2 ppnsteve Mar 30, 2016 23:46

And maybe it's better if we set more program execution time (or unlimited. Is that may likely to lead to other problem if we set set_time_limit(0)
? ) in installing and upgrading procedure.
@ppnsteve if the upgrade process stops for any reason, then you need to delete/rename the umaintenance.html file manually. It's not a cache thing.
@haharen execution time is already set to reasonable high limits on upgrade process, they are set to 1800 seg and even 2400 seg. If the upgrade stops earlier due to execution time exceeded, then maybe the hosting provider doesn't accept these kind of overrides. So, even trying set_time_limit(0)
would make no difference.
@mgsolipa
I get it. Thanks for explaining.
@mgsolipa yes I did rename / remove the umaintenance.html file manually.. the cache (evocache) was what was timing out the backup function during the upgrade process.
When the upgrade stops, you should do this:
OK this update problem is back...... I expressly set do not backup the -cache and _evocache folders in the advanced settings yet there it is backing it up and getting stuck again.
I think this part of the auto-update process needs to be looked at again.
Now lets figure out how to get around this now. update: Once again clearing the cache folders allowed the upgrade process to complete. not sure why it locks up there but its something to look into.
OK we'll check this
No matter what, please make sure you have the scheduled task that periodically clears your cache folder correctly set up.
@ppnsteve please post a screenshot of what checkboxes were selected in your backup options when you say "I expressly set do not backup the -cache and _evocache folders in the advanced settings"
UPDATE: Just a heads up for others.. I removed the umaintenance.html file and went back and cleared the cache folders.. then reran the auto-update. It completed without issue.
For François and the b2evo guys: Suggest adding a clear cache option checkbox in the upgrade process to prevent this kind of failure in the future.