Hi. Can you tell me how I can set up Bucket Explorer so that I can see the S3 buckets as a network or local computer drive? I use Vista as my OS.

Thanks

Jyoti

Jyoti,

We do not have plans to provide any support within Bucket Explorer to map a "virtual drive". In our evaluation and feedback from more than 300 customers that participated in our focused group, our users don't want to map a drive from S3 to a PC and use it as the primary storage media, the reasons given by our customers and us are very similar to the one posted by Mitch in this post at Amazon S3 forums http://developer.amazonwebservices.com/ ... 142&#65142 .

Could you explain the use case and may be we can recommend an approach to resolve that problem using Bucket Explorer or the new command line folder sync tool that we are working on?

In our view, its better to write locally and then sync to S3 in "near real time". We do plan to provide tools to do that in next few weeks. You can use the "subst" command to map any folder as a drive if you must use a drive letter.

Again, let us know the business requirement, and we will sure try to provide a solution as part of Bucket Explorer or the command line tool.

Saurabh

Hi Saurabh,

I do store key files locally on a storage drive but I would like to replicate the local storage drive on S3, as an on-line back-up. As far as I can tell, with Bucket Explorer, I would need to move the data manually from the local drive to S3. Ideally, I would like that to be automated. Can you tell me how to do that? I presume the sync tool you mention might help achieve this?

Note that in addition to synched folders on a storage drive, I also create back-ups on the storage drive using Memeo. I presume I can do this with S3 with or without Bucket Explorer. However, what I find compelling (as a non-technical person) is to use S3 directly rather than via back-up software. Is this a reasonable wish or would I be better off just using back-up software to populate S3?

Thanks

Jyoti

Thanks

Jyoti

Saurabh Dani wrote:Jyoti,

We do not have plans to provide any support within Bucket Explorer to map a "virtual drive". In our evaluation and feedback from more than 300 customers that participated in our focused group, our users don't want to map a drive from S3 to a PC and use it as the primary storage media, the reasons given by our customers and us are very similar to the one posted by Mitch in this post at Amazon S3 forums http://developer.amazonwebservices.com/ ... 142&#65142 .

Could you explain the use case and may be we can recommend an approach to resolve that problem using Bucket Explorer or the new command line folder sync tool that we are working on?

In our view, its better to write locally and then sync to S3 in "near real time". We do plan to provide tools to do that in next few weeks. You can use the "subst" command to map any folder as a drive if you must use a drive letter.

Again, let us know the business requirement, and we will sure try to provide a solution as part of Bucket Explorer or the command line tool.

Saurabh

Jyoti,

With our command line / sync interface you will be able to do what you have asked above. I you want to use S3 "directly", I would suggest to use it via API and make sure that every file which you are trying to write there is 100% uploaded and you get a commit back from S3. For anything else, its best to write locally and use a sync option.

A virtual drive where all writes from every applications are written directly to S3 is not possible unless the machine is hosted in Amazon's data center (EC2). Most applications which provide a "virtual disk" or "extended disk" first copy the data locally in a cache and then sync it with S3 in background. This gives you an illusion that all of the data is stored on S3, which may or may not be true if the background sync did not complete.

So, the sync tool which we plan to provide will clearly tell the user which files are replicated to S3 and which ones are only on local.

Thanks
Saurabh

Hi Saurabh,

When do you expect your command line / sync interface to be available? Thanks.

Jyoti

Before end of October.

We have added this to our roadmap:

http://www.bucketexplorer.com/forum/topic158.html

I'm actually a big fan of S3drive for this purpose. The only reason I switched from S3drive to BucketExplorer is their strange file naming scheme and lack of ACL command support.

Especially for windows users, have a mapped drive means that accessing s3's contents is literally a click away - I don't need to open another program, wait for it to open, and then use whatever interface its designers thought would be best. I just drag over my files and that is that.

Drive mapping is a bit of an advanced topic, in so far as people don't generally know what it is enough to know they want it - but if you did support this feature, and people saw their s3 drive on their computer appear just like any other hard drive, I'm certain many of those people you polled would find it extremely helpful.

If you do another focus group, you might consider asking people directly if they would like their s3 buckets to appear on their computer as drives - I'll bet if you ask that way many people would be interested.

In any case, I think this would be a useful feature and would place you ahead of the game in terms of s3 bucket managers.

My use case for this feature:

I use a program called Backup for Workgroups to backup multiple computers to a single server. For various disaster recovery reasons, I'd love to make an off site mirror of the repository. Amazon S3 seems like a good and cost effective way to do that.

Backup for Workgroups supports repository mirroring to a local drive letter. If I can map an Amazon S3 account to a drive letter on my machine, setting up the mirroring is as simple as point and click from within Backup for Workgroups.

I do not fully understand how Backup for Workgroups modifies the repository with each backup, however I would suspect using their built in sync tool is preferable to an outside tool.

Does Bucket Explorer offer an alternate method to accomplish this sort of sync? And how can it be automated?