[Swan-dev] should final.sh shut down pluto?

[Paul Wouters]

> On Jun 26, 2016, at 03:49, D. Hugh Redelmeier <hugh at mimosa.com> wrote:
> 
> I'm quite unhappy about uncaptured core files.  At the moment, I'm 
> not willing to say what a good solution would be.

That's not quite was it is. Not all tests perform a shutdown and so some errors in a shutdown that could cause a core would theoretically not be detected". Any core files made are always detected.

As antony said, it is VERY useful to run a single test and then ssh in while the tunnel is still up. We have a bunch of tests running shutdown but don't for the majority of tests. I think that's fine.

Paul



> 
> I did learn from Andrew that this is probably only due to the core being 
> generated during shutdown (or perhaps something else in final.sh).  
> Here's what he said in IRC:
> 
> <cagney> DHR, if pluto isn't explicitly shutdown (by final.sh) then both kvmrunner 
>    and swantest miss asserts generated while pluto shuts down as part of the 
>    KVM shutdown.  I sent something a while ago about this to the mailing 
>    list, issue is still open
> <cagney> same, but worse, can apply to core dumps
> 
> <antony> cagney swantest run with  TESTLIST should catch that
> <antony> 2091     if "is_testlist" in st:
> <antony> 2092         e = shut_down_hosts(args, test_hosts)
> 
> | From: Andrew Cagney <andrew.cagney at gmail.com>
> | Date: Fri, 5 Feb 2016 15:50:01 -0500
> | Subject: [Swan-dev] should final.sh shut down pluto?
> | 
> | While this question is kind of philosophical, there's a real problem behind it.
> | 
> | I've noticed that when running a postmortem on a test run I'm finding
> | assertion failures:
> | 
> | testing/pluto/newoe-05-hold-pass failed road:ASSERTION
> | testing/pluto/newoe-11-failike failed road:ASSERTION
> | testing/pluto/newoe-18-clear-poc failed east:ASSERTION
> | testing/pluto/newoe-18-clear-private failed east:ASSERTION
> | 
> | yet the result files seemed ok.  For swantest, if you happen to be
> | watching the console carefully you will see:
> | 
> | sanitizer output road Consoleoutput matched
> | east Consoleoutput matched
> | result newoe-05-hold-pass passed
> | copying results
> | grep 'ASSERTION FAILED'
> | /home/cagney/libreswan/testdir/testing/pluto/newoe-05-hold-pass/OUTPUT/road.pluto.log
> | "private-or-clear#7.0.0.0/8" #11: ASSERTION FAILED at
> | /source/testdir/programs/pluto/kernel.c:3413: LHAS(LELEM(CK_PERMANENT)
> | | LELEM(CK_INSTANCE), c->kind)
> | 
> | but if you're off doing other things you'll miss this :-(  If you run
> | just one test in isolation then even the above doesn't appear.
> | 
> | The problem is that pluto crashes while the VM is being shutdown and
> | that is well after the point where final.sh (which tries to capture
> | core files) has been run.
> | 
> | So, should final.sh be responsible for ensuring that pluto shutdown cleanly?
> _______________________________________________
> Swan-dev mailing list
> Swan-dev at lists.libreswan.org
> https://lists.libreswan.org/mailman/listinfo/swan-dev