Hitting BackBlaze B2 Class C transaction limits

Support
#1

So both myself and a friend who have Duplicacy setup very similarly are both running into Class C transaction limits on BackBlaze B2 storage.

He hits it more frequently than I do, and I went ahead and setup my transaction limit to $0.01/day (doubling the amount of class C transacations that way) and I still hit 100% usage last night and I guess I’m not sure why.

Here’s my Duplicacy config for backing up to BackBlaze, along with prune/check commands.
It faled during the prune, which I run an exhuastive prune one ot make sure we are pruning propely by removing everything that shouldn’t be there. – Perhaps the exhaustive prune is the issue though? Looking for some guidance here.

image
image1773×762 71.2 KB

Also here’s what my transactions were

image
image925×512 30.4 KB

Thank you in advance!

#2

I think you’re right. I used to run into something similar when I used -exhaustive on all of my scheduled prune operations, rather than only using it ad-hoc.

The -exhaustive option requires that all objects in storage be listed, which (AFAIK) is not something that’s necessary for a typical prune operation.

#3

Ok so I removed -exhaustive and I’m still using a boatload of Class C transactions on random days.

My backup jobs aren’t changing and I don’t have a huge file turnover, but some days I use a small fraction of class C (1k or less), and other days I’m using over 5000 (like last night)

Any advice on how I can reduce this, or at least understand why it’s happening some days and not others?

image
image925×496 30.2 KB

#4

If the prune job has many fossils to delete it may generate a lot of b2_list_file_versions calls. You need to look at the log to see if there are an unusual amount of activities on a day with high class C usage.

#5

So I’ve got thousands of lines of

2020-12-23 02:00:59.154 WARN CHUNK_DELETE Failed to remove (then the chunk file)

It seems odd though because I removed my daily cap, ran an exhaustive prune, and it deleted snapshots but I don’t see any chunks being deletd from it?

Would it make more sense to change my retention to just 0:30 to keep 30 days worth instead of 0:30 7:14? so it doesn’t have weird random daily fuctuations so high?

I guess I’m not fully understanding where the fossils are coming from unless it’s the 7:14 part of that prune job?

Thanks in advance!

#6

I do not understand how some days I am using under 200 class C, and other days I’m using over 5000, it does not make sense to me?

Can someone assist in explaining why there would be such a drastic difference from day to day with nothing much changing in the data set?

image
image1384×480 39.6 KB