Significantly reduced I/O throughput and increased swap use after kernel upgrade
Issue
- A server was upgraded to 2.6.32-358 last Friday. Immediately after the reboot, the load average was around 4 and it climbed slowly over the weekend to about 80. The server became progressively unresponsive and unable to start new processes or accept new connections.
- Reverted to 2.6.32-279 and the issue disappeared.
Environment
- Red Hat Enterprise Linux, version 6
- Kernel >= 2.6.32-358.el6.x86_64
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.