tag:blogger.com,1999:blog-145332831106508186.post2763616431987357577..comments2023-03-19T09:06:24.378-04:00Comments on /var/log/rant: Questions about GitDave Jacobhttp://www.blogger.com/profile/15052163927020492687noreply@blogger.comBlogger4125tag:blogger.com,1999:blog-145332831106508186.post-14823833549539649042011-12-06T13:32:19.491-05:002011-12-06T13:32:19.491-05:00gizmo: I know! I know! I know! I'm pretty sure...gizmo: I know! I know! I know! I'm pretty sure that it's RAID 5 set as conservatively as possible over a fairly large array, so my data is on several discs, so no worries unless I'm stupid. Problem is, I've been known to be stupid. I have my ~/bin and ~/dev directories on desktop shared to Dropbox, just because once I was stupid with code I cared about. But that's my desktop, not the server. <br /><br />Which is why I'm now so interested in learning git.Dave Jacobhttps://www.blogger.com/profile/15052163927020492687noreply@blogger.comtag:blogger.com,1999:blog-145332831106508186.post-32674771091209257162011-12-06T10:57:55.652-05:002011-12-06T10:57:55.652-05:00RAID is not data protection, it is hardware protec...RAID is not data protection, it is hardware protection. Assuming it's RAID 5 (or RAID 1 to a lesser extent) , that means if a drive fails you can easily swap it out.<br /><br />A nice side benefit is that if it is RAID 5 (or sort of RAID 1) that if a drive fails it isn't taking your data with it.<br /><br />RAID 0 is even worse, it just makes one big chunk of storage and if a drive fails, well so sorry about your data.<br /><br />Data is only protected if a copy or twenty exist elsewhere and a restore can be made from it.<br /><br />/end sysadmin rantgizmo mathboyhttps://www.blogger.com/profile/11021393775941673720noreply@blogger.comtag:blogger.com,1999:blog-145332831106508186.post-48599826212825538082011-12-06T09:25:02.346-05:002011-12-06T09:25:02.346-05:00development vs deployment: I was thinking that was...development vs deployment: I was thinking that was an issue, but what little I've done with git so far has pushed me toward that, creating ~/dev for working with my stuff and ~/bin for deploying them once I'm happy with them. That works for scripts, as the big difference between ~/bin and ~/dev is that ~/bin is in my path.<br /><br />Where it works less is the rest of the work, which is the web-based interface for our genomics lab. The scripts must be in the CGI bin to be tested, and we do not (yet) have a development server. Once we do, that'll help a lot. <br /><br />ThanksDave Jacobhttps://www.blogger.com/profile/15052163927020492687noreply@blogger.comtag:blogger.com,1999:blog-145332831106508186.post-53539969669000019702011-12-05T17:24:50.896-05:002011-12-05T17:24:50.896-05:00Or do we really, truly need a server?
Since git is...<i>Or do we really, truly need a server?</i><br />Since git is a DVCS, there is no server per se. You can easily set things up like in the SVN model and declare one repository the anointed one. If your machine is well-connected, it can be the one and no extra hardware is necessary. However, setting up on a dedicated server somewhere usually is a better idea so that the sysadmins can care for it properly, and so you can install some ecosystem around it, e.g. a browsable Web interface.<br /><br /><i>should I have those be the directories to run git in</i><br />You're mixing up working on a file/storing it in the VCS with deployment. Please do have a separate deployment step. In the most simple case that's a Makefile calling install (from GNU coreutils). If the project is very Perl-centric, following the standard practices and using Module::Build is not a bad idea, either. If you describe your software in detail, one may make specific recommendations; there's rarely something new under the sun and thus there is a diversity of pre-canned recipes for certain situations. Come to irc.perl.org #toolchain for advice.Anonymousnoreply@blogger.com