Node Crashes in Public Cloud / Reduced performance
Details
| Detail name | Value |
|---|---|
| Changelog Number | 9416 |
| Type | Bug |
| Status | Resolved |
| Affected Versions | Exasol 6.1.0, Exasol 6.2.0, Exasol 7.0.alpha1 |
| Fix Versions | Exasol 7.0.0, Exasol 6.1.9, Exasol 6.2.5 |
| Resolution Date | 2020-03-30 |
Issue
Due to an oversight, the transparent hugepage Linux kernel feature was mistakenly enabled in installations for the affected releases even though Exasol performs best with this feature disabled. In most cases this does not have adverse side-effects but can — depending on the workload and platform — cause the following problems:
- Depending on the specific workload, this can reduce system performance. Especially when no hugepages are explicitly configured. In extreme cases, this can lead to a database restart when the performance of one nodes degrades enough for it to be considered failed by the rest of the cluster
- Especially on the Azure cloud platform, this Issue has been identified as the cause for individual cluster nodes crashing due to a bug specific to this platform being triggered more often with this setting active.
Workaround
No workaround that can be implemented completely independently is available at this point but Exasol Support can provide affected customers with a hotfix that will correct this problem on affected systems until full update packages are released for the affected versions.
