[1906] in Moira
Re: host IP assigned out of range
daemon@ATHENA.MIT.EDU (Garry Zacheiss)
Mon Dec 10 14:13:31 2001
Message-Id: <200112101908.OAA10955@riff-raff.mit.edu>
To: Angie Kelic <sly@MIT.EDU>
cc: network@MIT.EDU, bug-moira@MIT.EDU
In-Reply-To: Your message of "Mon, 10 Dec 2001 12:17:18 EST."
<200112101717.MAA27011@prowler.mit.edu>
Date: Mon, 10 Dec 2001 14:08:56 -0500
From: Garry Zacheiss <zacheiss@MIT.EDU>
>> I have no idea how this could have happened unless moira somehow
>> calculated the address space incorrectly when told to assign a unique IP
>> address or if the network numbers changed at some point.
The latter was the case; on Feb 2, we had:
Feb 2 10:10:07 moirad[#66767]: query[2]: "update_subnet", "W84", "W84", "Building W84", "318439424", "-65536", "318439451", "318440953", "tang", "LIST", "ht-dorm-acl"
Feb 2 12:15:25 moirad[#67168]: query[2]: "update_subnet", "W84", "W84", "Building W84", "318439424", "-65536", "318439451", "318440697", "tang", "LIST", "ht-dorm-acl"
The interesting big number is the last one, which corresponds to
the high range. The first one set the high number to 18.251.5.249,
which overlapped with the DHCP range. The host you found was created
in the two hour window between then and when somebody noticed at noon
and set the high number to 18.251.4.249.
So there's no calculation bug; there's an arguable bug that the
subnet queries shouldn't let you change the range of a network if there
are any hosts on the network not in the new range. If someone in
network thinks that's desirable behavior and wants me to implement it,
they should let me know.
Garry
