[adding ceph-devel as this may also be an inconvenient to others]
On 09/09/2015 10:23, Ma, Jianpeng wrote:> Hi Loic:
> Today, I run test/cephtool-test-mds.sh, because my code has bug cause osd
> down. I only from the screen saw "osd o down ...." and so on. But I don't
> find the related osd log.
> This because in vstart_wrapper.sh
>>> function vstart_setup()
>>> {
>>> rm -fr $CEPH_DEV_DIR $CEPH_OUT_DIR
>>> mkdir -p $CEPH_DEV_DIR
>>> trap "teardown $CEPH_DIR" EXIT
> This " trap "teardown $CEPH_DIR" EXIT" will remove the test dir(osd log in
> this).
>
> I don' know how to resolve this. But I think this really is a bug. Or am
> missing something?
You're right, it is inconvenient for debug. What I usually do in these
situations is add "exit" or even a sleep 10000000 somewhere to bypass
everything and give me time to take a look at the logs. There should really be
a way to ask (option, environment variable ?) "please keep the logs when it
fails instead of cleaning up".
What do you think ?
--
Loïc Dachary, Artisan Logiciel Libre
signature.asc
Description: OpenPGP digital signature
