Run Show INNODB status. Look at ---------------------- BUFFER POOL AND MEMORY ---------------------- Total memory allocated 1299859045; in additional pool allocated 6113152 Buffer pool size 71936 Free buffers 59 Database pages 70898 Modified db pages 57113 Pending reads 1 Pending writes: LRU 0, flush list 0, single page 0 Pages read 379011342, created 2581822, written 233133461 58.62 reads/s, 0.12 creates/s, 61.24 writes/s ==> Buffer pool hit rate 981 / 1000
> -----Original Message----- > From: Emmett Bishop [mailto:[EMAIL PROTECTED] > Sent: Friday, April 23, 2004 8:01 AM > To: Dathan Vance Pattishall > Subject: RE: InnoDB Load Problem > > I've been keeping tabs on this thread and would just > like to know how to tell what the buffer pool ratio > is. What is it a ratio of? What command do I run to > take a look at it? > > Thanks, > > Tripp > > --- Dathan Vance Pattishall <[EMAIL PROTECTED]> > wrote: > > Look at your fsync stat and your buffer pool ratio. > > You may get better > > performance out of use O_DIRECT since it does not > > double buffer your log > > writes. > > > > Next make sure your buffer pool ratio is close to 1 > > (100%), if not raise > > your bugger pool if you can. Additionally make sure > > you transaction logs are > > large like 1/2 your buffer pool. Also note if your > > doing many fast small > > queries set innodb_thread_conncurency high (cpu+ > > number of disk)*2 > > > > > > For the hardware portion, you might need to use > > elvtune to get better > > throughput for your hard drive or update the kernel > > to a kernel that > > supports better interaction with your hardware > > makeup. > > > > This all assumes that your queries are already > > optimized. > > > > -- > > DVP > > > > > -----Original Message----- > > > From: Marvin Wright > > [mailto:[EMAIL PROTECTED] > > > Sent: Tuesday, April 20, 2004 5:13 AM > > > To: Mechain Marc; Marvin Wright; Dathan Vance > > Pattishall; > > > [EMAIL PROTECTED] > > > Subject: RE: InnoDB Load Problem > > > > > > Hi, > > > > > > To put the unique index on like you suggest is > > fine for this table but > > > this > > > table is just the top level of a hierarchy. > > > > > > table a has 1 record > > > table b has 100's of records linked to 1 table a > > record > > > table c has 100's of records linked to 1 table b > > record > > > > > > All the records in table b and c would need to be > > updated/deleted for a > > > new > > > record. > > > It think this would be very time consuming, and > > the clients that are > > > inserting are public internet users therefore I'd > > rather not slow these > > > down. > > > > > > > > > under load iostat -x 1 gives me this > > > > > > avg-cpu: %user %nice %sys %idle > > > 38.50 0.00 18.00 43.50 > > > > > > Device: rrqm/s wrqm/s r/s w/s rsec/s > > wsec/s rkB/s wkB/s > > > avgrq-sz avgqu-sz await svctm %util > > > /dev/hda 104.00 552.00 31.00 39.00 1088.00 > > 4728.00 544.00 2364.00 > > > 83.09 62.20 1174.29 141.43 99.00 > > > /dev/hda1 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > /dev/hda2 104.00 552.00 31.00 39.00 1088.00 > > 4728.00 544.00 2364.00 > > > 83.09 82.20 1174.29 75.71 53.00 > > > /dev/hda3 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > /dev/hda5 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > > > > avg-cpu: %user %nice %sys %idle > > > 44.50 0.00 16.50 39.00 > > > > > > Device: rrqm/s wrqm/s r/s w/s rsec/s > > wsec/s rkB/s wkB/s > > > avgrq-sz avgqu-sz await svctm %util > > > /dev/hda 6.00 838.00 1.00 58.00 64.00 > > 7168.00 32.00 3584.00 > > > 122.58 3.30 393.22 169.49 100.00 > > > /dev/hda1 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > /dev/hda2 6.00 838.00 1.00 58.00 64.00 > > 7168.00 32.00 3584.00 > > > 122.58 23.30 393.22 23.73 14.00 > > > /dev/hda3 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > /dev/hda5 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > > > > avg-cpu: %user %nice %sys %idle > > > 2.00 0.00 0.00 98.00 > > > > > > Device: rrqm/s wrqm/s r/s w/s rsec/s > > wsec/s rkB/s wkB/s > > > avgrq-sz avgqu-sz await svctm %util > > > /dev/hda 195.00 162.00 58.00 8.00 2080.00 > > 1392.00 1040.00 696.00 > > > 52.61 44.40 740.91 128.79 85.00 > > > /dev/hda1 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > /dev/hda2 195.00 162.00 58.00 8.00 2080.00 > > 1392.00 1040.00 696.00 > > > 52.61 64.40 740.91 151.52 100.00 > > > /dev/hda3 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > /dev/hda5 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > > > > avg-cpu: %user %nice %sys %idle > > > 8.00 0.00 3.00 89.00 > > > > > > Device: rrqm/s wrqm/s r/s w/s rsec/s > > wsec/s rkB/s wkB/s > > > avgrq-sz avgqu-sz await svctm %util > > > /dev/hda 174.00 0.00 60.00 5.00 1856.00 > > 8.00 928.00 4.00 > > > 28.68 50.00 1235.38 147.69 96.00 > > > /dev/hda1 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > /dev/hda2 174.00 0.00 60.00 5.00 1856.00 > > 8.00 928.00 4.00 > > > 28.68 70.00 1235.38 153.85 100.00 > > > /dev/hda3 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > /dev/hda5 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > > > > avg-cpu: %user %nice %sys %idle > > > 29.50 0.00 16.50 54.00 > > > > > > Device: rrqm/s wrqm/s r/s w/s rsec/s > > wsec/s rkB/s wkB/s > > > avgrq-sz avgqu-sz await svctm %util > > > /dev/hda 102.00 71.00 40.00 6.00 1088.00 > > 616.00 544.00 308.00 > > > 37.04 5.60 671.74 193.48 89.00 > > > /dev/hda1 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > /dev/hda2 102.00 71.00 40.00 6.00 1088.00 > > 616.00 544.00 308.00 > > > 37.04 25.60 671.74 163.04 75.00 > > > /dev/hda3 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > /dev/hda5 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > > > > avg-cpu: %user %nice %sys %idle > > > 57.50 0.00 20.00 22.50 > > > > > > Device: rrqm/s wrqm/s r/s w/s rsec/s > > wsec/s rkB/s wkB/s > > > avgrq-sz avgqu-sz await svctm %util > > > /dev/hda 0.00 398.00 0.00 28.00 0.00 > > 3408.00 0.00 1704.00 > > > 121.71 42949657.76 171.43 357.14 100.00 > > > /dev/hda1 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > /dev/hda2 0.00 398.00 0.00 28.00 0.00 > > 3408.00 0.00 1704.00 > > > 121.71 4.80 171.43 14.29 4.00 > > > /dev/hda3 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > /dev/hda5 0.00 0.00 0.00 0.00 0.00 > > 0.00 0.00 0.00 > > > 0.00 0.00 0.00 0.00 0.00 > > > > > > avg-cpu: %user %nice %sys %idle > > > 39.00 0.00 9.50 51.50 > > > > > > Device: rrqm/s wrqm/s r/s w/s rsec/s > > wsec/s rkB/s wkB/s > > > === message truncated === > > > > > > __________________________________ > Do you Yahoo!? > Yahoo! Photos: High-quality 4x6 digital prints for 25¢ > http://photos.yahoo.com/ph/print_splash -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]