zulip/docs/production/upload-backends.md

244 lines
9.3 KiB
Markdown
Raw Normal View History

# File upload backends
Zulip in production supports a couple different backends for storing
files uploaded by users of the Zulip server (messages, profile
pictures, organization icons, custom emoji, etc.).
The default is the `LOCAL_UPLOADS_DIR` backend, which just stores
files on disk in the specified directory on the Zulip server.
Obviously, this backend doesn't work with multiple Zulip servers and
doesn't scale, but it's great for getting a Zulip server up and
running quickly. You can later migrate the uploads to S3 by
[following the instructions here](#migrating-from-local-uploads-to-amazon-s3-backend).
We also support an `S3` backend, which uses the Python `boto` library
to upload files to Amazon S3 (or an S3-compatible block storage
provider supported by the `boto` library).
## S3 backend configuration
Here, we document the process for configuring Zulip's S3 file upload
backend. To enable this backend, you need to do the following:
1. In the AWS management console, create a new IAM account (aka API
user) for your Zulip server, and two buckets in S3, one for uploaded
files included in messages, and another for user avatars. You need
two buckets because the "user avatars" bucket is generally configured
as world-readable, whereas the "uploaded files" one is not.
1. Set `s3_key` and `s3_secret_key` in /etc/zulip/zulip-secrets.conf
to be the S3 access and secret keys for the IAM account.
Alternately, if your Zulip server runs on an EC2 instance, set the
IAM role for the EC2 instance to the role.
1. Set the `S3_AUTH_UPLOADS_BUCKET` and `S3_AVATAR_BUCKET` settings in
`/etc/zulip/settings.py` to be the names of the S3 buckets you
created (e.g. `"exampleinc-zulip-uploads"`).
1. Comment out the `LOCAL_UPLOADS_DIR` setting in
`/etc/zulip/settings.py` (add a `#` at the start of the line).
1. If you are using a non-AWS block storage provider,
you need to set the `S3_ENDPOINT_URL` setting to your
endpoint url (e.g. `"https://s3.eu-central-1.amazonaws.com"`).
For certain AWS regions, you may need to set the `S3_REGION`
setting to your default AWS region's code (e.g. `"eu-central-1"`).
1. Finally, restart the Zulip server so that your settings changes
take effect
(`/home/zulip/deployments/current/scripts/restart-server`).
It's simplest to just do this configuration when setting up your Zulip
server for production usage. Note that if you had any existing
uploading files, this process does not upload them to Amazon S3; see
[migration instructions](#migrating-from-local-uploads-to-amazon-s3-backend)
below for those steps.
uploads: Serve S3 uploads directly from nginx. When file uploads are stored in S3, this means that Zulip serves as a 302 to S3. Because browsers do not cache redirects, this means that no image contents can be cached -- and upon every page load or reload, every recently-posted image must be re-fetched. This incurs extra load on the Zulip server, as well as potentially excessive bandwidth usage from S3, and on the client's connection. Switch to fetching the content from S3 in nginx, and serving the content from nginx. These have `Cache-control: private, immutable` headers set on the response, allowing browsers to cache them locally. Because nginx fetching from S3 can be slow, and requests for uploads will generally be bunched around when a message containing them are first posted, we instruct nginx to cache the contents locally. This is safe because uploaded file contents are immutable; access control is still mediated by Django. The nginx cache key is the URL without query parameters, as those parameters include a time-limited signed authentication parameter which lets nginx fetch the non-public file. This adds a number of nginx-level configuration parameters to control the caching which nginx performs, including the amount of in-memory index for he cache, the maximum storage of the cache on disk, and how long data is retained in the cache. The currently-chosen figures are reasonable for small to medium deployments. The most notable effect of this change is in allowing browsers to cache uploaded image content; however, while there will be many fewer requests, it also has an improvement on request latency. The following tests were done