3

We're looking into putting up a wiki or CMS for internal use by our IT department. One of the big things we want to use it for is disaster recovery procedures.

Given that a disaster, such as a power or network outage, might render the wiki inaccessible, it seems sensible to to host the wiki in two places so that if one is inaccessible, we can fall back to the other.

Are there any wikis or CMSes that synchronize (or an alternate way to achieve a similar end)?

6 Answers 6

3

Assuming you don't have tons of entries per minutes in your wiki, You can also achieve synchronization by using NoSql db CouchDB or Tokyo as a backend. There is also redis. They are replication-friendly.

An experimental solution could be also TiddlyWiki on a Dropbox, really decentralized ! (or maybe too crazy but you can test It :) )

About the Wiki and CMS you can use with those backends Github is the place to search. really really simple but extendable simple sinatra couchdb wiki Instiki (Rails) could be adapted with Ohm

You can also consider sqlite3 asa synchronizable backend since the db It's just a file. Frameworks like Rails and Django support It out of the box.

MORE: After some thinking It comes to my mind that of course you can use git too, since it's a decentralized thing by nature. If you build your wiki on top of git, You can add hooks to every push to fire synchronization with other cloned repositories of your wiki or just luanch a pull every few minutes. And here yuo have an example of a wiki-over-git.

I've found also zim that you could combine, again with git.

4
  • Any idea which wikis would support using those databases as a backend? May 19, 2010 at 17:56
  • I really get a kick out of the dropbox idea! May 19, 2010 at 17:57
  • Added more stuff to answer you.
    – microspino
    May 19, 2010 at 18:36
  • Added a git solution too.
    – microspino
    May 21, 2010 at 11:13
2

i use dokuwiki with file storage. works very well, and rsync as synchronization metod does a trick.

i trust rsync with all my backups, why would i not trust with wiki replication?

plain-text files in which this wiki stores pages are perfectly readable and searchable directly from shell/mc which is another disaster-recovery advantage.

1

Wikis like MoinMoin don't use database servers, only local files that can be synced easily with i.e. rsync.

1
1

If you like minimalistic file based frameworks you can try werc. It follows the plan9 philosophy so expect everything to be a file. That way, sinchronization is easy.

1

I think you could use fossil for this, even if it is not its main goal to be a CMS. It has wiki and tickets, it can sync over network, it is a single executable, it can run in offline mode it has history of changes (since it is a distributed version control tool).

0

ScrewTurn Wiki also supports file-based storage, however I would advise you not to rely on home-brew replication strategies (like rsync). I would just opt for a database server that supports replication and failover clustering (like SQL Server).

3
  • 1
    rsync is a weird replication strategy? Care to explain?
    – Sven
    May 19, 2010 at 17:45
  • ScrewTurn is an ASP.NET application. I suspect that on Windows, using rsync is a little bit unusual. May 19, 2010 at 17:55
  • @SwenW rsync is "weird" in a sense that it might not replicate all the data in a way that makes sense. What if the application modifies the data while you're replicating? You'd end up with a broken replica. That's why I suggested to use SQL Server (or any other more robust system that supports hot replication). May 20, 2010 at 5:36

Not the answer you're looking for? Browse other questions tagged .