Something isn’t updating the used space… still. I don’t know if it’s trash or something else. But…
I ran startup FW on my nodes; the dashboard was corrected and the graph on the left showed the same thing as the numbers on the right (even a little bit higher - see pic).
Now, after a week, once again it’s a discrepancy, and this difference increases each day, meaning the used space on the right is bigger than the graph on the left.
I just finished a FW on a node and I will post here the stats now and after a week or two, just to prove my point. The nodes are working fine, all the walkers are finishing with no errors (retain, GC, startup FW, etc.). This happens on all nodes, even on Ubuntu server with 32GB RAM. I run mostly Synologys with 18GB RAM and 2 nodes. The startup FW is disabled after a succesful scan.
Agree - just checked my logs and used-space file walker has not run since I restarted my node on 2 Sept, which was to force an update to the dashboard then following the corrected garbage collection.
Is this a known issue with v1.111.4 does anyone know?
Thanks
CC
The node in the example is at 112, but I believe the problem is much older.
Then it’s updated. There will be always a discrepancy between reports from the satellites and the local used space calculation - GC can move only 90% of the garbage to the trash, so there is always would be a discrepancy at least of 10%.
But does the used space on the right is matching the usage reported by the OS (you need to use SI measure units)? If yes - then you do not have the issue of this topic.
@snorkel Did your piece_expiration.db shrink from 28GB after the TTL deletes ?
Wouldn’t the discrepancy then be the other way around? In this case the right side should report disk usage around 11.13 TB (110% of the reported usage by satellites as 10% will not be garbage collected?)
Yes, but there is also a delay between reports from the satellites and a TTL collector for example.
But accordingly
The right pie chart has more used space showed than the graph on the left. I would not lean on an average, but on the last successful report from all satellites to see it as accurate as possible without involve of the averaging.
Something seems strange with the average. It is not really moving since days.
I cannot say for 100% but I see used space increasing which means that average should follow the trend.
I do see a last report as growing as well. The average would follow if each next report would grow, however, it’s an average and thus would change much slower.
After 2 weeks it seems that the used space is updated and displayed correctly. It didn’t go past the average on the left.