A quick google search on the retrans-line pointed me to an old email
from this list:

> Answer_retrans: retrans of 8253 requested while Aru is 8411
>
> This happens when two networks are *very* out of sync, and an old
> token carrying retransmission requests for messages we don't have
> anymore comes in. Too bad.

Bad. Yeah.
After all, this gave me the hint to change the spread/network setup. Our
LAN is in deed quite busy, and using broadcast wasn't really improving
the situation. Now I use IP multicast, this seems to fix the problem.


> Note that this race condition still exists. It is *very* unlikely to
> happend now, but could happen under very specific conditions.

Obviously i've got the jackpot. Perhaps I should go to the casinos more
often :o)

regards,
Martin