Stuart Barkley <[email protected]> writes: >> Why don't the intermediate accounting records cover that? > > Didn't know about them. Are they in the accounting file or just the > reporting file?
Sorry, just the latter. I don't remember what was in the original, but it's in http://arc/SGE/htmlman/htmlman5/reporting.html >> Is it worth worrying about any real accuracy in accounting? Tyring >> to keep track of everything relevant, like system problems that >> clobber jobs, just doesn't seem worthwhile to me. > > This is a very important basic question and strongly it interacts with > the accuracy of the collected information. I don't think it has much to do with what GE reports, but it seems I don't understand. > I think that as a practical > matter, the data collected by SGE is not good enough for precise use. > It misses any system usage outside of SGE (and depending heavily on > details I have not fully explored can miss internal usage). I'm surprised the resource manager would be expected to keep track of resources it doesn't manage, and that you run stuff on the nodes outside it -- that sounds like asking for trouble. Nevertheless, as I said, GE can track anything for which you can write a load sensor, and put it in the reporting database. >> There are assorted scripts floating around for doing that, including >> analyze.rb (?) in the distribution. Note that they need >> consideration of the configuration in general, such as the one >> slot/per node loosely integrated parallel jobs that were originally >> configured here... > I do assume each organization will need some customization (at a > minimum someone will want some color changed) I'm sure that's most important. > Do you mean applying compensation for over committed nodes and > dedicated nodes with only a single SGE job (but maybe using multiple > cores)? I can't remember the specifics off-hand, related to different numbers of accounting records per job and possible multiple counting that people fell foul of. > It can also get complicated when nodes have different characteristics > (processor speed, software licenses, etc). Yes. I reckon it's just not worth worrying about on a system like ours, not that we require more than fairly cursory accounting. > The problem is that it already exists and people will find references > to it and say "install it". The good news is that it was only prebult > for the Sun licensed version (as I understand it). I don't know why everything needs to come pre-built, but there was an arco `courtesy' tarball. The webconsole itself is separate, though. Dbwriter is from the same packaging/repo, and at least some of us think that's useful (even if it's still a bit of a Java horror). > I used qacct to generate some basic .csv files. Someone else used > excel (or something else) to make graphs for a management report. I wouldn't know anything about excel... > We will probably also need to do a web page showing some sort of > historical reporting. This will need to integrate into our web > content management system. Do share anything you can that's sufficiently general to be useful others directly or as examples, of course. _______________________________________________ users mailing list [email protected] https://gridengine.org/mailman/listinfo/users
