Storage I/O could remain blocked for up to the configured timeout - 30 minutes by default - after an inactive SDDC node failure
Details
| Detail name | Value |
|---|---|
| Changelog Number | 33027 |
| Type | Bug |
| Status | Open |
| Affected Versions | Exasol 8.0.0 |
Description
In SDDC deployments (see SDDC: Overview), volume redundancy segments are placed on inactive nodes. If such a node fails while a storage operation is in progress, the operation could continue waiting for a response from the failed redundancy segment, even after the node failure was detected. The issue was timing- and operation-dependent. After the timeout, the database could remain online or the affected database process could restart automatically.
Workaround
If the problem occurs frequently, contact Exasol Support to discuss adjusting the storage timeout.
Fix
Fixed handling of in-flight storage operations after an inactive SDDC node failure so they no longer wait unnecessarily for the full storage timeout.