Forum Moderators: open
That you can access some documents, but most of the system is not reachable, is a little bit more difficult to eplain. The public servers are a cluster of machines which take the data from the main server (which is down) and cache it for some days to reduce load. So if documents are still in their cache (like some of the static documents), you will be able to access them for some time. Because www.dmoz.org is a cluster, several attempts to access a file might lead to different result.
How can they possibly rely on a single server as their main data store and not have a backup or a plan for a failover of any sort? How can they not have the "public server" cluster able to continue to serve up the directory during the time the "main server" is down?
How do you know this, are you privvy to inside info? Is the info on this downtime being published anywhere other than this thread? Is the DMOZ's architecture publicly available?
I'm really just curious, but an acquaintance of mine from another webmaster's forum (foolishly?) relies on the DMOZ for a substantial portion of his website and traffic.
Thanks.
reports of it's demise were somewhat premature ..( and despite the wishfull thinking of many )..it appears to be online ..even if that upsets some folks around here ;-)
nothing is perfect ..dmoz is no exception ..get over it ..that isn't what it's there for ..yay! for diversity!
it isnt the fault of dmoz if google uses it as some sort of bench mark ...and even less if one cant get in ones preferred category and have ones preferred place there within ..
anyone basing their traffic on their place in dmoz is being woefully naive ..sacrificing chickens at dawn might well be just as relevant ..and one would be equally as entitled to ( female dog ) at the chickens if it didnt result in a better result in serps ..
[edited by: Leosghost at 11:01 pm (utc) on Oct. 22, 2006]
It's published first in internal editor forums.
>Is the DMOZ's architecture publicly available?
There have been some public discussions of its network. (It's really quite simple--fewer than a dozen servers, most of which the public doesn't see. Don't think of it as a portal: think of it as a content-development workshop -- to take another well-known online example, dmoz.org is not like Project Gutenberg, it's like Distributed Proofreaders.) As is appropriate for a development system, the focus is very much on data integrity rather than on data availability, with the expectation is that licensees will provide always-available "public access" from their own servers.)
>an acquaintance ... (foolishly?) relies on the DMOZ for a substantial portion of his website and traffic.
If his concern is on data availability, and a lot of traffic is involved, he should (definitely!) be hosting a clone of the ODP via RDF-publication rather than depending on parasitic syphons on the public servers.
[edited by: Webwork at 12:47 pm (utc) on Oct. 23, 2006]
[edit reason] Please initiate a new thread to introduce entirely new issues [/edit]
[edited by: Webwork at 3:15 pm (utc) on Oct. 24, 2006]
[edit reason] Emphasis Added [/edit]
If anyone wishes to raise new issues relating to the outtage - such as consequences of the outtage or the need for redundancy - please initiate a new thread and the same will be promptly activated if it is in compliance with the Charter.
[edited by: Webwork at 3:14 pm (utc) on Oct. 24, 2006]