Flashes (Alerts)
Abstract
An issue has been found which can cause a self-recovering node warmstart. In rare cases, a warmstart can occur on both nodes in an I/O group, leading to a temporary interruption in volume access.
This issue (APAR IT14917) is present in the v7.4.0.9, v7.5.0.7, v7.6.0.4, v7.6.1.1 and v7.6.1.3 PTFs. Earlier PTFs on these code streams are not affected.
Content
This issue is caused by a defect related to a small timing window during the transition from a volume's cache being empty, to the cache being non-empty. Volumes which have a bursty I/O workload, where there are frequent idle periods followed by a burst of I/O, are therefore more likely to experience this issue.
This issue is fixed in the v7.4.0.10, v7.5.0.8 and v7.6.1.4 PTF releases.
Was this topic helpful?
Document Information
Modified date:
28 March 2023
UID
ssg1S1005863