with a non-AWS client in SFO, a server and S3 storage in us-east-1, and with 100 requests after 10 requests of warm-up (to fill the nginx cache). The mean and standard deviation are shown. | | Redirect to S3 | Caching proxy, hot | Caching proxy, cold | | ----------------- | ------------------- | ------------------- | ------------------- | | Time in Django | 263.0 ms ± 28.3 ms | 258.0 ms ± 12.3 ms | 258.0 ms ± 12.3 ms | | Small file (842b) | 586.1 ms ± 21.1 ms | 266.1 ms ± 67.4 ms | 288.6 ms ± 17.7 ms | | Large file (660k) | 959.6 ms ± 137.9 ms | 609.5 ms ± 13.0 ms | 648.1 ms ± 43.2 ms | The hot-cache performance is faster for both large and small files, since it saves the client the time having to make a second request to a separate host. This performance improvement remains at least 100ms even if the client is on the same coast as the server. Cold nginx caches are only slightly slower than hot caches, because VPC access to S3 endpoints is extremely fast (assuming it is in the same region as the host), and nginx can pool connections to S3 and reuse them. However, all of the 648ms taken to serve a cold-cache large file is occupied in nginx, as opposed to the only 263ms which was spent in nginx when using redirects to S3. This means that to overall spend less time responding to uploaded-file requests in nginx, clients will need to find files in their local cache, and skip making an uploaded-file request, at least 60% of the time. Modeling shows a reduction in the number of client requests by about 70% - 80%. The `Content-Disposition` header logic can now also be entirely shared with the local-file codepath, as can the `url_only` path used by mobile clients. While we could provide the direct-to-S3 temporary signed URL to mobile clients, we choose to provide the served-from-Zulip signed URL, to better control caching headers on it, and greater consistency. In doing so, we adjust the salt used for the URL; since these URLs are only valid for 60s, the effect of this salt change is minimal.
2022-11-22 20:41:35 +01:00
## S3 local caching
For performance reasons, Zulip stores a cache of recently served user
uploads on disk locally, even though the durable storage is kept in
S3. There are a number of parameters which control the size and usage
of this cache, which is maintained by nginx:
- `s3_memory_cache_size` controls the in-memory size of the cache
_index_; the default is 1MB, which is enough to store about 8 thousand
entries.
- `s3_disk_cache_size` controls the on-disk size of the cache
_contents_; the default is 200MB.
- `s3_cache_inactive_time` controls the longest amount of time an
entry will be cached since last use; the default is 30 days. Since
the contents of the cache are immutable, this serves only as a
potential additional limit on the size of the contents on disk;
`s3_disk_cache_size` is expected to be the primary control for cache
sizing.
These defaults are likely sufficient for small-to-medium deployments.
Large deployments, or deployments with image-heavy use cases, will
want to increase `s3_disk_cache_size`, potentially to be several
gigabytes. `s3_memory_cache_size` should potentially be increased,
based on estimating the number of files that the larger disk cache
will hold.
You may also wish to increase the cache sizes if the S3 storage (or
S3-compatible equivalent) is not closely located to your Zulip server,
as cache misses will be more expensive.
## nginx DNS nameserver configuration
The S3 cache described above is maintained by nginx. nginx's configuration
requires an explicitly-set DNS nameserver to resolve the hostname of the S3
servers; Zulip defaults this value to the first nameserver found in
`/etc/resolv.conf`, but this resolver can be [adjusted in
`/etc/zulip/zulip.conf`][s3-resolver] if needed. If you adjust this value, you
will need to run `/home/zulip/deployments/current/scripts/zulip-puppet-apply` to
update the nginx configuration for the new value.
[s3-resolver]: system-configuration.md#nameserver
## S3 bucket policy
The best way to do the S3 integration with Amazon is to create a new IAM user
just for your Zulip server with limited permissions. For both the user uploads
bucket and the user avatars bucket, you'll need to adjust the [S3 bucket
policy](https://awspolicygen.s3.amazonaws.com/policygen.html).
