Chris Lu

icon indicating a website link https://github.com/chrislusf/seaweedfs

icon location San Francisco icon location https://github.com/chrislusf/seaweedfs SeaweedFS the distributed file system and object store for billions of small files ...

Results 785 comments of Chris Lu

data loss - filer IO Error; 500 Internal Server Error when reading files after some time - needle_read.go:58 entry not found1

I do not know now. Could you use `see_dat` and `see_idx` under `https://github.com/seaweedfs/seaweedfs/tree/master/unmaintained` to see what is inside the `.dat` and `.idx` files? Or you can upload these two files...

data loss - filer IO Error; 500 Internal Server Error when reading files after some time - needle_read.go:58 entry not found1

does `volume 573` actually exist in `volume.list` output? can you remove `volume.deleteEmpty -quietFor=24h -force` from scheduled jobs?

data loss - filer IO Error; 500 Internal Server Error when reading files after some time - needle_read.go:58 entry not found1

re-upload should work well.

data loss - filer IO Error; 500 Internal Server Error when reading files after some time - needle_read.go:58 entry not found1

can you try to ec decode the volume 29 and see whether it helps? btw: `.ecj` contains the deleted file ids, and can be different on different volume servers since...

data loss - filer IO Error; 500 Internal Server Error when reading files after some time - needle_read.go:58 entry not found1

> i am mostly using it via the filer web api, typically via POST to "http://10.10.200.201:8888/example.jpg?mode=0666&uid=33&gid=33". Typically only write once and then read a few times. The mount is mostly...

the size becomes double after ec.decode

> the size from s3.bucket.list is same as the size in real. EC would not possibly achieve this. You can only use replication `000` and vacuum.

no more space for volume, even though disks are only at 80% full and still have > 1T free

@wzrdtales what is the version? Try to upgrade.

no more space for volume, even though disks are only at 80% full and still have > 1T free

@kmod-midori Thanks for the pointer! 4ec6eab004ff500b849f10f1985e83ce0855c70c this commit may fix this problem due to not including EC shards.

MongoDB container with data on SeaweedFS mount hangs

anyway to reproduce the workload? With some scripts, assuming a "weed mount" is running. Or use docker compose?

MongoDB container with data on SeaweedFS mount hangs

It would be good to see some logs. I am testing this way: * use "weed mount" to mount a folder, and run `docker run -d -p 27017:27017 -v ~/tmp/mounted_directory:/data/db...