Perhaps this can help put things into perspective:
It is quite evident that it’s not sarcasm on my end. The only real testing that has been done in the past 18 days has been updating the storj logo and making sure it’s pretty. There has been absolutely NO testing performed on the trash improvements, otherwise this testing would have caught the (number of) bugs mentioned in this thread. The only reason I was able to catch the trash cleanup not updating the databases (which a simple diff on the resulting database files would have caught instantly past merge-testing on your end) was because the monitoring system I use showed decreasing disk space, yet the nodes were still reporting full.
I’m not being sarcastic: It is very evident where storj’s priorities are. We (as SNOs) can only follow those priorities. And the current priorities are:
- make the product look pretty
- test the possibility of having larger clients onboard by migrating data from the production satellites to the testing satellite
What is NOT a priority:
- making sure that nodes don’t constantly die left, right, above, below and center
- keeping an accurate track of used, trash and free space
- actually cleaning up the trash
- expecting the team to open a bug internally with all the provided logs, df, du outputs already provided
Some will still argue that what I’m saying above is not the case, not true or FUD. The quoted parts are there, as are all the replies. To each their own, other SNOs as well as I, are only trying to keep the (remaining) nodes online to the best of our abilities.