Hi, I would ask these questions if you were doing things on jBASE level: 1) Was it a distributed file? (LIST-DISTRIB <file>) What was distribution algorithm? 2) What was kind of jBASE file before writes performed? (jstat -v) 3) Did you have any triggers / indexes? 4) How (badly) (part)file(s) are sized?
You did however testing on AIX level completely and question would be more storage related (and I am not a storage specialist :() I will try to ask guys to do similar testing and give you some "our" numbers (if they agree). Question 1: How disks are connected? (SATA / FATA / FC?) How many disk IO/sec are you able to reach? Question 2: What kind of disks you have and how many physical? (eg. 4 x 500 GB 15K RPM FC) Question 3: How many Ranks did you implement? Question 4: What is stripe size in your storage? As far as I remember guys were experimenting with that long time ago. Question 5: is this SAN shared (affected by non-T24 traffic) or exlusively used for T24? Question 6: can you afford for SAN / IBM specialist? :))))))) Can you give you good one from IBM Poland. I can not advise antyhing as this is not a field of my interest :D Please check that http://www.google.pl/url?sa=t&source=web&cd=15&ved=0CDAQFjAEOAo&url=http%3A%2F%2Fbpic.com.my%2Fxiv%2Flinks%2FSSG%2FDisk%2FBenchmarks%2FIBM%2520System%2520Storage%2520DS6800%2520and%2520DS8000.pdf&rct=j&q=disk%20storage%20ranks&ei=sod1TJ2yE8mdOOa5geYG&usg=AFQjCNHmuvFwYlFJIPbRjuvWBo9-ln0l1g&cad=rja however and "google" more! Kind regards Pawel Dnia 25-08-2010 o godz. 15:58 aft3rgl0w napisał(a): > Hello all. > > i would like to see what sort of speeds do you get generally get on a > system running jbase. more specifically, we have a power 5 570 machine > that runs AIX 5.3 TL 11 SP4 and we are running jbase 4.1.5.29 on top > of that. > the machine 8 CPUs at 1900Mhz and 16GB of memory. thats our production > environment and its attached on a IBM DS4800 storage. our jbase and > production environment are on the same filesystem that resides on 4 > physical disks of the storage which are on RAID 10. a dd if=/dev/zero > of=testfile0 bs=4096 count=1000000 gives me speeds of around and more > than 110MB/sec. > now running from jsh several SELECT statements or COUNT etc and > looking through the topas utility but also nmon, the average read disk > speeds will never go beyond 7-8MB/sec. rarely i will notice bursts of > 20MB/sec but usually it runs around 5MB/sec or even less.I understand > that theres also logical processing that needs to be done from the CPU > so its not only disks, but still i feel that 7-8MB/sec is too slow > since the system is capable of a lot more.running nmon to collect date > throughout the duration of a COB i get average read speed of about 1 > MB/sec and a max of 74MB/sec. that 74MB/sec sounds very good but its > probably at a burst time period and lasted for less than a blink of > they eye. Write speeds arent any much different either.all the speeds > i mentioned are noticed while im the only user logged on the server. > it would be nice to know what sort of speeds you are running on and if > im the only one running so slow. > Thanks. -- Please read the posting guidelines at: http://groups.google.com/group/jBASE/web/Posting%20Guidelines IMPORTANT: Type T24: at the start of the subject line for questions specific to Globus/T24 To post, send email to [email protected] To unsubscribe, send email to [email protected] For more options, visit this group at http://groups.google.com/group/jBASE?hl=en
