failed to unpack `/var/lib/reth/db/mdbx.dat`

[b0a7]

Describe the bug

I tried 3 times to get the download to work, but each time I got:

2026-04-10T05:26:11.272924Z INFO Downloads complete downloaded=687.93 GB total=597.92 GB elapsed=5h 26m
Error: failed to unpack /var/lib/reth/db/mdbx.dat

Caused by:
0: failed to unpack db/mdbx.dat into /var/lib/reth/db/mdbx.dat
1: request or response body error
2: error reading a body from connection
3: end of file before message length reached

Location:
crates/cli/commands/src/download/mod.rs:1061:13

Possibly related, earlier in the logs I see:
2026-04-10T04:52:41.272424Z INFO Extracting remaining archives archives=687/688 downloaded=637.90 GB
2026-04-10T04:52:44.272495Z INFO Extracting remaining archives archives=687/688 downloaded=638.02 GB
2026-04-10T04:52:44.424387Z WARN Streaming extraction failed, retrying url=https://snapshots-r2.reth.rs/reth-1-archive-stable-24843216-1775755010/state.tar.zst attempt=9 max=10 err=failed to unpack /var/lib/ reth/db/mdbx.dat
2026-04-10T04:52:47.272916Z INFO Extracting remaining archives archives=687/688 downloaded=638.03 GB
2026-04-10T04:52:49.424983Z INFO Retrying streaming download from scratch url=https://snapshots-r2.reth.rs/reth-1-archive-stable-24843216-1775755010/state.tar.zst attempt=10 max=10
2026-04-10T04:52:50.272914Z INFO Extracting remaining archives archives=687/688 downloaded=638.03 GB

Steps to reproduce

reth download --chain mainnet --datadir=/var/lib/reth --storage.v2 --full

Node logs

Platform(s)

Linux (x86)

Container Type

Not running in a container

What version/commit are you on?

2.0.0

What database version are you on?

v2

Which chain / network are you on?

mainnet

What type of node are you running?

Full via --full flag

What prune config do you use, if any?

No response

If you've built Reth from source, provide the full command you used

No response

Code of Conduct

• I agree to follow the Code of Conduct

[yorickdowne]

Try with --resumable, that should do it. This isn't a bug, just a missing parameter.

[b0a7]

that is possible. I found the release notes and documentation about upgrading to v2 storage to be quite lacking to be honest. It took some digging through docs just to find out that a resync was needed to migrate. On the v2 storage page that param is not included

https://reth.rs/run/faq/storage-v2#storage-v2-experimental

perhaps resumable should be the default?

[mark0-cn]

#23451 Merged

[yorickdowne]

Qualifier: Maybe it's a bug? State consistently gets interrupted, in every single test I see something like:

2026-04-11T06:58:50.300856Z  INFO Download interrupted, retrying in 5s... file=state.tar.zst
2026-04-11T06:58:52.753043Z  INFO Downloading archives=1185/1186 progress=95.1% downloaded=746.89 GB total=785.75 GB eta=2m 55s
2026-04-11T06:58:55.301024Z  INFO Retry attempt 3/10 - resuming from 56413257728 bytes file=state.tar.zst

--resumable resolves that. And, it's odd that I haven't ever seen a success without a resume. Maybe there's a timeout in the download code somewhere, for large files?

[Rjected]

Resumable is now the default and we just merged #23028 which prevents us from requesting large files from r2 to begin with (assuming this is the root of these failures and retries). If you would like, please try it out, otherwise this may be fixed in the next release.