Use S3 for image upload/storage

[eprothro]

If we roll our own, we'll need an S3 directory for us to use for paper-cliping (e.g. not blocking our server during upload) images and files?

[eprothro]

@tubaman, would you mind creating protege-production on S3 for us? I assumed you have the credentials based on our CDN convo this AM :).

[eprothro]

@tubaman Also, FYI, this repo is public, so if you can email me credentials instead of posting them here, that would be lovely. They'll only live on the production server as ENV vars.

[tubaman]

What problem are we trying to solve? What do you mean blocking the
server? I'm not sure about the web server that you're running but on
Apache, each request get's it's own thread. Usually upload requests
are more long-lived than download requests but they're not very frequent
and shouldn't block other stuff.

All that being said, I can certainly create an S3 bucket if that's what
you want.

• Ryan

On Mon, Aug 13, 2012 at 04:01:09PM -0700, eprothro wrote:

Hey @tubaman, could you create an S3 directory for us to use for papercliping (e.g. not blocking our server during upload) images and files? I assume you have the credentials to do so based on our convo this AM :).

---

Reply to this email directly or view it on GitHub:
#2

[eprothro]

There are other reasons for wanting to use S3 than performance (e.g. not using a lot of the free database to store images and files, content assets coming from S3 instead of more requests to the free app), but also the Heroku infrastructure isn't inherently multi-threaded the same way it would be if we were running on a dedicated server/VM. If it's not too much trouble/cost, an S3 bucket would be great. Happy to discuss if you want.

Thanks,

Evan Prothro
YouVersion Web // evan.prothro (Skype/gTalk)

On Aug 14, 2012, at 11:35 AM, Ryan Nowakowski wrote:

What problem are we trying to solve? What do you mean blocking the
server? I'm not sure about the web server that you're running but on
Apache, each request get's it's own thread. Usually upload requests
are more long-lived than download requests but they're not very frequent
and shouldn't block other stuff.

All that being said, I can certainly create an S3 bucket if that's what
you want.

• Ryan

On Mon, Aug 13, 2012 at 04:01:09PM -0700, eprothro wrote:

Hey @tubaman, could you create an S3 directory for us to use for papercliping (e.g. not blocking our server during upload) images and files? I assume you have the credentials to do so based on our convo this AM :).

---

Reply to this email directly or view it on GitHub:
#2

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/2#issuecomment-7731602.

[tubaman]

Makes sense. Do we want a new AWS account for this S3 bucket? Use the
current ACF AWS account?

• Ryan

On Tue, Aug 14, 2012 at 10:12:49AM -0700, eprothro wrote:

There are other reasons for wanting to use S3 than performance (e.g. not using a lot of the free database to store images and files, content assets coming from S3 instead of more requests to the free app), but also the Heroku infrastructure isn't inherently multi-threaded the same way it would be if we were running on a dedicated server/VM. If it's not too much trouble/cost, an S3 bucket would be great. Happy to discuss if you want.

Thanks,

Evan Prothro
YouVersion Web // evan.prothro (Skype/gTalk)

On Aug 14, 2012, at 11:35 AM, Ryan Nowakowski wrote:

What problem are we trying to solve? What do you mean blocking the
server? I'm not sure about the web server that you're running but on
Apache, each request get's it's own thread. Usually upload requests
are more long-lived than download requests but they're not very frequent
and shouldn't block other stuff.

All that being said, I can certainly create an S3 bucket if that's what
you want.

• Ryan

On Mon, Aug 13, 2012 at 04:01:09PM -0700, eprothro wrote:

Hey @tubaman, could you create an S3 directory for us to use for papercliping (e.g. not blocking our server during upload) images and files? I assume you have the credentials to do so based on our convo this AM :).

---

Reply to this email directly or view it on GitHub:
#2

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/2#issuecomment-7731602.

---

Reply to this email directly or view it on GitHub:
#2 (comment)

[eprothro]

I'm fine with the existing if you are.

Evan Prothro
YouVersion Web // evan.prothro (Skype/gTalk)

On Aug 14, 2012, at 1:38 PM, Ryan Nowakowski wrote:

Makes sense. Do we want a new AWS account for this S3 bucket? Use the
current ACF AWS account?

• Ryan

On Tue, Aug 14, 2012 at 10:12:49AM -0700, eprothro wrote:

There are other reasons for wanting to use S3 than performance (e.g. not using a lot of the free database to store images and files, content assets coming from S3 instead of more requests to the free app), but also the Heroku infrastructure isn't inherently multi-threaded the same way it would be if we were running on a dedicated server/VM. If it's not too much trouble/cost, an S3 bucket would be great. Happy to discuss if you want.

Thanks,

Evan Prothro
YouVersion Web // evan.prothro (Skype/gTalk)

On Aug 14, 2012, at 11:35 AM, Ryan Nowakowski wrote:

What problem are we trying to solve? What do you mean blocking the
server? I'm not sure about the web server that you're running but on
Apache, each request get's it's own thread. Usually upload requests
are more long-lived than download requests but they're not very frequent
and shouldn't block other stuff.

All that being said, I can certainly create an S3 bucket if that's what
you want.

• Ryan

On Mon, Aug 13, 2012 at 04:01:09PM -0700, eprothro wrote:

Hey @tubaman, could you create an S3 directory for us to use for papercliping (e.g. not blocking our server during upload) images and files? I assume you have the credentials to do so based on our convo this AM :).

---

Reply to this email directly or view it on GitHub:
#2

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/2#issuecomment-7731602.

---

Reply to this email directly or view it on GitHub:
#2 (comment)

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/2#issuecomment-7735762.