Efcore: Published to Centos7, memory has been growing until the process is killed

[netnr]

Asp.net core 2.1 EF

Centos7 1Core 1G

Previously, Windows was running on IIS, using the sqlserver database. In the first two weeks, it was replaced by Centos, using nginx as a proxy, and the database was replaced by sqlite3.

At the beginning, the Centos system swap space is about 200MB, the site runs for one day, the memory accounts for 90%, the swap runs out, the dotnet process is killed; the swap set 2GB later, persists for about a week, the swap runs out, and the dotnet process is killed.

As long as you visit the site, the memory footprint will gradually increase and will not be released; IIS runs completely without these problems on Windows.

The difference between them is that the database is different, one is sqlserver and the other is sqlite3;

[ajcvickers]

@netnr It's unlikely we will be able to figure out what is happening without a solution/project or code listing that reproduces what you are seeing. Are you able to distill the scenario down into something that we can use to investigate?

[netnr]

After testing, using SQL to query memory without EF will not always increase

[netnr]

https://github.com/netnr/netnrf/tree/master/src/v2/Netnr.ResponseFramework/Netnr.Data

[netnr]

[linkinshi]

@netnr
optionsBuilder.UseLoggerFactory(new LoggerFactory(new[]
{
new ConsoleLoggerProvider((category, level)
=> category == DbLoggerCategory.Database.Command.Name
&& level == LogLevel.Error, true)
}));
don't create new LoggerFactory each time.
you can see this https://docs.microsoft.com/en-us/ef/core/miscellaneous/logging
so，it's you code bug.
国内很多文档在打印日志都这么写，然后上线之后内存暴涨，坑人不浅。

[netnr]

@linkinshi Resolved issue, thanks