Re: {bb} "bbnet Timeout with server" messages and purple conn tests

[Daniel O Winkless <wink@usgs.gov>]

Hi Jason,

        Exactly.  By default, a purple test means no status for that test has been received for at least 30 minutes.  I would expect your purples to turn red (slowly, one at a time), then return to purple 30 minutes later while the tests continue to fail.

        If you increase the number of threads the BBNET is allowed to use, you can shorten the time needed to run the test.  Many BB users have implemented fping, which can test many targets simultaneously -- just to avoid this purple problem when the network is behaving badly.

                                Dan

Jason Signalness <jason@btiadmin.net>
Sent by: owner-bb@bb4.com

12/21/2005 10:02 AM

Please respond to
bb@bb4.com

To	bb@bb4.com
cc
Subject	Re: {bb} "bbnet Timeout with server" messages and purple conn tests

My BBNET and BBDISPLAY servers are one and the same.
 So it doesn't seem 
likely that the client (the BBNET server in this case, right?) can't 
connect to the BBDISPLAY server (it's the same server). 

Is it possible that too many hosts became unreachable at the same time,
causing a backlog of network tests that was over 1/2 an hour, so things
went purple?

Thanks,
Jason


Sean MacGuire wrote:

> Jason Signalness wrote:
>
>> We've been having some unusual problems in our network, most likely
>> switch related.  But the odd thing is, our Big Brother server
doesn't 
>> react in any logical way.
>>
>> For example, a bunch of the "conn" tests go purple,
not red.  Since 
>> we don't get paged on purple, we don't get paged.  I noticed
these 
>> alerts in the BBOUT file:
>>
>> Wed Dec 21 04:20:42 2005 bbnet Timeout with server 1.2.3.4 on
port 22
>> Wed Dec 21 04:21:02 2005 bbnet Timeout with server 5.6.7.8 on
port 110
>> Wed Dec 21 04:21:22 2005 bbnet Timeout with server 9.10.11.12
on port 25
>> ... and so on.
>>
>> Can anyone explain why conn tests go purple, not red?  Under
normal 
>> operation (when our switch or whatever the cause of this issue
is 
>> behaves) if BB can't ping a host, it marks it red and we are notified.
>
>
> The network test info didn't make it back to the server, therefore
> the server just sees that it hasn't heard from the conn test for
> a while, and marks it purple.
>
> That's what those 'timeout' messages are - the clients can't
> reach the server to report in (for SSH/22 POP3/110 and SMTP/25).

--
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-==-=-=-=-=-=-=-=-=-=-=-=
To unsubscribe from this list, or to subscribe to the bb-digest list
send e-mail to mailto:majordomo@bb4.com with unsubscribe bb -and/or-
subscribe bb-digest in the BODY of the message.