RestInCloud
simple restfully dockerized distributed open source cloud backup server ;-)
Install Server
to install the ric-server check deployment dir for different enviroments (docker, apache, nginx)
Client for CLI
if your ric-servers are running, you can use the smart "ric" cli tool to backup etc. and manage the cluster
Help
Api Overview
Show Help
GET /
GET /?help
File:
PUT /filename # backup/upload a file
POST /filename # refresh file if exists
GET /filename # restore/download a file
GET /filename?check # check exists, replicas, freshness, size of the file
GET /filename?list # list all versions of the file
DELETE /filename # delete a selected or all version of the file
Server Management
GET /?list # list all files on this server
GET /?info # show server info
GET /?health # show cluster health
POST /?addServer # add a remote server to server (internal for cluster management)
POST /?removeServer # remove a server from connected server (internal for cluster management)
POST /?joinCluster # add the server to existing cluster
POST /?leaveCluster # leave a cluster
POST /?removeFromCluster # drop a other server out of cluster
POST /filename push=>server # push file to server
Api Details
- GET http://ric1.server/?help - show this help
- GET http://ric1.server/?list - list all files ... &pattern=~regEx~i&limit=100&start=100&showDeleted=1 (ordered by random!)
- GET http://ric1.server/?listDetails - list all files with details - parameters like ?list
GET http://ric1.server/?info - show server infos (and quota if set)
-
PUT http://ric1.server/error.log - upload a file to the store
- use ?timestamp=1422653.. to set correct modificationTime [default:requestTime]
- use &retention=last3 to select the backup retention strategy [default:last3]
- use &noSync to suppress syncronisation to replication servers (used for internal sync)
- retention strategies (versions sorted by timestamp): {retentionList}
- with curl: curl -X PUT --upload /home/www/phperror.log http://ric1.server/error.log curl -X PUT --upload "/home/www/phperror.log" "http://ric1.server/error.log&retention=last7×tamp=1429628531"
-
POST http://ric1.server/error.log?sha1=23423ef3d..×tamp=1422653.. - check and refresh a file
- checks if version exists and updates timestamp
- returns 1 if version was updated, 0 if version not exists
- if 1 is returned, there is no need to upload the same version
- use &noSync to suppress syncronisation to replication servers (used for internal sync)
- if &noSync is not set, refresh is also performed on the replication servers (with &noSync)
POST http://ric1.server/error.log ?action=delete - delete a file !! Attention if version is omitted, ALL Versions will be deleted
- POST http://ric1.server/error.log ?action=push&targetServer=ricx.example.com - push file to other server (with admin pw)
DELETE http://ric1.server/error.log - delete a file !! Attention if version is omitted, ALL Versions will be deleted
GET http://ric1.server/error.log - download a file (etag and lastmodified supported)
- GET http://ric1.server/error.log?&version=13445afe23423423 - version selects a specific version, if omitted the latest version is assumed (must not: ...error.log?version... )
- GET http://ric1.server/error.log?list - show all (or &limit) versions for this file; (ordered by latest); &showDeleted=1 to include files marked for deletion
-
GET http://ric1.server/error.log?check&minSize=40000&minReplicas=2&minTimestamp=14234234&sha=1234ef23
- check that the file
- (1) exists,
- (2) size >40k [default:1],
- (3) fileTime>=minTimestamp [default:8d],
- (4) min 2 replicas (3 files) [default:max(1,count(servers)-1)]
- (5) sha1, (if sha give, the size is irrelevant)
- returns json result with status: OK/WARNING/CRITICAL, a msg and fileInfo
admin Commands: Post: with parameters: [action: addServer, addServer: s1.cs.io:3723]
- addServer s1.cs.io:3723 - add Server to local list,
- removeServer s1.cs.io:3723 - remove Server from local list
- removeServer all - remove all Servers from local list
- joinCluster s2.cs.io - join to existing cluster (or join a single node and create a cluster)
- leaveCluster - leave a cluster (send removeServer=self to all cluster nodes an clear my servers list)
-
removeFromCluster s3.cs.io - kick a server from the servers list of all known nodes (use this if the server is unresponsive and you can't send a leaveCluster)
auth (only as parameter supported yet)
-
use &token=YourAdminToken to authenticate as admin or writer or reader (e.g. for info command)
serverVersion
- use &minServerVersion=1.4.0 to require a minimal ServerVersion - it also acts reverse - if the client major version not matches the server major version - the request is rejected - to protect old clients from doing mad things
Usecases
Backup a Dir
dir to back up: /home/www/configs/ server identification: myServer encryption password: fooSecret ricServer: ric1.server.de ricWriterToke: barSecret
store the passwords
echo "fooSecret" > /home/www/ricPassFile.txt
chmod 600 /home/www/ricPassFile.txt
echo "barSecret" > /home/www/ricWriterFile.txt
chmod 600 /home/www/ricPassFile.txt
cronjobs (with sic monitoring)
*/5 * * * * /usr/local/sbin/ric backup /home/www/configs/ myServer-configs.tar.bz2 --retention=last7 --passFile=/home/www/ricPassFile.txt --prefix=myServer- --authFile=/home/www/ricWriterFile.txt --server=ric1.server.de 2>&1 >/dev/null | /usr/local/sbin/sic /myServer/ric-backup --STDINasCRITICAL
*/2 * * * * /usr/local/sbin/ric check myServer-configs.tar.bz2 --prefix=myServer- --authFile=/home/www/ricWriterFile.txt --server=ric1.server.de --minTimestamp=-300 2>&1 >/dev/null | /usr/local/sbin/sic /myServer/ric-backup/check --STDINasCRITICAL
Server Manifest
- no daemon type, only react on request, no long running requests
- all response is json, except download and errors
- the base is a normal filesystem, files in a hashed directory, with original filenames extended by it's sha1
- no other persistent storage for attributes no database
- it's seen as critical, if a file has no replicas
License
The MIT License (MIT)