- Knowledge Base
- Configurations
-
Zmanda Enterprise
-
Zmanda Recovery Manager (ZRM)
-
Backend Operations
-
Backup
-
Server Installation
-
Client Configuration
-
Account and Licenses
-
Sources
-
Storage
-
Schedules
-
Backup Sets
-
Backup Operations
-
Restore Operations
-
Monitoring and Reporting
-
Configurations
-
ZMC/UI
-
Whitepaper
-
Zmanda Cloud Storage
-
VMWare
-
Key Vault
-
Disaster Recovery
-
Zmanda Support
-
Product lifecycle
-
Zmanda Pro
Amazon S3 : clock skew
Resolution for restore failures due to misalignment of machine time between Amanda server and Amazon
This article is for Amanda Enterprise (AE).
Issue Type: Tip
Info Description:
Amanda Enterprise server machine time must be in sync with Amanda clients. If the date on the Amanda server is different from the Amanda client, then restore may fail due to a clock skew between the Amanda server and Amazon. You can get the following errors:
While trying to read tapestart header: The difference between the request time
and the current time is too large. (RequestTimeTooSkewed) (HTTP 403)
DEVICE_ERROR
SOLUTION:
The fix is to keep the date and time of Amanda server in sync with Amazon. Amazon's clocks are within a few milliseconds of the global time standard - they use Xen, not Vmware. Configure Amanda Server and Amanda client as ntp clients.