Singlestore6.8 can't start in LXC environment

Help
1

Hello,
thank you for your help。
I build a LXC (Linux containers) environment，use ubuntu20 and centos7 LXC OS deploy memsql 6.8.
at run “memsql-deploy cluster-in-a-box” command failure ：

superright:~# /usr/bin/memsqlctl --json --yes create-node --no-start --port 3306 --bind-address 127.0.0.1 --timeout 60s -vvv
2021/01/28 08:35:22 state_manager.go:97 Acquiring exclusive lock on /var/lib/memsql/nodes.hcl.lock to permit read/write access to /var/lib/memsql/nodes.hcl
2021/01/28 08:35:22 main.go:66 Running command /opt/memsql-server-6.8.24-8e110b7bed/memsqld --defaults-file /var/lib/memsql/1088dca4-b25d-49fc-982c-25385d044ee2/memsql.cnf --user 110 --version
2021/01/28 08:36:22 main.go:66 Command /opt/memsql-server-6.8.24-8e110b7bed/memsqld --defaults-file /var/lib/memsql/1088dca4-b25d-49fc-982c-25385d044ee2/memsql.cnf --user 110 --version failed with exit code -1 in 59.991093972s
Error creating node: memsqld --version failed: : error running command: "/opt/memsql-server-6.8.24-8e110b7bed/memsqld" "--defaults-file" "/var/lib/memsql/1088dca4-b25d-49fc-982c-25385d044ee2/memsql.cnf" "--user" "110" "--version": signal: killed
(run memsql7.3 , same error messages…)

This is my OS and install software information:
superright:~# dpkg -l |grep memsql
ii memsql-client 5.1.17 amd64 MemSQL database client binary
ii memsql-server6.8.24-8e110b7bed 6.8.24 amd64 MemSQL database server binaries

superright:~# lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 20.10
Release: 20.10
Codename: groovy

Has anyone encountered the same problem?

2

@bendchan,

It looks like you are using the cluster deployment guide for Docker cluster in a box, with Linux Containers. Since Docker containers are built on top of LXC this is worth a shot, but be aware it may not work perfectly. It looks like you are doing the online cluster in a box command line method of deployment, is that right?

The cluster deployment is failing on the step when it tries to create a node. Specifically it indicates memsqld --version failed due to the signal being killed.

For an online deployment, you can specify the version in the deployment statement as in the example below. Is this how you are specifying 6.8/7.3?

memsql-deploy cluster-in-a-box --license <license> \
--version 7.3 --password <secure-password>

Additionally, be sure you’ve set port forwarding correctly so the container has access to the internet, otherwise you can download the memsql-server binary separately, and point to it in the command with the flag --file-path <memsql-server-package>. Note that specifying a file path for the local binary should also avoid the check for the current version, which would otherwise require internet access.

For any Toolbox commands, you can use the flag -vvv for a very very verbose output. This is typically useful in debugging, since it includes more details about what parts of the command were accomplished, where the failure occurred, and the error code/message returned from the failure.