Restore Error from B2

Support
#1

I am testing the restore from B2 (now that download is free) and encountered the error below. I encountered this with 10 threads and 3 threads so I don’t think the threads setting matter. I am running v1.6.3 and the data is on B2 encrypted. On the last note, I tested restore a simple file and that came down without any problems and unencrypted so I know that is not the issue. Is this a Backblaze issue, the destination storage (12TB HDD), or something else? Thanks.

Error Message in dialog box
The restore command encountered an error:
Failed to download the chunk c65372d62b11d89cd4822cdc2fcdd4b1e6d2f3dba4b5f6ad699f41f13fa7f51d: read tcp 10.0.7.106:52826->149.137.128.16:443: wsarecv: A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond.

Exit code: 100

OkView log

Error Line in Log File
2023-10-08 01:02:10.744 ERROR DOWNLOAD_CHUNK Failed to download the chunk c65372d62b11d89cd4822cdc2fcdd4b1e6d2f3dba4b5f6ad699f41f13fa7f51d: read tcp 10.0.7.106:52826->149.137.128.16:443: wsarecv: A connection attempt failed because the connected party did not properly respond after a period of time, or established connection failed because connected host has failed to respond.

#2

It’s a network communication issue somewhere between duplicacy and B2, inclusively.

Does the issue immediately occur on the very first file or few files download successfully and then it fails after a while? Did it ever work on that machine?

Are you using a proxy? Anything unusual about your setup? Are you running windows version right on windows? Any other filters on the environment? Do you have anything like docker, kubernetes, virtual box, hyper-V etc installed on the machine?

What does this refer to? You can restore with 1 thread by not with 3 or 0?

#3

This never happened immediately. It could be GB of data already downloaded before it happened. As mentioned, I have successfully restored a small folder on the same setup without any problems and this is what I alluded to, in that I know the whole setup works by way of confirming the restored files being viewable, i.e. encryption/decryption not being the issue.

Yes, this is running right on Windows 11 bare metal, plain vanilla Windows and w/o proxy to reach the internet.

I am trying to test restore many TBs of data so using 1 thread is not really feasible. I tested with 3 threads and had the same problem. I should add that the internet connection is on a widely used business ISP and I am in a major metro location. That is to say, I am not relying on a flaky ISP.

#4

Thanks. Yes, with Backblaze you are actually expected to run multithreaded downloads, each thread is around 10Mbps. 10, 50, 100 threads should work just fine.

It’s not clear why does not Duplicacy retry on failure. Is this the only failure reported in the log and then Duplicacy stops? I would wait for @gchen’s comments.

Backblaze provides S3 endpoint as well, in addition to B2. I would try to add the same storage via S3 to Duplicacy and restore form there, to see if that is more stable. Implying that there could have been some stability issues with B2 api on their end. In parallel, you can reach out to Backblaze support and ask them about any issues they could have logged for your account on that endpoint.

This is assuming you don’t have any overzealous security software that may be randomly blocking connections. I understand the windows is vanilla, but there could be something on the router.

#5

Yes, this is the only and last failure every time this happened. I am calling on anyone with backup to B2 with a decent backup to test restore their data so that we have more data point on this. Again, Backblaze has changed their policy to allow free download up to 3x your storage size.

I could try to play with the S3 endpoint as you suggested.

And yes, nothing crazy installed on that Windows 11 machine and it’s sitting behind a typically-configured firewall. Like I said, I could download the data just find for a while (up to 60GB once out of 3TB) and then it would encounter the error.

My only guess is that everyone is doing what I am doing and hitting the B2 pipes hard with the downloads since the policy change on 10/3.

1 Like
#6

Ah, they replaced the plan even for existing customers? Wow. I abandoned wasabi when they hiked price three times in 18 months all while degrading reliability. Now b2 goes the same route. Not exactly surprised: with all my negativity towards wasabi they are at least profitable, unlike Backblaze, and never had issues like returning bad data, like b2 did in the past.

On the other hand — the only scenario where this arrangement (fixed amount of free egress) is suitable are in fact backup applications….

If you are looking for a cloud to move to — consider Storj. I’ve been using it for a last year — super happy (as long as your network equipment is more or less decent to use native integration, or if you use their s3 gateway).