[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference molar::snmp

Title:SNMP
Moderator:MKNME::DANIELE
Created:Mon Apr 02 1990
Last Modified:Tue May 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:289
Total number of notes:900

286.0. "Sendto failed? results?" by CSC32::PITT () Wed Mar 26 1997 10:23

    
    Question.
    On a 4.0 system. If a network managment station sends out his
    snmpgets but is OUT of disk space so that the responses can't
    be written back to the NMS (in this case, it's a vendors NMS 
    running on a wnt node), would you expect the snmpd on the client
    (unix) node to fail with:
    
    Feb 21 16:47:20 rmicprdsrv1 snmpd[526]: Sendto failed: 2
    
    ???
    
    
    And what would be the result of that error?
    Would we expect snmp to just report the sendto error and
    go on as if nothing happened? or??
    
    thanks!
T.RTitleUserPersonal
Name		DateLines
286.1CSC32::PITTThu Mar 27 1997 10:4718
    
    Hi. 
    I'll restate my question since it might be unclear. 
    Real life example. 
    I have a PC running a monitoring product. It is set to a 15 minute
    polling interval. I ran out of disk space on the PC. 
    It continues to poll, but cannot receive the responses. 
    The unix node, who was polled and tried to respond, logs an
    snmp error (2). 
    I need to know if the snmp error (2) is enough to cause snmpd to go
    into some "wierd state" while he's trying to respond, or it's simply 
    reporting that it was unable to respond, so what, go on....
    I cannot reproduce the error (2) thru normal means and I am unable to
    get supporting information from the customer. 
    
    Please let me know if I need to IPMT this problem. 
    thanks!
    Cathy
286.2SMURF::DANIELEMon Mar 31 1997 15:1933
>    I have a PC running a monitoring product. It is set to a 15 minute
>    polling interval. I ran out of disk space on the PC. 
>    It continues to poll, but cannot receive the responses. 

>    The unix node, who was polled and tried to respond, logs an
>    snmp error (2). 

It's hard for me to believe the PC being out of disk space is in any
way related to snmpd logging Sendto failed: 2.  This is errno 2, 
"No such file or directory".

This would happen if a subagent shut down or cored, etc.

>    I need to know if the snmp error (2) is enough to cause snmpd to go
>    into some "wierd state" while he's trying to respond, or it's simply 
>    reporting that it was unable to respond, so what, go on....

No, snmpd should keep running just fine.

>    I cannot reproduce the error (2) thru normal means and I am unable to
>    get supporting information from the customer. 
    
>    Please let me know if I need to IPMT this problem. 
>    thanks!

Can you access MIB variables via snmpd?  Try logging into
the unix system and running snmp_request, for instance.

I expect it to still be running fine, in which case you don't need to
IPMT.

Regards,
Mike
286.3thanksCSC32::PITTTue Apr 01 1997 11:5718
    
    
    unfortunatly, what I gave you is all that I have. THe customer is
    unwilling to provide any additional information so we are left
    trying to piece together what we can. All we do know is that when
    (according to what he says), the PC sends out 'some' snmp request,
    but it is out of disk space, the errno 2 appears in his log file. 
    I can't provide more at this time, sorry...but I do appreciate your
    response on my note. I wasn't sure if the errno 2 would result because
    the file we were trying to write to was no longer writable. Maybe snmp
    would see it as an invalid file....I wasn't sure. But if what the
    customer says is true, then this is what is happening......
    I will add more detail if/when I can. 
    If you have any other ideas on this, feel free!
    
    thanks again. 
    
    Cathy