Re: TUNING:aio parameters on AIX

Art S. Kagel <kagel@bloomberg.net> wrote in article <36002794.5F4A@bloomberg.net>... > Dave Thacker wrote: > > > > Platform: IBM Rs/6000 J30 AIX V4.1.5 > > Engine Online DSA 7.23 > > > > From much reading of vmstat, iostat, and xmperf we appear to be I/0 bound > > on our production machine. > > While investigating the problem I came across a configuration for aio > > servers. We are currently set to the default and IBM docs recommend that > > we increase them to MAX= (10* count of disks using aio) and > > MIN = (MAX * .5). I believe that OL 7.23 does use the aio libraries. Has > > someone else in an AIX shop changed this configuration? Did it help? > > WHOA HORSEY! Informix's recommendations for 7.10 was 2-3 AIO VPs per > chunk, for 7.14 they made some improvements and changed that to 1.5-2 > AIO VPs per chunk, as of 7.21 the recommendation, after another round > of optimizing AIO VPs is 1-1.5 AIO VPs per chunk (leaning toward 1), I > have not seen any recommendation for 7.3 but with the other speedups to > the 7.3 code stream I'd say that 1 AIO VP/chunk was now a maximum. I > have always found that these recommendations are MORE than sufficient. > I actually only count active chunks when determining AIO VP needs. For > example on one pair of servers I have 279 chunks and growing but many > of these are holding one months data for four years of history (ie all > September data in one dbspace made of 8-12 2GB chunks) so only one out > of twelve dbspaces ( and their chunks) are active and I do very well > with 127 AIO VPs (io/wp 0.6 -> 1.3 with most at 0.8 -> 1.1). IBM's > recommendation is excessive in the extreme. > > Run onstat -g ioq to monitor q length by chunk and onstat -g iov to > monitor the activity of the KAIO and AIO VPs (BTW you can determine > which are being used this way). If almost all of your AIO VPs are > performing at least 1.0 io/wp with many showing 1.5 or more then you > need more AIO VPs if not, not. A value of io/wp means that an AIO VP > was awakened to handled a pending IO but before it could get to the > required page another, previously busy, AIO VP finished what it was > doing and took care of it. In effect having some of your VPs below > 1.0 means you could even do without those VPs except at peak. > Conversely, if you have any AIO VP with <0.5 io/wp you can probably > reduce the number of AIO VPs accordingly since more than half the time > these VPs awaken there is nothing for them to do, they are just wasting > cycles and taking up swap space. > > If it turns out, as I expect, that the VPs are not the culprit look to > your disk farm. Do you use singleton disk drives on a single > controller? Work toward the ultimate setup RAID1+0 of at least eight > mirrored pairs, spread across at least 4 controllers (2 primary, 2 > mirror, preferably with 4 additional backup controllers handling the > same drives), 8 or 16K stripe size, 500MB-1GB cache per controller. > Expensive, but no I/O problems. Anyway RAID1+0, small stripe size. > > Post some more information about your configuration and someone may be > able to help further. > > Art S. Kagel > Actually, I believe Dave was refering to AIX kernel parameters for KAIO server threads, NOT number of Informix AIO VPs. I'm not sure on what IBM's recommendations are based, but if I recall correctly (it's been a while since I read the book) the reasoning had something to do with providing enough KAIO threads that no process requesting KAIO would get blocked waiting for an available thread, thus making the I/O not all that asynchronous. I did raise the parameters as IBM suggests on an AIX 4.1.4/Informix 7.23 server about a year ago or so. I never did any detailed benchmarks, but there was no immediately apparent performance improvement. (It didn't seem to hurt either.) I have not experimented with KAIO vs. regular Informix AIO VPs either. I have heard reports on this newsgroup in the past claiming that Informix AIO was often superior to HP/UX KAIO. I don't know if the same holds true for AIX. Has anyone done any benchmarking on this? Irwin Goldstein Objective Software Systems, Inc. http://www.objectsoft.com