Created attachment 768240 [details] engine.log Description of problem: rhevm is stucked while loading an unaccessible config file during startup. Not sure how rhevm should handle this... rhevm loads files in $ENGINE_ETC/conf/osinfo.conf.d directory, if i create a symlink in this directory pointing to problematic nfs share, rhevm would never load correctly, it's just stucked... Version-Release number of selected component (if applicable): is4 How reproducible: 100% Steps to Reproduce: 1. create a symlink in $ENGINE_ETC/conf/osinfo.conf.d pointing to a nfs share 2. disable access to this nfs share (iptables -I INPUT -s $your_nfs_host -j DROP) 3. restart rhevm Actual results: rhevm stucked while starting up Expected results: no sure... but what about some timeout and then fail with an ERROR? Additional info: mount $server:$export /mnt echo 'foobar' > /mnt/test ln -sf /mnt/test $ENGINE_ETC/conf/osinfo.conf.d/11-link-to-unavailfs.properties iptables -I INPUT -s $server -j DROP service ovirt-engine restart check log around 12:30 and later...les
what do you mean stuck? if your nfs share is "hard" it will hang until you fix it. That's exactly the behavior it should have if your nfs is using "soft" it will time out and error out. Not sure - worth a check if it fails loading osinfo file only or exits completely. The latter would be best and if it's not the case this bug should be fixed - engine should fail to start.
Yes, if using 'soft' option for nfs mount, after timeout engine continues in startup and finishes. The original question was if engine should guess it is starting up for long time thus being stucked somewhere. Feel free to close it as user problem.
Closing old bugs. If this issue is still relevant/important in current version, please re-open the bug.