| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| | |
|
| | |
|
| |
|
|
|
|
|
| |
Make the (request) response timeout adapt to the host
using the snmp_factor.
Also if a timeout occurs; display the content of the
message queue.
|
| |
|
|
|
|
|
|
|
|
|
| |
On some of our slow (and/or "unstable) machines, time checks may
produce results that are untrustworthy.
So, as a sort of sanity check, we check how long time it took
to get the snmpEngineTime. This should normally only take a couple
of milli sec (at most 100ms). So, if it takes more tha a second
then we can no longer trust the results. Therefor SKIP instead.
Also update copyright end date.
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
| |
Make it possible to configure agent transport (net-if) to
use the inet_backend gen_udp option.
OTP-17526
OTP-17410
|
| | |
|
| |
|
|
|
|
|
|
| |
A system that "stood still" for more than 20 minutes, caused
a test case to time out and then be terminated (when the test
server finally started "running" again).
By checking for system events (in this c ase also) we *may*
have time to skip the case before its killed...
|
| | |
|
| |
|
|
|
| |
Add a utility function(s) (and macro) for 'peeking' at the
message queue of a process.
|
| |
|
|
|
|
|
| |
When an request, or any message, does not arrive in time,
an error message is written. Now it also includes the
current message queue (the expeected message did not arrive,
but what did).
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
| |
Add transport kind test case(s).
*Reqests* sent from the manager to the agent (which sends a reply)
and *traps* sent from the agent to the manager. Both handled on the
respective transports.
Also adjusted the info test case (net-if info has been change).
OTP-16649
|
| |
|
|
|
|
|
|
|
| |
It "looks like" an exit signal (intended to pass a result)
from a test runner, passed the starter process and went
directly to the test server, which promptly failed the
test case. There was no sign that the starter process, which
supervises the test runner, died, but that is the only
reason I can think of...
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
We skip all IPv6 test case(s) oín the agent suite on
NetBSD. UDP on IPv6 is glitchy on the only NetBSD machine.
It could be because of configuration, but since the socket
suite works...anyway, its not a snmp problem, so skip
until the root cause is solved.
Improved handlng of udp error(s) in the test manager (used in
the agent suite). If sunch an error occurs, we exit. But since
its not an snmp error we don not fail the test case. Instead
we skip.
|
| | |
|
| |
|
|
|
| |
Add some and improve existing printouts in the agent
set/get (expect) wrapper functions.
|
| |
|
|
| |
OTP-15764
|
| |
|
|
|
|
|
| |
Use of the deprecated module random has been replaced
by the module rand.
OTP-15331
|
| |
|
|
|
|
|
|
| |
Improved test printouts (more with timestamps), also
fixed the printouts of the (fake) local-db start
result printout.
Also updated the copyright end-dates.
|
| | |
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
When testing the agent, it is started and then used for a series
(v1, v2, v3, ...) of tests. But the individual test cases are
use one instance of the test manager. Because of this, the requerst
id cannot be a counter starting from 1 or any counter for that
matter. Instead every request id was generated using random.
But this is obiously not a guarantee fo uniqueness, it merely
makes it unlikely that the request ids will be reused.
To get around this problem (without having to rewrite the
entire agent test suite) a simple (global) counter server is
introduced. It is started at the start of the agent test suite
(init_per_suite) and stopped at the end (end_per_suite).
|
| | |
|
| |
|
|
| |
OTP-9851
|
| | |
|
| |
|
