Enhancements, Fixes, and Known Issues 05/11/2023
The following enhancements, fixes, and known issues are found in this release, dated 05/11/2023:
- Elastio mount was stabilized to avoid IO errors when performing Iscan operations. Also, additional logging was added.
- Certain volumes can be excluded from the EC2 backup now if they are granted a tag with name “elastio:nobackup” and value - “true”. The volumes with this tag would not get backed up, all others attached to the EC2 instance would.
- App-consistent snapshot creation logic was improved to avoid duplication tags and creating both app- and crash- consistent snapshots in rare cases.
- Kernel dump was added to improve logging in filesystem check operations.
- Vault job worker instances list was extended and prioritized. This way the most effective instance types will be tried first, and if those are not available from AWS spot capacity in the region, lower priority instance types from the list will be used.
- Background job worker for PostgreSQL backup / restore was extended to support PostgeSQL versions 10 and newer.
- Background job worker AMIs were upgraded to use latest Linux distros with
- EC2 and EBS backups would sometimes fail due to
VolumeInUseerror during snapshot creation. The backup process was adjusted to avoid this failure mode.
- IO errors would occur in some cases when Iscan was running. This was fixed.
- Elastio mount used to fail with “Failed to mount ‘volume-id’ device due to error: The export with ID ‘blob-id’ already exists”. Mount was hardened against this failure mode.
- Iscan would skip AL2023 root volume and return “Failed to mount device due to error: No such file or directory (os error 2)” error. Filesystem detection was extended to accommodate for the updated AL2023 volume structure.
- On some occasions EBS backup would fail with ‘The snapshot ‘snap-id’ does not exist” if previous backup job failed and did not produce a snapshot. The snapshot cleanup logic was improved to always leave a valid previous snapshot for incremental backups.
- Iscan would fail with “Failed to mount ‘name’ device due to error: No reader implementation has been found for the range” when a block recovery point was scanned. This was fixed.
- PostgreSQL restore hang was removed.
- Iscan of the file recovery point was improved to avoid the “BlkId library failed to probe device” error.
- Previously, if one of the EBS volumes attached to the EC2 was without file system, the snapshots of all the volumes would not have “app-consistent” tag on them even with VSS enabled. This behavior was improved to always tag the snapshots according to the type of snapshot that was taken.
- In rare cases, elastio.exe would crash when updating or uninstalling Elastio CLI. This was fixed.
- EC2 restore in Elastio CLI would sometimes fail trying to add a duplicate tag to the restored entity. This behavior was improved to avoid tag duplication.
- Iscan would silently skip files on Windows volumes without reporting them as unable to scan. This was fixed to make sure all files are scanned, and if we are unable to scan any of them, they are reported back as such.