Today I suddenly experienced issues with deploying my multisite to my staging environment. Everything has been running fine for the last month, but I tried to deploy today and ran into an issue (don’t remember exact what issue) with deployment. When I tried to debug it, I found out I didn’t have any space left on my DO droplet. Normally my site takes up something like 6.4GB, so something used up the remaining 25GB.
I rebooted the droplet and the capacity of the HD returned back to normal, but without me doing anything, I noticed the HD slowly filling back up. I tried to re-provision, but got some errors (don’t remember what errors).
The next thing I tried was a deploy rollback. That seemed to work. From that moment on, my HD wasn’t filling back up again.
After that I was able to re-provision. I removed my last commit, so my code is the same as the working code deployed to production. But when I try to deploy again, it takes an eternity on the task [deploy : Get WP theme template and stylesheet roots] and when it finally continues and goes to the task [deploy: Remove WordPress transient containing old release path], I get the following error (replaced my IP address):
System info:
Ansible 2.7.13; Linux
Trellis 1.2.0: October 11th, 2019
---------------------------------------------------
The conditional check 'site_transient_theme_roots.stdout != ''' failed. The
error was: error while evaluating conditional
(site_transient_theme_roots.stdout != ''): 'dict object' has no attribute
'stdout'
fatal: [this.is.my.ip]: FAILED! => {}
to retry, use: --limit @/var/www/nexus_institute/trellis/deploy.retry
While it ran, the space on my HD was filling up again. When I check top on my droplet, I see 4 or 5 PHP processes taking up a lot of CPU (together probably almost 100%). I have no idea what’s going on. Any ideas?