I decided to host blog in OracleCloud Always free object storage. 50Gb is plenty and does not cost me a penny (for now).

I plan to use Rclone to sync local site to the cloud.

I was loosely following guide from Oracle on RClone Cloud Shell refused to start for me, so I just pull all info manually form the GUI.

  1. Create bucket named blog in OracleCloud
  2. Upload something, and check “item properties” to get full URL https://objectstorage.eu-amsterdam-1.oraclecloud.com/n/axpvighhs5ai/b/blog/o/comics-pic.jpeg
    eu-amsterdam-1 is a <REGION>
    /n/axpvighhs5ai/b/blog/o/ is a prefix in my case
  3. record Namespace: form bucket information, it looks like axpvighhs5ai
  4. Create new user in Oracle Cloud and denerate access_key/Secret_key under
    Identity -> Users -> Username -> Customer Secret Keys
    It is BAD that user does not need to have access to everything n OracleCloud. But at this moment I do not know how to set permissions in Oracle. Things would be simpler if I use AWS S3.
  5. put /n/axpvighhs5ai/b/blog/o as prefix in _config_yaml and https://objectstorage.eu-amsterdam-1.oraclecloud.com as base URL and regenerate site. It will break local browsing.
  6. pkg install rclone
     blog:/root@[16:42] #  man rclone
     No manual entry for rclone

    This is a fail. No man page? RLY?

  7. create ~/.rclone.conf for whatever user which doing uploads.
     type = s3
     provider = Other
     env_auth = false
     access_key_id = <ACCESS KEY>
     secret_access_key = <SECRET KEY>
     endpoint = <NAMESPACE>.compat.objectstorage.<REGION>.oraclecloud.com
  8. rclone sync --dry-run -i /usr/local/www/thttpd/ oracle:/blog to see what happens
  9. rclone sync -i /usr/local/www/thttpd/ oracle:/blog
  10. Get back to bucket and get url for index.html. Browse to It. You should be able to see your site.

It is still really bad, URL is ugly, CNAME redirection will not work. It hardly can be called a hosting.

But enough for today.