Docker container running SQL Server exits on start

Question

I'm running an instance of SQL Server in a Docker container. It exits on start automatically. The logs say:

SQL Server 2019 will run as non-root by default.
This container is running as user mssql.
Your master database file is owned by mssql.
To learn more visit https://go.microsoft.com/fwlink/?linkid=2099216.
sqlservr: Unable to read instance id from /var/opt/mssql/.system/instance_id: File: pal.cpp:438 [Status: 0xC0000034 Object name not found errno = 0x2(2) No such file or directory]
/opt/mssql/bin/sqlservr: PAL initialization failed. Error: 101

If I am reading this right SQL Server will not run with root user. But then both my container and volume are running as the same non-root user. So I'm confused where the problem is coming from. Can someone help me decode this error message?

Answer 1

I have got same this error and I resolved it by way remove the file /var/opt/mssql/.system/instance_id and then restart service again

Answer 2

UPD 2025:

I once again had this issue, and its due to 0 free space on disk. In this case you delete or update instance_id but after container start the issue will repeat itself and instnce_id file will remain empty, and SQL wont start ☝️

2023:

I had this exact issue today. It happened after I resized disk partition somehow.

At first I inspected docker volume and data seemed to be ok, however

/var/opt/mssql/.system/instance_id

wich is

/var/lib/docker/volumes/[YOUR-VOLUME]/_data/.system/instance_id

was empty!

Then I googled this thread https://github.com/microsoft/mssql-docker/issues/615 suggested to run chown 10001:0 on _data folder wich did not help.

Fortunately my docker is running inside Hyper-V vm. So I created checkpoint and reverted to previous one to see whats inside instance_id file

It turned out to be guid and 20 digit number

dfb10ba0-dc20-4022-ab31-0f52be4dcabe
16537894658124376490

I copied that and returned to latest checkpoint. Stopped container, updated instance_id with recovered data and container started seamlessly.

SQL Server on Linux stores the sequential UUID seed in /var/opt/mssql/.system/instance_id and increments it during startup. Take a backup of the instance_id file in case of system failure. If the file is lost, the seed is missing, and the new seed is regenerated. The initial seed generation is based on a random bit pattern and a UUID to avoid collisions. However, the new seed that has to be sequentially ordered may not be sequentially ordered after the seed is lost.

https://support.microsoft.com/en-us/topic/kb4078097-fix-newsequentialid-function-generates-duplicate-guid-after-sql-server-2017-on-linux-is-restarted-829b97f7-6359-8f66-6581-37aac819c210