A very large number of deleted pieces today

As in the title, I looked some time at the mesmerizing tail -f node.log and
I noticed an unusual amount of deleted files.

Yesterday:

cat node.log | grep 2023-02-08 | grep "uploaded" | wc -l
121692
cat node.log | grep 2023-02-08 | grep "downloaded" | wc -l
99864
cat node.log | grep 2023-02-08 | grep "delete piece sent to trash" | wc -l
34332

Today so far:

cat node.log | grep 2023-02-09 | grep "uploaded" | wc -l
25767
cat node.log | grep 2023-02-09 | grep "downloaded" | wc -l
31808
cat node.log | grep 2023-02-09 | grep "delete piece sent to trash" | wc -l
65388

Is there anything to worry about?

1 Like

Same here… about 100GB went to the trash in the past 12h and it’s still happening with a 10GB / hour rate.

Things happens :stuck_out_tongue:

1 Like

There was a very large spike in trash at the end of Nov, beginning of Dec, Jan, and now Feb. It seems like theres a large customer (or multiple) that delete(s) a lot of data fairly regularly. Its just how they use it.

5 Likes

Almost 600GB in the trash across my 3 nodes right now. I think that’s the most I’ve seen yet. I will not complain though. That means it will be replaced with data more likely to be uploaded to customers.

I don’t think so…
Almost all deleted files are the size of 2,212,608B and only my oldest node is shrinking. I think storj is cleaning up the test data. It was bound to happen. I think they won’t do it all at once so that the node operators won’t be too disappointed with the shrinking nodes…
I also think they shouldn’t do it while I’m migrating my largest node. Right now I’m rsyncing 600GB of trash for no good reason, it will all be deleted in a few days…

With my log parser it looks like this:


with + 300GB of trash on single node.

I’ll agree with the opinion from humbfig i guess

How old is your node @Dunc4n1d4h0 ?

It started in 2020-06.

Also an older one thanks.

I started 2019-11 and can confirm ~300 to 600GB trash per node in 3 days.

My rsync looks like this:

Yes same here my two oldest node got each trash around 500GB, so i think testdata gets deleted.

That’s a good thing if the data is from a customer who don’t access the files regularly, because there will be free space for fresh data which hopefully will be downloaded more frequently :slight_smile:

1 Like

You can’t erase data that is not accessed regularly… c’mon… get real…

Could you please explain why you backup your files?

My assumption is: If files are corrupt or lost, you get disqualified. You can’t react fast to prevent the disqualification. Is this correct?

1 Like

It’s correct, however @humbfig wrote

so they do a migration to a new location, not backup.

1 Like

I see deleted over 10TB over my nodes, but i dont care about it, more active data will come.

1 Like

image

Wow… 8% deletion

1 Like

Overall, a question related to the topic,
in the latest version, pieces are copied or moved to the trash?
I feel that if copied, a lot of people have too little space left for trash and number of active nodes will drop.

Well, if they are not moved (very inefficient), at least they get deleted after the copy. So, no problem whatsoever with available space.

1 Like

how do you like it Ilon !?