Skip to content

Latest commit

 

History

History
135 lines (100 loc) · 3.73 KB

README.md

File metadata and controls

135 lines (100 loc) · 3.73 KB
Raw

Prerequisites

1. qemu-kvm installed:

apt-get install qemu-kvm

2. user is added into kvm group

usermod -a -G kvm $USER

How to build and run

1. Building

To build run the following command from the repository root folder:

./ya make -r -- cloud/filestore/buildall

2. Configuring

  • prepare bin directory
cd cloud/filestore/bin
./setup.sh

3. Running services

  • to avoid dynamic link errors you need to set LD_LIBRARY_PATH to include all necessary libraries:
export REPO_ROOT=<path-to-repo-root>

export FILESTORE_APPS_PATH=cloud/filestore/buildall/cloud/filestore/apps
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$REPO_ROOT/$FILESTORE_APPS_PATH/client
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$REPO_ROOT/$FILESTORE_APPS_PATH/server
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:$REPO_ROOT/$FILESTORE_APPS_PATH/vhost
  • use initctl.sh to format and initialize kikimr service
./initctl.sh format initialize
  • then you should create filestore and can start vhost endpoint
./initctl.sh create startendpoint

4. Accessing filestore

  • After creating endpoint you can manually run qemu from the test suite in interactive mode
./run_test_qemu.sh

#Login/Password: qemu/qemupass
qemu@test> mkdir mnt
qemu@test> sudo mount -t virtiofs fs0 mnt    # creates mount point for the filestore
  • or you can mount filestore localy
./initctl.sh mount

Thanks for flying NFS

Windows virtiofs

1. Preparing windows

  • Initial instruction is here

Please note that upon installing virtio drivers you should select only virtiofs or you will render vm unusable after restart

2. Preparing environment

  • Follow steps 1..3 from build and run
  • Either make your own image or prepare existing one
ya make cloud/storage/core/tools/testing/qemu/win-image

3. Prepare and start windows

  • After creating endpoint you can manually run qemu w windows
./run_win_qemu.sh
  • To connect you can use default RDP port 3389 via Administrator/qemupass!77

If you are running qemu inside dev machine in the cloud with limited external ports you'll need:

  • to proxy via any open port(22,80,443)
  • to proxy ipv6 to ipv4 net
sudo socat TCP6-LISTEN:80,fork TCP4:127.0.0.1:3389

4. Accessing filestore

  • By design there should be service mounting filestore starting from drive Z:

  • Or you can manually run

C:\Program Files\Virtio-Win\VioFS>virtiofs.exe -D - -d -1

5. Bugs and limitations

  • So far it supports only one virtiofs drive with no way to specify specific tag
  • It doesn't work: NBS-3031

initctl help

NOTE: you can chain commands but they are order sensitive.

./initctl.sh stop           # stop all running services
./initctl.sh kill           # same but SIGKILL
./initctl.sh format         # format kikimr pdisk

./initctl.sh start          # launches kikimr itself, kikimr storage service & vhost
./initctl.sh initialize     # same as above but initializes kikimr, required at first run and after format cmd
./initctl.sh startlocal     # launches passthrough to local fs storage service & vhost
./initctl.sh startnull      # launches dummy storage service & vhost

./initctl.sh create         # creates filestore, required for local/kikimr services at first run or after format
./initctl.sh mount          # mounts filestore via fuse driver usually at ~/nfs unless configuration changed
./initctl.sh startendpoint  # creates virtiofs endpoint sock usually at /tmp/vhost.sock
./initctl.sh stopendpoint   # stops endpoint

NOTES

  • due to multiple grpc issues under thread sanitizer you should run them with TSAN_OPTIONS='report_atomic_races=0' ya make ...