Backups to cloud are susceptible to transient cloud provider errors
Details
| Detail name | Value |
|---|---|
| Changelog Number | 9439 |
| Type | Bug |
| Status | Resolved |
| Affected Versions | Exasol 7.1.0, Exasol 8.0.0 |
| Fix Versions | Exasol 8.29.0 |
| Resolution Date | 2024-07-12 |
Issue
Backups to cloud based remote archive volumes are susceptible to transient errors within the corresponding cloud provider's infrastructure. (e.g. HTTP 5xx error codes)
While Exasol implements a retry mechanism to for cases in which the cloud infrastructure is momentarily unable to accept the backup, this can retry the operation too quickly. The backup can then fail if the cloud infrastructure is still unavailable.
Fix
We will add a mechanism to wait longer before retrying the operation. This decreases the likelihood that the backup will fail outright.
Workaround
These errors are generally transient, so performing another backup should be successful.
In the event that the error persists, it is likely that the cloud provider is always under heavy load when the backup is scheduled. In this case, using a different bucket or slightly moving the backup schedule can help.
