The OOM Killer Cometh

Kevin Feasel

2017-10-18

Linux

Venu Cherukupalli shows how to keep Linux’s Out of Memory Killer from taking down SQL Server:

When an index rebuild was kicked off on a large table (around 25GB), the reindex operation terminated, and the availability group had failed over to the other replica.

Upon further investigation, we discovered that the SQL Server process terminated at the time reindex operation was run and this resulted in the failover.

To determine the reason for the unexpected shutdown, we reviewed the Linux System Logs (/var/log/messages on RHEL) & pacemaker logs. From the pacemaker logs and system logs, we saw entries indicating that oom-killer was invoked, and as a result SQL Server process was terminated.

Read on for the two solutions.  I was hoping for a solution that involved making the SQL Server executable immune from oom-killer’s wily ways, but not so much in this post.

Related Posts

Building A Kubernetes Cluster With Kubespray

Chris Adkin continues a series on Kubernetes clusters: In essence Kubespray is a bunch of Ansible playbooks; yaml file that specify what actions should take place against one or more machines specified in a hosts.ini file, this resides in what is known as an inventory. Of all the infrastructure as code tools available at the time […]

Read More

Attaching Databases Via Dockerfile

Andrew Pruski shares a better technique for attaching database files held outside of a Docker container: Now this works a treat. It waits ten seconds for the SQL instance to come up within the container and then runs the sqlcmd script below.The problem with this is, it’s a bit of a hack. The HEALTHCHECK command isn’t designed […]

Read More

Categories

October 2017
MTWTFSS
« Sep Nov »
 1
2345678
9101112131415
16171819202122
23242526272829
3031