If you want to build a real high available mail system with the POPular server suite, some things have to be considered. The POPular proxy and server in itself are not enough to get a highly available system. This documenation can't go into all details of building a highly available system, but some issues to get you startet will be discussed.
For high availability you want to install more than one server running the POPular proxy software. You will need some way of failing over the IP address to the other server. Depending on your configuration of the proxy server you will have to change the configuration when this failover occures.
Each storage server has access to only a subset of all mailboxes. If a storage server fails, no access to this subset is possible anymore. All the mailboxes stored on other servers are unaffected. There is no good way around this. Even if you use a big dedicated NFS server for storing all the mailboxes, this server can fail. The only way to make this really redundant would be by saving all the mails on more than one hard drive attached to separate hosts, which would induce a hugh overhead of transaction processing to keep the two copies in sync. For most uses the best solution is probably to use a multihost capable RAID system for storing the mails. This RAID system is attached to more than one host. In case of failure of one host the other host attached to the same RAID mounts the disks of the failed host after doing a file system check. A similar solution could be build with a SAN (storage area network) based on fibrechannel or a similar technology.
Whichever system you use, when failing over to another storage server, the proxy servers needs to know which server to access. You can either move the IP number of the failed storage server over to the new server or change the configuration of the proxy server to the new IP number or host name.
Independently of the exact storage architecture used, you obviously want to use a RAID system instead of simple disks. Disks are by far the most unreliable part of any system and you should be sure to handle disk failures without downtimes.
Exactly how you get the mail delivered to the storage server is not part of this documentation. Depending on what mechanism is used for deciding which mail to deliver where, the fail over of a storage server has to be taken into account.