INFORMIX DBA Survival Guide
Has anyone out there attempted to use the book from INFORMIX/press called INFORMIX DBA Survival GUIDE? Is it worth the $'s for someone who has taken a the most of the Informix training classes? Any comments? Thanks Sent via Deja.com http://www.deja.com/ Share what you know. Learn what you don't.
In article <7mg41b$s4q$1@nnrp1.deja.com>, coburn@aottsd.uscourts.gov wrote: > Has anyone out there attempted to use the book from INFORMIX/press > called INFORMIX DBA Survival GUIDE? Is it worth the $'s for someone > who has taken a the most of the Informix training classes? Any > comments? First, make sure you get the second edition of the book. You will probably have a hard time finding a serious Informix DBA that doesn't have (or wish he had) this book. Get it. I've been doing Informix for about 15 years, have taken Informix training classes, and I still find good things in the book. I do have one complaint, though. There are a lot of places where it will say something like "If you are experiencing excessive bufwaits, ..." without giving any kind of clue on what would constitute "excessive". It would really be nice to have even a ballpark estimate of what would seem like too high a value for this and all of the other measurements. I guess I have two complaints. The index isn't the greatest. Oops, I'm going to go ahead and add this third one, even though I don't that Joe had any real say in the matter and it's one of those "don't look a gift horse in the mouth" kind of things. The CBT course "chosen especially for this book" that comes free on the included CD is "MS Exchange Server 5.5:Connectors". It seems to me like they could have put a little more effort into finding a relevant course to include. I won't even bother looking at this one to see if I might want to try any of their other courses. -- --Marty Allred Sent via Deja.com http://www.deja.com/ Share what you know. Learn what you don't.
I'm going to presume to speak for Joe here since I tech reviewed the book for him. Excessive bufwaits is defined as a bufwait percent of more than a few percentage points. Percent of what? Here is the formula: bufwait% = (bufwaits / (pagreads + bufwrits)) * 100.00 This is a formula that I developed 2 years ago to track bufwaits when I first upgraded to 7.xx and discovered the LRUS bug. It is based on the concept that bufwaits are caused by LRU or buffer contention. Since you can see buffer contention in lokwaits I concentrated on LRU contention. A session has to latch an LRU, and might have to wait, in order to read a clean page into a buffer (pagreads) and to write to a page dirtying it (bufwrits). Informix has never acknowledged whether this is a good calc but it always works for me. For a 5.0 server the % is almost always <2% for a healthy server (I have one 5.0 server, my last, that hovers around 0.3% bufwait%). IDS 7.xx accumulates bufwaits differently but the target is to keep bufwait% between 3 and 5% with any value > about 7% indicating that there is LRU or buffer contention going on. Art S. Kagel Marty Allred wrote: > > In article <7mg41b$s4q$1@nnrp1.deja.com>, > coburn@aottsd.uscourts.gov wrote: > > Has anyone out there attempted to use the book from INFORMIX/press > > called INFORMIX DBA Survival GUIDE? Is it worth the $'s for someone > > who has taken a the most of the Informix training classes? Any > > comments? > > First, make sure you get the second edition of the book. > > You will probably have a hard time finding a serious Informix DBA that > doesn't have (or wish he had) this book. Get it. > > I've been doing Informix for about 15 years, have taken Informix > training classes, and I still find good things in the book. I do have > one complaint, though. There are a lot of places where it will say > something like "If you are experiencing excessive bufwaits, ..." without > giving any kind of clue on what would constitute "excessive". It would > really be nice to have even a ballpark estimate of what would seem like > too high a value for this and all of the other measurements. > > I guess I have two complaints. The index isn't the greatest. > > Oops, I'm going to go ahead and add this third one, even though I don't > that Joe had any real say in the matter and it's one of those "don't > look a gift horse in the mouth" kind of things. The CBT course "chosen > especially for this book" that comes free on the included CD is "MS > Exchange Server 5.5:Connectors". It seems to me like they could have put > a little more effort into finding a relevant course to include. I won't > even bother looking at this one to see if I might want to try any of > their other courses. > > -- > --Marty Allred > > Sent via Deja.com http://www.deja.com/ > Share what you know. Learn what you don't.
Art S. Kagel <kagel@bloomberg.net> wrote: : I'm going to presume to speak for Joe here since I tech reviewed the : book for him. Excessive bufwaits is defined as a bufwait percent of : more than a few percentage points. Percent of what? Here is the : formula: Well, I guess I'm going to have to claim success for the book. Art Kagel finally admitted publicly that he was my tech reader! ;) Good info on the bufwaits statistics, Art. RE: the CBT course. The CBT course on the CDROM is the correct one. The page describing the course was the wrong one. If you go into the CDROM, you'll find one very good Informix course. The wrong insert page will be corrected in the next printing. Hope to see y'all at the User Conference in San Diego. I'm doing a session on the Survival Guide on Thursday at 4:00 (talk about a prime spot.... I always wanted to be the last presentation in a 4-day show! Joe
For those of us who will not be able to attend, could someone post some notes / happenings to c.d.i.? John Carlson Informix DBA WHSmith USA Joe Lumbley wrote: > > Art S. Kagel <kagel@bloomberg.net> wrote: > : I'm going to presume to speak for Joe here since I tech reviewed the > : book for him. Excessive bufwaits is defined as a bufwait percent of > : more than a few percentage points. Percent of what? Here is the > : formula: > > Well, I guess I'm going to have to claim success for the book. Art Kagel > finally admitted publicly that he was my tech reader! ;) > > Good info on the bufwaits statistics, Art. RE: the CBT course. The CBT > course on the CDROM is the correct one. The page describing the course was > the wrong one. If you go into the CDROM, you'll find one very good Informix > course. The wrong insert page will be corrected in the next printing. > > Hope to see y'all at the User Conference in San Diego. I'm doing a > session on the Survival Guide on Thursday at 4:00 (talk about a prime spot.... > I always wanted to be the last presentation in a 4-day show! > > Joe