G8 H8

Bence Romsics bence.romsics at gmail.com
Mon Jan 28 10:31:14 UTC 2019


Hi,

On Mon, Jan 28, 2019 at 9:53 AM Lajos Katona <lajos.katona at ericsson.com> wrote:
> On 2019. 01. 26. 18:16, Slawomir Kaplonski wrote:
> >> Wiadomość napisana przez Matt Riedemann <mriedemos at gmail.com> w dniu 26.01.2019, o godz. 01:47:
> >> * Probably our biggest issue right now is test_subnet_details failing: http://status.openstack.org/elastic-recheck/#1813198. I suspect that is somehow related to using cirros 0.4.0 in devstack as of Jan 20. I have a tempest patch up for review to help debug that when it fails https://review.openstack.org/#/c/633225 since it seems we're not parsing nic names properly which is how we get the mangled udhcpc..pid file name.
> > I was looking at logs from failed job [1] and what I noticed in tempest log [2] is fact that couple of times this command returned proper „eth0” interface and then it once return empty string which, looking at command in tempest test means IMO that IP address (10.1.0.3 in above example) wasn’t configured on any interface. Maybe this interface is losing its IP address during renew lease process and we just should make tempest test more proof for such (temporary I hope) issue.
> I tried to do the same in a loop from cirros 0.4.0, but I can't remove
> IP from interface. Of course it's possible that something else happens
> there out of the command executed from tempests.

For what it's worth I think the problem can be reproduced like this:

1) take a cirros image (either 0.3.5 or 0.4.0)

2) boot a vm with it (I booted it by libvirt, didn't even use openstack)

3) look up the current ip of eth0 manually: ip a (here: 100.109.0.64)

4) run this command once:

sudo /bin/kill -USR1 $( cat /var/run/udhcpc.$( ip -o addr | awk '/100.109.0.64
/ {print $2}' ).pid )

There's no apparent error.

5) run the same command in a tight loop:

while true ; do sudo /bin/kill -USR1 $( cat /var/run/udhcpc.$( ip -o addr | aw
k '/100.109.0.64/ {print $2}' ).pid ) ; done

This reliably produces error messages like:

cat: can't open '/var/run/udhcpc..pid': No such file or directory
kill: you need to specify whom to kill
cat: can't open '/var/run/udhcpc..pid': No such file or directory
kill: you need to specify whom to kill
cat: can't open '/var/run/udhcpc..pid': No such file or directory
kill: you need to specify whom to kill
cat: can't open '/var/run/udhcpc..pid': No such file or directory
kill: you need to specify whom to kill
cat: can't open '/var/run/udhcpc..pid': No such file or directory
kill: you need to specify whom to kill

That's how far I got in debugging this at the moment.

Cheers,
Bence



More information about the openstack-discuss mailing list