Bug #2078
Interface cleanup - not correctly removed from UPnP when deleted
| Status: | New | Start date: | 01/10/2012 | |
|---|---|---|---|---|
| Priority: | Low | Due date: | ||
| Assignee: | - | % Done: | 0% |
|
| Category: | uPNP | |||
| Target version: | - | |||
| Affected version: | 2.0.x | Affected Architecture: | All |
Description
When deleting an interface that had previously been active with UPnP, that interface apparently is not correctly removed from the UPnP config, or possibly the service just isn't restarted or HUP'd as appropriate.
I deleted an interface quite a while ago, and I happened to look at system.log today while updating to 2.0.1, and I saw many many many lines of this (at different dates/times obviously):
Jan 10 00:59:49 pf miniupnpd20514: sendto(udp_notify=13, 172.16.42.1): No buffer space available
172.16.42.1 was the IP of the interface that I deleted.
There didn't seem to be any impact to the system, but the "no buffer space available" message did worry me a little (was there potential for some resource exhaustion here and a problem due to that?) I was going to mark "Very Low" due to the lack of any obvious impact, but I'm leaving it at "Low" in the event that it could've been eating resources somehow.
I unfortunately can't provide a full copy of system.log because it is gone due to the restart, but I had some of the lines in my console scrollback (I had SSHd in).
History
#1
Updated by Chris Buechler over 1 year ago
"no buffer space available" in scenarios like that is just a generic "I can't do that" complaint. Won't hurt anything, but should be fixed.
#2
Updated by Braden McGrath over 1 year ago
Gotcha. Feel free to alter it down to "very low" unless you guys save that for things like spelling errors. :) I figured it wasn't a big problem but I wanted to report it regardless.