Mirror or not to Mirror?
May. 6th, 2009 09:40 am![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
kerravonsenI'm beginning to think that it isn't worth the effort of mirroring www.katspace.org at too.katspace.org; because it is too large, and, alas, not a straightforward case of just doing an rsync, because (a) part of the site is a wiki and (b) I can't seem to get the wiki working on too.katspace.org. I have tried a few different approaches, and they all failed in various ways.
Note that part of the difficulty is that www.katspace.org is running on Apache, and too.katspace.org is running on Zeus.
A. Mirror the whole site by copying it with wget
1) Takes too long, because the site is so big
2) For some reason, Apache isn't giving "Last Modified" headers, which means that wget insists on copying every file every time, so every time takes just as long as the first time.
B. Mirror the non-wiki part with rsync, and the wiki part with wget
1) Still have the "Last Modified" problem with the wiki part
2) Broken links in the non-wiki part, because I use Multiviews to have friendly URLs, and the webhost does not support Multiviews.
C. Mirror the non-wiki part with rsync, and also the wiki part with rsync, running the wiki on the webhost
1) Still have the broken links problem with the non-wiki part
2) The wiki relies on Apache mod_rewrite to make its links transparent, so there are more broken links
C. Mirror the non-wiki part with rsync, and the wiki part with rsync, and try other approaches to try to get "clean" URLs for the wiki, though not as "clean" as using mod_rewrite
1) Still have the broken links problem with the non-wiki part
2) The alternative "clean URLs" approaches would still end up with broken links
3) But I couldn't get the alternative approaches to work anyway, in fact, I couldn't get the wiki to work at all.
All of these approaches highlight why I decided to host my website on my own system in the first place; it is just too difficult to do advanced things with the Zeus webserver. Oh, I can understand why they use it; it is very fast, and very good at serving simple sites. But utterly useless if one wishes to manipulate URLs to make them more friendly.
So what should I do with too.katspace.org? I'd like to do something with it. Because it is a much faster and more reliable system than my own. I mean, I get power failures at home while I'm at work and then all of www.katspace.org is down for hours and there's nothing I can do about it. Should I mirror a subset of www.katspace.org on too.katspace.org? Such as the gallery, which would probably benefit from having a faster server. Should I make too.katspace.org a whole different site? Take some of the stuff that's on www.katspace.org off it and put it solely on too.katspace.org? (Such as the gallery, again)
What do you think?
Note that part of the difficulty is that www.katspace.org is running on Apache, and too.katspace.org is running on Zeus.
A. Mirror the whole site by copying it with wget
1) Takes too long, because the site is so big
2) For some reason, Apache isn't giving "Last Modified" headers, which means that wget insists on copying every file every time, so every time takes just as long as the first time.
B. Mirror the non-wiki part with rsync, and the wiki part with wget
1) Still have the "Last Modified" problem with the wiki part
2) Broken links in the non-wiki part, because I use Multiviews to have friendly URLs, and the webhost does not support Multiviews.
C. Mirror the non-wiki part with rsync, and also the wiki part with rsync, running the wiki on the webhost
1) Still have the broken links problem with the non-wiki part
2) The wiki relies on Apache mod_rewrite to make its links transparent, so there are more broken links
C. Mirror the non-wiki part with rsync, and the wiki part with rsync, and try other approaches to try to get "clean" URLs for the wiki, though not as "clean" as using mod_rewrite
1) Still have the broken links problem with the non-wiki part
2) The alternative "clean URLs" approaches would still end up with broken links
3) But I couldn't get the alternative approaches to work anyway, in fact, I couldn't get the wiki to work at all.
All of these approaches highlight why I decided to host my website on my own system in the first place; it is just too difficult to do advanced things with the Zeus webserver. Oh, I can understand why they use it; it is very fast, and very good at serving simple sites. But utterly useless if one wishes to manipulate URLs to make them more friendly.
So what should I do with too.katspace.org? I'd like to do something with it. Because it is a much faster and more reliable system than my own. I mean, I get power failures at home while I'm at work and then all of www.katspace.org is down for hours and there's nothing I can do about it. Should I mirror a subset of www.katspace.org on too.katspace.org? Such as the gallery, which would probably benefit from having a faster server. Should I make too.katspace.org a whole different site? Take some of the stuff that's on www.katspace.org off it and put it solely on too.katspace.org? (Such as the gallery, again)
What do you think?
