Yes and no. As @BrightSilence alluded to in one of his earlier posts, there have been quite a few bugs caught because of incongruences seen on the dashboards.
It’s a bit of a canary in the coal mine so they do have their uses 
7 Likes
Great analogy!
I understand there is some small benefit — but at what cost?
If the canary is sick all the time on its own and sometimes dies — it’s of questionable value, too many false positives. It’s best to get an electronic CO monitor and let the poor animal go, along with overhead of feeding, caring, and attending to her needs.
A lot of io is wasted on these databases. While my nodes don’t care — evidently, a lot of people here do (those that complain they the recent tests overwhelm their setups).
6 posts were merged into an existing topic: Updates on Test Data
I use some of the data the node reports. At some point, the log timestamps were changed to remove the millisecond part (now it only shows seconds) and this annoys me. Not that I can do much about it.
Also, used bandwidth by satellite (in the database) used to be updated constantly, now it’s updated every hour or so. That screwed up some of my graphs as well.
I have had these mixed folder too. I moved all these xy folders into a 2024-04-27 folder and the next filewalker deleted it.
3 Likes
What matters most to me is when the garbage does not get deleted. When a node cannot accept ingress because a mountain of garbage does not get deleted that is the worst case. We have this because the new date folder structure is botched on some nodes and we have this because bloomfilters have been deleted.
One of my nodes has date folders back to 22-04 and they do not get deleted.
One node is showing me it has 3.4 TB trash. I have no idea if that is correct. I have no idea about the current size of node as database does not get updated.
I use the API and the data that is in there and I want/need them to show correct values.
The reported bandwitdh thing is another issue. Nodes that accept uploads succesfully suddenly show only ingress of bandwidth of 130 bytes.
And then I suddenly discover a node that has 2 trash folders for the same satellite. With absolutely no idea what will happen from that.
On other nodes I see still a very big discrepancy between the disk usage and average disk usage and it does not seem to reduce. There are no bloomfilter files in the retain folder but I have date folders in the trash. It seems that processing got interrupted the latest date folder only has subfolders up to av. Interrupting probably has caused the bloomfilter to get deleted. So trash remains in the blobs folders.
No I don’t think these are just gimmicks.
1 Like
There is also a TTL information. Without it the pieces which could be deleted right away will be collected by a garbage collector later and will stay in the trash one more week.
And I see them mostly for for US-1 satellite folder.
Second often is EU-1 folder.
As the US-1 folder is the largest, it must have to do something with it.
Yes but now I have manually deleted all the other folders. I did wait for a month.
1 Like
Probably the only way…
1 Like