PowerStore: Node reboot due to rasdaemon memory leak (OOM)

Symptoms may include:

• A single node reboot (the node automatically recovers after the reboot).
• In rare cases, both nodes may be affected by this issue simultaneously, resulting in a short service disruption.

This issue only affects PowerStore appliances running releases older than PowerStoreOS 3.2.

---

PowerStoreOS includes a process called rasdaemon which is responsible for monitoring and reporting various CPU errors to the upper layers. The rasdaemon utility comes bundled with the PowerStoreOS. An issue has been identified that may cause a memory leak when an underlying hardware issue exists. These underlying hardware issues are typically correctable errors on DIMMs and CPU. Every time an issue is encountered the process may leak memory until no memory is left.

---

This issue is fixed in PowerStoreOS 3.2.0 and later.

To determine if this issue has occurred, contact Dell Technical Support or your authorized service provider and reference this article.

---