Restored Database Size is 0.

[zzerjae]

Hello, first of all, thank you for providing such a great tool. I have been using it well.

Recently, I encountered an issue when trying to restore a database from a remote storage to my local environment. The issue is restored database's size is 0. My setup is as follows:

• ruby 3.3.1
• rails 7.1.3.2
• litestream_ruby 0.10.1
• litestream_rails 0.4.1
• I am using CloudFlare R2 Storage as the remote storage.

The following commands provide normal results:

bin/rails litestream -- --database=storage/development.sqlite3

name  generation        lag     start                 end
s3    7cb3016377add0bf  -3m2s   2024-07-07T12:45:42Z  2024-07-07T12:45:43Z
s3    87a822e9c2b78ddd  -808ms  2024-07-07T12:38:47Z  2024-07-07T12:42:41Z
name  generation        lag     start                 end
s3    7cb3016377add0bf  -3m2s   2024-07-07T12:45:42Z  2024-07-07T12:45:43Z
s3    87a822e9c2b78ddd  -808ms  2024-07-07T12:38:47Z  2024-07-07T12:42:41Z

bin/rails litestream -- --database=storage/development.sqlite3

replica  generation        index  size    created
s3       7cb3016377add0bf  0      532584  2024-07-07T12:45:42Z
s3       87a822e9c2b78ddd  0      532554  2024-07-07T12:38:47Z
replica  generation        index  size    created
s3       7cb3016377add0bf  0      532584  2024-07-07T12:45:42Z
s3       87a822e9c2b78ddd  0      532554  2024-07-07T12:38:47Z

The restore command also works as follows:

bin/rails litestream:restore -- --database=storage/development.sqlite3 -o=storage/development-replica.sqlite3

time=2024-07-07T21:53:46.648+09:00 level=INFO msg="restoring snapshot" db=/Users/me/webapp/storage/development.sqlite3 replica=s3 generation=7cb3016377add0bf index=0 path=storage/development-replica.sqlite3.tmp
time=2024-07-07T21:53:47.274+09:00 level=INFO msg="restoring snapshot" db=/Users/me/webapp/storage/development.sqlite3 replica=s3 generation=7cb3016377add0bf index=0 path=storage/development-replica.sqlite3.tmp
time=2024-07-07T21:53:47.519+09:00 level=INFO msg="restoring wal files" db=/Users/me/webapp/storage/development.sqlite3 replica=s3 generation=7cb3016377add0bf index_min=0 index_max=0
time=2024-07-07T21:53:47.840+09:00 level=INFO msg="downloaded wal" db=/Users/me/webapp/storage/development.sqlite3 replica=s3 generation=7cb3016377add0bf index=0 elapsed=320.626584ms
time=2024-07-07T21:53:47.844+09:00 level=INFO msg="applied wal" db=/Users/me/webapp/storage/development.sqlite3 replica=s3 generation=7cb3016377add0bf index=0 elapsed=4.216583ms
time=2024-07-07T21:53:47.844+09:00 level=INFO msg="renaming database from temporary location" db=/Users/me/webapp/storage/development.sqlite3 replica=s3
time=2024-07-07T21:53:48.129+09:00 level=INFO msg="restoring wal files" db=/Users/me/webapp/storage/development.sqlite3 replica=s3 generation=7cb3016377add0bf index_min=0 index_max=0
time=2024-07-07T21:53:48.477+09:00 level=INFO msg="downloaded wal" db=/Users/me/webapp/storage/development.sqlite3 replica=s3 generation=7cb3016377add0bf index=0 elapsed=347.545584ms
time=2024-07-07T21:53:48.478+09:00 level=INFO msg="applied wal" db=/Users/me/webapp/storage/development.sqlite3 replica=s3 generation=7cb3016377add0bf index=0 elapsed=1.373792ms
time=2024-07-07T21:53:48.478+09:00 level=INFO msg="renaming database from temporary location" db=/Users/me/webapp/storage/development.sqlite3 replica=s3

However, the restored result is 0 in size as follows. What could be the issue?

ls -al storage | grep development

drwxr-xr-x@   4 me  staff      128  7  7 21:45 .development.sqlite3-litestream
-rw-r--r--@   1 me  staff        0  7  7 21:53 development-replica.sqlite3
-rw-------@   1 me  staff    32768  7  7 21:53 development-replica.sqlite3.tmp-shm
-rw-r--r--@   1 me  staff  1785856  7  7 21:42 development.sqlite3
-rw-r--r--@   1 me  staff    32768  7  7 21:38 development.sqlite3-shm
-rw-r--r--@   1 me  staff    37112  7  7 21:42 development.sqlite3-wal

[zzerjae]

update)

If run the litestream directly, get normal results, as shown below:

LITESTREAM_REPLICA_BUCKET=*** LITESTREAM_ACCESS_KEY_ID=*** LITESTREAM_SECRET_ACCESS_KEY=*** litestream restore --config=/Users/me/webapp/config/litestream.yml -o=storage/development-replica.sqlite3 storage/development.sqlite3

time=2024-07-07T22:38:35.679+09:00 level=INFO msg="restoring snapshot" db=/Users/me/webapp/storage/development.sqlite3 replica=s3 generation=7cb3016377add0bf index=0 path=storage/development-replica.sqlite3.tmp
time=2024-07-07T22:38:35.857+09:00 level=INFO msg="restoring wal files" db=/Users/me/webapp/storage/development.sqlite3 replica=s3 generation=7cb3016377add0bf index_min=0 index_max=0
time=2024-07-07T22:38:35.994+09:00 level=INFO msg="downloaded wal" db=/Users/me/webapp/storage/development.sqlite3 replica=s3 generation=7cb3016377add0bf index=0 elapsed=137.237167ms
time=2024-07-07T22:38:36.001+09:00 level=INFO msg="applied wal" db=/Users/me/webapp/storage/development.sqlite3 replica=s3 generation=7cb3016377add0bf index=0 elapsed=6.322208ms
time=2024-07-07T22:38:36.001+09:00 level=INFO msg="renaming database from temporary location" db=/Users/me/webapp/storage/development.sqlite3 replica=s3

ls -al storage | grep development

drwxr-xr-x@   4 me  staff      128  7  7 21:45 .development.sqlite3-litestream
-rw-------@   1 me  staff  1785856  7  7 22:38 development-replica.sqlite3
-rw-------@   1 me  staff    32768  7  7 22:38 development-replica.sqlite3.tmp-shm
-rw-------    1 me  staff        0  7  7 22:38 development-replica.sqlite3.tmp-wal
-rw-r--r--@   1 me  staff  1785856  7  7 22:19 development.sqlite3
-rw-r--r--@   1 me  staff    32768  7  7 21:38 development.sqlite3-shm
-rw-r--r--@   1 me  staff    37112  7  7 21:42 development.sqlite3-wal

[fractaledmind]

This is quite bizarre to me. Do you want to setup a call for us to debug together?