MCorange/wings
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
1,374 Commits 42 Branches 73 Tags 43 MiB
Commit Graph

36 Commits

Author SHA1 Message Date
Matthew Penner
d1c0ca5260
server(filesystem): rebuild everything imaginable 
This wonderfully large commit replaces basically everything under the
`server/filesystem` package, re-implementing essentially everything.

This is related to
https://github.com/pterodactyl/wings/security/advisories/GHSA-494h-9924-xww9

If any vulnerabilities related to symlinks persist after this commit, I
will be very upset.

Signed-off-by: Matthew Penner <me@matthewp.io>
 2024-03-13 12:27:29 -06:00
Daniel Barton
d1fd0465e4
server(filesystem): fix file names when extracting archives (#165) 2023-10-17 20:47:36 -06:00
Matthew Penner
57e7eb714c
Streaming Transfers (#153) 2022-11-14 18:25:01 -07:00
Matthew Penner
eb4df39d14
server(filesystem): fix inaccurate archive progress (#145) 2022-11-06 13:38:30 -07:00
Matthew Penner
7245791214
backups(s3): send uploaded parts in complete request 2022-09-26 11:14:57 -06:00
Matthew Penner
5cd43dd4c9
archive: keep timestamps when extracting 2021-09-01 09:54:41 -06:00
Matthew Penner
31ff3f8b56 server(fs): keep file mode when extracting archive 2021-07-15 15:37:38 -06:00
Dane Everitt
ddfd6d9cce Modify backup process to utilize contexts and exponential backoffs 
If a request to upload a file part to S3 fails for any 5xx reason it will begin using an exponential backoff to keep re-trying the upload until we've reached a minute of trying to access the endpoint.

This should resolve temporary resolution issues with URLs and certain S3 compatiable systems such as B2 that sometimes return a 5xx error and just need a retry to be successful.

Also supports using the server context to ensure backups are terminated when a server is deleted, and removes the http call without a timeout, replacing it with a 2 hour timeout to account for connections as slow as 10Mbps on a huge file upload.
 2021-05-02 12:28:36 -07:00
Matthew Penner
0e3778ac47 transfers: use backup archiver 2021-03-07 11:04:15 -07:00
Dane Everitt
98c68142cd
Remove all of the remaining API logic and port it all to the remote.Client type 2021-02-01 21:28:46 -08:00
Dane Everitt
6ef0bd7496
Get general concept for backup resotration using a unified interface implemented 2021-01-18 21:20:58 -08:00
Dane Everitt
66b6f40b61
Fix import cycle issue 2021-01-17 21:05:51 -08:00
Dane Everitt
902f9f5944
Include more (and customizable) context with backup logs 2020-12-27 16:16:40 -08:00
Matthew Penner
b26db99ee7 Fix part_id field in logs 2020-12-27 13:02:50 -07:00
Matthew Penner
de4d2f4724 Make logs follow a more consistent format 2020-12-27 13:01:34 -07:00
Matthew Penner
640e30de8a Ignore symlinks with missing target, add better backup logging, update copyright year to 2021 2020-12-27 12:54:18 -07:00
Matthew Penner
5c78cb9ab3
Add transfer logging output (#77) 
Co-authored-by: Dane Everitt <dane@daneeveritt.com>
 2020-12-25 13:32:41 -08:00
Matthew Penner
901ab1157d
Rework archiving logic to be more consistent and less impactful on disk IO (#79) 
Co-authored-by: Dane Everitt <dane@daneeveritt.com>
 2020-12-25 11:52:57 -08:00
Matthew Penner
b52c3fb61e Cleanup backup/backup_s3.go 2020-12-06 15:25:11 -07:00
Matthew Penner
ee08829a28 s3 backups: handle CompleteMultipartUpload and AbortMultipartUpload on the panel 2020-12-06 13:56:17 -07:00
Matthew Penner
de51fd1c51
Error handling improvements (#71) 
* Remove `emperror.dev/errors`, remove all `errors#Wrap` and `errors#WithStack` calls
* Improve logging in `server/backup.go`
 2020-11-28 16:57:10 -07:00
Dane Everitt
be9d1a3986
Replace error handling package with emperror; add better reporting for errors escaping server root 2020-11-08 13:52:20 -08:00
Matthew Penner
05cfb59e18 Add better logging for s3 backup uploads 2020-11-01 15:36:16 -07:00
Matthew Penner
5475cb02c1 Cleanup s3 backup uploads, add request timeouts 2020-11-01 15:23:46 -07:00
Matthew Penner
1239b1c0ca Switch to s3 multipart uploads for backups 2020-11-01 15:23:46 -07:00
Dane Everitt
de30e2fcc9
Dont attempt to get size within archive process, will return empty; ref pterodactyl/panel#2420 
The stat call is operating against an unflushed file if called in the archive function, so you'll just get the emtpy archive size, rather than the final size.

Plus, we only used the file stat in one place, so slight efficiency win?
 2020-09-27 11:16:38 -07:00
Matthew Penner
7ba32aca84 Fix some typos and run gofmt on all .go files 2020-09-05 13:08:40 -06:00
Dane Everitt
0e474c8b24
Better error captures 2020-08-27 19:57:22 -07:00
Dane Everitt
c1e591c99b
Support generating a compressed archive for a server via the API 2020-07-11 13:13:49 -07:00
Dane Everitt
7d4a8d7f7e
Update everything expect transfers & sftp to not use zap 2020-06-13 10:40:26 -07:00
Dane Everitt
0bd28a4480
Cleanup S3 support; send actual backup details in response 2020-05-09 19:24:30 -07:00
Dane Everitt
b2797ed292
Abstract out shared backup functionality 2020-05-02 15:02:02 -07:00
Dane Everitt
507d0100cf
Hilariously rough code to get a backup pushed into S3 2020-04-26 17:20:26 -07:00
Dane Everitt
91d12ab9a7
More abstract support for backups & misc code cleanup in that area 2020-04-26 16:43:18 -07:00
Dane Everitt
df6d98bbda
Add support for ignoring directories/files; fix compression of archives 2020-04-18 23:26:23 -07:00
Dane Everitt
4ff7bd2777
Abstract out more of the backup functionality 2020-04-17 13:46:36 -07:00