It's very hard to debug problems with daemon. For example when exception here[1] occurs, you should print traceback, so we know what's the problem. I believe it can be improved on more places. [1] https://github.com/oVirt/ovirt-imageio/blob/master/daemon/ovirt_imageio_daemon/server.py#L160
Tentatively suggesting for 4.1. If we see this is non-intrusive enough, we ca definitely backport to 4.0.z too.
Please at least log the tracebacks of exceptions in 4.0.z. It's impossible to debug anything, because exceptions are swalloed.
Ondra, what are the versions of ovirt-imageio-daemon, ovirt-imageio-common in your host?
It's following: [vagrant@host1 ~]$ rpm -qa | grep imageio ovirt-imageio-daemon-0.5.0-0.201611201242.gitb02532b.el7.centos.noarch ovirt-imageio-common-0.5.0-0.201611201242.gitb02532b.el7.centos.noarch
Hi Ondra, We don't log traceback for 4xx status since it's originated from a user error (i.e. bad parameters), so a full traceback won't help. For 5xx statuses, the traceback should be logged. Did you encounter any other scenario that should have the traceback logged?
Hi Daniel, This one is most important I think, I didn't encounter any other.
@Nir - (In reply to Ondra Machacek from comment #6) > Hi Daniel, > > This one is most important I think, I didn't encounter any other. @Nir - do we have anything else to do in this case besides logging a warning?
(In reply to Daniel Erez from comment #7) > @Nir - (In reply to Ondra Machacek from comment #6) > > Hi Daniel, > > > > This one is most important I think, I didn't encounter any other. > > @Nir - do we have anything else to do in this case besides logging a warning? I still do not understand this bug. We don't have any details about the error that is not easy enough to debug without a traceback. If we don't get more data we should close this bug. The attached patch should be reverted.
(In reply to Nir Soffer from comment #8) > (In reply to Daniel Erez from comment #7) > > @Nir - (In reply to Ondra Machacek from comment #6) > > > Hi Daniel, > > > > > > This one is most important I think, I didn't encounter any other. > > > > @Nir - do we have anything else to do in this case besides logging a warning? > > I still do not understand this bug. We don't have any details about the > error that > is not easy enough to debug without a traceback. > > If we don't get more data we should close this bug. The attached patch > should be > reverted. Closing. Please re-open if a traceback is mandatory for another error.
Feel free to keep it closed if you think the logging of the daemon is good, I've just opened this bug, because it was really hard to work with logs for me, but maybe it's just me, so it's OK if the bug is closed.