mglazer: wtf? could you please do that somewhere else?
code: I asked ask if he was going to do some statistics for metamark. He said that he had planned to, and that he had some sql in the TODO, but that he had not. He was kind enough to provide the code. However, it wants to talk to mysql, and I'm pretty much a postgres kinda guy.
So I know that its sort of one of those jihads among OSS geeks, postgres or mysql. Both camps seem firmly convinced that theirs is better (the postgres camp: "mysql isnt a real database", the mysql camp: "postgres is slow and hard to use"), but I've never seen what I considered an honest comparison of the two of them. So I've been doing that myself now I guess. My first impressions were not good. The build process is fairly easy. After getting it installed, I was appalled to know that mysql has a "root" user and it needs its password set. My first thought was, "oh, christ, I can only imagine how many users out there think that mysql root is system root and vice-versa." (note: I once gained root on a machine by finding the user's "root" user password in the database [for php nuke or something silly], and su'ing with it in the shell)
Next, after the software has been installed, it makes assumptions about where the data should go. In my case, I think this was /opt/alex/mysql/opt/var. This bugged me, because the assumption is that users should keep their data where their binaries are. That's just flawed. Additionally, when trying to find out how to pass it arguments (just passing it arguments did not get a 'usage' message), I had to pore through some bash source.
Ok, so after mysql was installed, and had its own directory (on /global/database/mysql64, naturally), it was fairly easy to get running and poke at in mysql's interpreter. It seems somewhat friendlier than postgres' psql(1), but it wasn't particularly useful to me, since I couldn't issue "help create" ("\h create" in psql). The sql differences are minor, but different enough to be a pain. Additionally, using mysql_setpermission, I was able to "create a user and database," but nothing really happened, so I have an xrl database which I created, and an xrl user that can't connect to it.
I'm pretty frustrated by all this, and the people in #mysql on freenode were not exactly helpful. In fact, I mostly just sat through arguments about why php did or did not suck. I'm probably going to poke on this some more today, but I've just got a bad taste in my mouth from it, and I'm pondering just porting metamark to postgres to be done with it
geek: my Octane showed up. I've been struggling with getting an OS on it. My only wide scsi cable [ed: I now have ultra 320 disks coming out of my ears. the world sucks...) is attached to a raid on my Ultra, so I can't attach a CDROM to it, and Octane's just don't have internal cd's. So I was trying to netboot it off my Challenge, but quickly tired of it. The heat produced by an Octane and a Challenge in my apartment is just obscene, and I could only go at it for a few hours at a time. A friend of mine fedex'd me a new scsi cable though, so I should be able to get media installed tonight without booting the challenge.
pudge is a cool dude. As it turns out, many large networks I frequents have large populations of Mac users, and I'm pondering going file poaching. I suppose its only a matter of time before Apple sets up some cryptographic challenge or something to discourage curl/wget users
Job: Well, we have a meeting with DISA for next Thursday, and I've been hearing good things about OSD, but I'm not sure what I'm going to do to come up with another month of money. I'm sick of being unemployed.
Next, after the software has been installed, it makes assumptions about where the data should go. In my case, I think this was /opt/alex/mysql/opt/var. This bugged me, because the assumption is that users should keep their data where their binaries are. That's just flawed. Additionally, when trying to find out how to pass it arguments (just passing it arguments did not get a 'usage' message), I had to pore through some bash source.
Ok, so after mysql was installed, and had its own directory (on /global/database/mysql64, naturally), it was fairly easy to get running and poke at in mysql's interpreter. It seems somewhat friendlier than postgres' psql(1), but it wasn't particularly useful to me, since I couldn't issue "help create" ("\h create" in psql). The sql differences are minor, but different enough to be a pain. Additionally, using mysql_setpermission, I was able to "create a user and database," but nothing really happened, so I have an xrl database which I created, and an xrl user that can't connect to it.
I'm pretty frustrated by all this, and the people in #mysql on freenode were not exactly helpful. In fact, I mostly just sat through arguments about why php did or did not suck. I'm probably going to poke on this some more today, but I've just got a bad taste in my mouth from it, and I'm pondering just porting metamark to postgres to be done with it
geek: my Octane showed up. I've been struggling with getting an OS on it. My only wide scsi cable [ed: I now have ultra 320 disks coming out of my ears. the world sucks...) is attached to a raid on my Ultra, so I can't attach a CDROM to it, and Octane's just don't have internal cd's. So I was trying to netboot it off my Challenge, but quickly tired of it. The heat produced by an Octane and a Challenge in my apartment is just obscene, and I could only go at it for a few hours at a time. A friend of mine fedex'd me a new scsi cable though, so I should be able to get media installed tonight without booting the challenge.
pudge is a cool dude. As it turns out, many large networks I frequents have large populations of Mac users, and I'm pondering going file poaching. I suppose its only a matter of time before Apple sets up some cryptographic challenge or something to discourage curl/wget users
Job: Well, we have a meeting with DISA for next Thursday, and I've been hearing good things about OSD, but I'm not sure what I'm going to do to come up with another month of money. I'm sick of being unemployed.