Indicate Glacier restore in progress

[cyberduck]

d8d6957 created the issue

It would be nice to have some indication where in the restore process a particular file is at. Thus when getting information on a S3 object, has a restore request been placed, has it now been restored and is ready for download? Even just a history of restore requests would be nice. As it is, when I place a restore request via CyberDuck I never know when it is ready for download other than to try to download it. With a few objects, this is ok but with dozens it's quite tedious. Thx!

[dirkpetersen]

I agree, this is quite confusing right now, as a first step it would be great to change the error message when you try to download an object for which a restore request has already been made, something along the lines of

        try {
            // Attempt to get the object
            s3Client.getObject(BUCKET_NAME, OBJECT_KEY);
        } catch (AmazonServiceException e) {
            // Check if it's a 403 error and the operation is not valid for the object's storage class
            if (e.getStatusCode() == 403 && e.getErrorCode().equals("InvalidObjectState")) {
                LocalDateTime restoreInitiationTime = restoreInitiationTimes.get(OBJECT_KEY);
                
                if (restoreInitiationTime == null) {
                    // Initiate a restore request if one has not been initiated before
                    RestoreObjectRequest restoreObjectRequest = new RestoreObjectRequest(BUCKET_NAME, OBJECT_KEY)
                            .withExpirationInDays(7) // Keep the restored data for 7 days
                            .withGlacierJobParameters(new RestoreObjectRequest.GlacierJobParameters()
                                    .withTier(RestoreObjectRequest.Tier.Bulk)); // Initiate a bulk retrieval job

                    s3Client.restoreObject(restoreObjectRequest);
                    
                    // Store the current time as the initiation time for the restore
                    restoreInitiationTimes.put(OBJECT_KEY, LocalDateTime.now());

                    System.out.println("A restore request has been triggered. Please come back in 3-10 hours when the data will be accessible.");
                } else {
                    // Calculate the time passed since the initiation
                    long hoursPassed = java.time.Duration.between(restoreInitiationTime, LocalDateTime.now()).toHours();
                    long hoursRemaining = Math.max(3, 12 - hoursPassed); // Assume max time is 12 hours

                    // Format the initiation time
                    String formattedTime = restoreInitiationTime.format(DateTimeFormatter.ofPattern("dd MMM yyyy 'at' HH:mm"));

                    System.out.println("A restore request was already triggered on " + formattedTime + 
                        ". You may have to wait another " + hoursRemaining + " hours until the object is available for download.");
                }
            } else {
                // Other exceptions can be handled here
                e.printStackTrace();
            }
        }

Also when requesting it is not clear what type of request was made. This should default to "Bulk" -> the cheapest.