The planning of a new system architecture at Hcoop, Inc. is motivated by a set of concerns with the current infrastructure. This page details these concerns in order to improve ourselves in Hcoop, Inc. v. 2.

1. What deficiencies in our current set-up do we want to correct?

• If one machine goes down, then our services will go down, too.
• No way for members to buy arbitrary amounts of disk space
• No automated off-machine backups

  • Until we get new machine(s) set up, would it be possible to rsync the server contents off-site? I would be willing to contribute space on my linux workstation to this purpose... or do bandwidth constraints make this unreasonable? I imagine that our content is relatively static so that might make this possible. --JustinLeitgeb

• We don't have a human administration set-up where everyone with an admin job:
  • Feels compensated in proportion to the work he does
  • Is accessible on a predictable schedule
• PHP scripts have a time limit. This created problems with Squirrel Mail:
  • It is often impossible to attach large files or download large attachments.
  • When Fyodor is running its backup, the system becomes slow. Sometimes folder listings don't display because the php script times out before the folder is read from the imapd.

    • Can we correct this by "nicing" the backup process? Might be a quick fix. --JustinLeitgeb

    • Already done, unfortunately. --AdamChlipala

• Downloads from www.hcoop.net/~username are not counted toward the user's bandwidth usage

  • Actually, they are counted as much as any other web activity, though we don't have any formal idea of "bandwidth quota" ATM. Do you have any information to indicate otherwise? --AdamChlipala