Skip Menu |
 
rt.cpan.org will be shut down on March 1st, 2021.

This queue is for tickets about the Net-SNMP CPAN distribution.

Report information
The Basics
Id: 78221
Status: open
Priority: 0/
Queue: Net-SNMP

People
Owner: Nobody in particular
Requestors: Jonas.Olsson [...] metaswitch.com
Cc:
AdminCc:

Bug Information
Severity: (no value)
Broken in: (no value)
Fixed in: (no value)



Subject: SNMP requests fails without error after timeout.
Date: Thu, 5 Jul 2012 14:36:03 +0000
To: "bug-Net-SNMP [...] rt.cpan.org" <bug-Net-SNMP [...] rt.cpan.org>
From: Jonas Olsson <Jonas.Olsson [...] metaswitch.com>
Download (untitled) / with headers
text/plain 1018b
Hello, and thanks for providing this module. I think I have found a bug/issue: If a blocking get_request is aborted halfway through the Dispatcher::loop function, in my case caused by a alarm in my code, the $INSTANCE->{_active} flag remains set to TRUE. This in turn means that all following calls to get_request, even from new sessions, instantly return an empty hash without sending any actual requests or raising any errors. In my code, I check for this behaviour and reset the flag, but this feels like an ugly hack. It would be nice to have a public method that resets the entire Net::SNMP library to its original state, even if this would mean losing some unsent requests. I am using perl 5.8.8, and Net:SNMP 6.0.1 (but this issue seems to exist in at least 5.1.0 as well). The code that encountered this problem was running on a Red Hat server, I don't know what version (and it is currently down for maintenance). Once again, thank you for providing this highly useful module! Sincerely, Jonas Olsson
Download (untitled) / with headers
text/plain 522b
I agree, it would be nice to have a function to clean up the dispatch queue. Or better yet, each Net::SNMP session instance should have it's own dispatcher instance. As it is, if you have a SIGALRM interupt a Net::SNMP session which is waiting on an unresponsive device, the dispatch queue will be in an inconsistent state, causing the next request to automatically fail. My fix for this is to have the following: $Net::SNMP::Dispatcher::INSTANCE = undef; $Net::SNMP::DISPATCHER = Net::SNMP::Dispatcher->instance();


This service is sponsored and maintained by Best Practical Solutions and runs on Perl.org infrastructure.

Please report any issues with rt.cpan.org to rt-cpan-admin@bestpractical.com.