There's an upgrade to the job scheduler that every unix-server-using group must go thru here at work. I've been assigned to that project for our team. (Heck, I may be a grease monkey, but I'm a good grease-monkey.) Basically, instead of each server handling its own scheduling, it is done by an intranet interface that has 3 servers linked it to it, and none of those servers is a group's actual server. This apparently will make it easier to handle failures. Maybe it will. Yesterday though, I was talking to someone in another group who's also been dealing with that and who is quite leery of the underlying complexity, and that brought to mind Scotty's derisive comment about how easy it was for him to knock out the Excelsior's warp drive.
no subject
Date: Oct. 23rd, 2008 01:05 am (UTC)no subject
Date: Oct. 23rd, 2008 01:34 pm (UTC)There's an upgrade to the job scheduler that every unix-server-using group must go thru here at work. I've been assigned to that project for our team. (Heck, I may be a grease monkey, but I'm a good grease-monkey.) Basically, instead of each server handling its own scheduling, it is done by an intranet interface that has 3 servers linked it to it, and none of those servers is a group's actual server. This apparently will make it easier to handle failures. Maybe it will. Yesterday though, I was talking to someone in another group who's also been dealing with that and who is quite leery of the underlying complexity, and that brought to mind Scotty's derisive comment about how easy it was for him to knock out the Excelsior's warp drive.