The file uploads bucket should have a policy of:
```json
{
"Version": "2012-10-17",
"Id": "Policy1468991802320",
"Statement": [
{
"Sid": "Stmt1468991795370",
"Effect": "Allow",
"Principal": {
"AWS": "ARN_PRINCIPAL_HERE"
},
"Action": [
"s3:GetObject",
"s3:DeleteObject",
"s3:PutObject"
],
"Resource": "arn:aws:s3:::BUCKET_NAME_HERE/*"
},
{
"Sid": "Stmt1468991795371",
"Effect": "Allow",
"Principal": {
"AWS": "ARN_PRINCIPAL_HERE"
},
"Action": "s3:ListBucket",
"Resource": "arn:aws:s3:::BUCKET_NAME_HERE"
}
]
}
```
The file-uploads bucket should not be world-readable. See the
[documentation on the Zulip security model](security-model.md) for
details on the security model for uploaded files.
However, the avatars bucket is intended to be world-readable, so its
policy should be:
```json
{
"Version": "2012-10-17",
"Id": "Policy1468991802321",
"Statement": [
{
"Sid": "Stmt1468991795380",
"Effect": "Allow",
"Principal": {
"AWS": "ARN_PRINCIPAL_HERE"
},
"Action": [
"s3:GetObject",
"s3:DeleteObject",
"s3:PutObject"
],
"Resource": "arn:aws:s3:::BUCKET_NAME_HERE/*"
},
{
"Sid": "Stmt1468991795381",
"Effect": "Allow",
"Principal": {
"AWS": "ARN_PRINCIPAL_HERE"
},
"Action": "s3:ListBucket",
"Resource": "arn:aws:s3:::BUCKET_NAME_HERE"
},
{
"Sid": "Stmt1468991795382",
"Effect": "Allow",
"Principal": {
"AWS": "*"
},
"Action": "s3:GetObject",
"Resource": "arn:aws:s3:::BUCKET_NAME_HERE/*"
}
]
}
```
## Migrating from local uploads to Amazon S3 backend
As you scale your server, you might want to migrate the uploads from
your local backend to Amazon S3. Follow these instructions, step by
step, to do the migration.
1. First, [set up the S3 backend](#s3-backend-configuration) in the settings
(all the auth stuff), but leave `LOCAL_UPLOADS_DIR` set -- the
migration tool will need that value to know where to find your uploads.
2. Run `./manage.py transfer_uploads_to_s3`. This will upload all the
files from the local uploads directory to Amazon S3. By default,
this command runs on 6 parallel processes, since uploading is a
latency-sensitive operation. You can control this parameter using
the `--processes` option.
3. Once the transfer script completes, disable `LOCAL_UPLOADS_DIR`, and
restart your server (continuing the last few steps of the S3
backend setup instructions).
Congratulations! Your uploaded files are now migrated to S3.
**Caveat**: The current version of this tool does not migrate an
uploaded organization avatar or logo.
## S3 data storage class
In general, uploaded files in Zulip are accessed frequently at first, and then
age out of frequent access. The S3 backend provides the [S3
Intelligent-Tiering][s3-it] [storage class][s3-storage-class] which provides
cheaper storage for less frequently accessed objects, and may provide overall
cost savings for large deployments.
You can configure Zulip to store uploaded files using Intelligent-Tiering by
setting `S3_UPLOADS_STORAGE_CLASS` to `INTELLIGENT_TIERING` in `settings.py`.
This setting can take any of the following [storage class
value][s3-storage-class-constant] values:
- `STANDARD`
- `STANDARD_IA`
- `ONEZONE_IA`
- `REDUCED_REDUNDANCY`
- `GLACIER_IR`
- `INTELLIGENT_TIERING`
Setting `S3_UPLOADS_STORAGE_CLASS` does not affect the storage class of existing
objects. In order to change those, for example to `INTELLIGENT_TIERING`, perform
an in-place copy:
aws s3 cp --storage-class INTELLIGENT_TIERING --recursive \
s3://your-bucket-name/ s3://your-bucket-name/
Note that changing the lifecycle of existing objects will incur a [one-time
lifecycle transition cost][s3-pricing].
[s3-it]: https://aws.amazon.com/s3/storage-classes/intelligent-tiering/
[s3-storage-class]: https://aws.amazon.com/s3/storage-classes/
[s3-storage-class-constant]: https://docs.aws.amazon.com/AmazonS3/latest/API/API_PutObject.html#AmazonS3-PutObject-request-header-StorageClass
[s3-pricing]: https://aws.amazon.com/s3/pricing/