Skip to content

QSFS losing connection during operation #19

@mohamedamer453

Description

@mohamedamer453

After connecting to a deployed Vm with a QSFS using terraform provider on testnet.
Using the example from threefoldtech/terraform-provider-grid
with the following config for the QSFS

First Example

   name = "qsfs"
   description = "description6"
   cache = 256 
   minimal_shards = 2
   expected_shards = 4
   redundant_groups = 0
   redundant_nodes = 1
   max_zdb_data_dir_size = 64
   encryption_algorithm = "AES"
   encryption_key = "4d778ba3216e4da4231540c92a55f06157cabba802f9b68fb0f78375d2e825af"
   compression_algorithm = "snappy"

VM Config:

vms {
    name = "vm"
    flist = "https://hub.grid.tf/tf-official-apps/base:latest.flist"
    cpu = 4
    memory = 2048
    entrypoint = "/sbin/zinit init"
    planetary = true
    env_vars = {
      SSH_KEY = "ssh-rsa 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 root@mohamed-Inspiron-3576"
    }
    mounts {
        disk_name = "qsfs"
        mount_point = "/qsfs"
    }
  }

According to TC199 after setting the cache to 256MB and trying to create a larger file with the following command

dd if=/dev/urandom of=tmp.txt bs=1M count=1024

Firstly it gives an i/o error and only creates a 256MB file

dd: error writing 'tmp.txt': I/O error
256+0 records in
255+0 records out

after that using the command df -h to check the disk size or trying any other command on /qsfs it gives a socket not connected error

df: /qsfs: Socket not connected

and after disconnecting from the vm and reconnecting again. still there is no /qsfs

vm:~# cd /qsfs 


-ash: cd: can't cd to /qsfs: Socket not connected

Logs
metrics.txt

**note: after the qsfs disconnected i tried to redeploy using terraform init && terraform apply -parallelism=1 and it also did not work

Second Example

Config is the same as the first example except:

    cache = 1024 
    max_zdb_data_dir_size = 256

VM Config is the same as the first example.

According to TC214 after setting the cache to 1GB and trying to create a 2GB file with the following command

dd if=/dev/urandom of=tmp.txt bs=1M count=2048

It gives an i/o error and only creates a 1GB file

dd: error writing 'tmp.txt': I/O error
1008+0 records in
1007+0 records out

Without disconnecting.

but after trying the same command again to create another file it disconnects during the creation of the file.

dd: error writing 'tmp2.txt': Connection aborted
755+0 records in
754+0 records out

it only created a 750MB file and disconnected during creation giving Connection aborted and after that error trying any command on /qsfs gives Socket not connected

Disconnecting and reconnecting , redeploying also did not work.

Logs
metrics.txt

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions