BBTest Error 🔗 link

On Fri, Jun 30, 2006 at 06:11:17PM -0400, David Gilmore wrote:

Pardon my ignorance, but what does this mean?
 
Error output:
Timeout waiting for data from child, killing it
Timeout waiting for data from child, killing it
Child process terminated with signal 15

On Fri, Jun 30, 2006 at 06:11:17PM -0400, David Gilmore wrote:

Pardon my ignorance, but what does this mean?

Error output:

Timeout waiting for data from child, killing it Timeout waiting for > data from child, killing it Child process terminated with signal 15

It is a message that can occur for the network tests (ntp, rpc and
traceroute) that rely on external programs rather than being implemented natively by Hobbit. It means that the program did not finish testing in 30 seconds, so Hobbit assumes it's hung waiting for some network I/O, and decides to treat this as a timeout error.

It's a safety measure to prevent e.g. a hung traceroute from stopping all of the network tests.

If 30 seconds is not enough, you can raise that with the --cmdtimeout=N option for bbtest-net.

And can I apply it to my five year old when she is not doing what I > ask?  :)

Perhaps ... but you'd better check with your local child-abuse authorities first :-)


Regards,
Henrik

On 11/20/06, Larry Barber <user-6ef9c2864140@xymon.invalid> wrote:

I am occasionally seeing the following error report from bbtest:

IO error in SSL_connect to https (47873/tcp) on host x.y.z.a: error:00000000:lib(0):func(0):reason(0)


Does anybody know what this signifies?

(my apologies for the duplicate post)

Thanks,
Larry Barber

On 11/20/06, Dan Vande More <user-f3c4c62d9d50@xymon.invalid> wrote:

I used to see this error w/ squid when using self-generated or chained
certs. I always figured it was something to do with the client not accepting
the cert or having an error checking the chain of trust.


On 11/20/06, Larry Barber <user-6ef9c2864140@xymon.invalid> wrote:

I am occasionally seeing the following error report from bbtest:

IO error in SSL_connect to https (47873/tcp) on host x.y.z.a: error:00000000:lib(0):func(0):reason(0)


Does anybody know what this signifies?

(my apologies for the duplicate post)

Thanks,
Larry Barber

On Mon, Nov 20, 2006 at 04:13:16PM -0600, Larry Barber wrote:

That doesn't seem to apply here, the cert is from Verisign and I don't
believe that it is chained. These are just intermittent failures that last
from 1 round of testing to several minutes, then go away. The websites
associated with the test also have problems when this error is reported, I
don't think this is a Hobbit problem, I'm just trying to help with the
trouble-shooting.

On 11/20/06, Henrik Stoerner <user-ce4a2c883f75@xymon.invalid> wrote:

On Mon, Nov 20, 2006 at 04:13:16PM -0600, Larry Barber wrote:

That doesn't seem to apply here, the cert is from Verisign and I don't
believe that it is chained. These are just intermittent failures that
last
from 1 round of testing to several minutes, then go away. The websites
associated with the test also have problems when this error is reported,

I

don't think this is a Hobbit problem, I'm just trying to help with the
trouble-shooting.

I've also seen this occasionally with https tests. The frustrating thing
is that OpenSSL reports an error, but doesn't provide any information
whatsoever about what the error is (that's the "error:0...." message).
But since it happens when calling SSL_connect, I suppose there is some
sort of problem with the SSL handshake (this error happens *after*
Hobbit managed to connect at the network - TCP/IP - layer, so the
connection to the webserver has been established). It could be a lack of
entropy (randomness) on the Hobbit server so it cannot generate a
session key, but that is just a wild guess - and it doesnt really sound
like something that would happen.


Regards,
Henrik

On 11/20/06, Larry Barber <user-6ef9c2864140@xymon.invalid> wrote:

I don't think it has anything to do with Hobbit or the Hobbit server, as
the website goes down when we see this error as well.

Thanks,
Larry Barber

On 11/20/06, Henrik Stoerner <user-ce4a2c883f75@xymon.invalid> wrote:

On Mon, Nov 20, 2006 at 04:13:16PM -0600, Larry Barber wrote:

That doesn't seem to apply here, the cert is from Verisign and I don't
believe that it is chained. These are just intermittent failures that
last
from 1 round of testing to several minutes, then go away. The websites
associated with the test also have problems when this error is
reported, I
don't think this is a Hobbit problem, I'm just trying to help with the

trouble-shooting.

I've also seen this occasionally with https tests. The frustrating thing
is that OpenSSL reports an error, but doesn't provide any information
whatsoever about what the error is (that's the "error:0...." message).
But since it happens when calling SSL_connect, I suppose there is some
sort of problem with the SSL handshake (this error happens *after*
Hobbit managed to connect at the network - TCP/IP - layer, so the
connection to the webserver has been established). It could be a lack of

entropy (randomness) on the Hobbit server so it cannot generate a
session key, but that is just a wild guess - and it doesnt really sound
like something that would happen.


Regards,
Henrik