Odd Performance
IDS 10.00.UC8 Solaris 10 I am playing with our new Solaris server - I just increased the VPCLASS parameter to VPCLASS cpu,num=32,aff=64-95,noage from cpu,num=5,aff=13-21,noage. Our new server has 132 virtual processors available: uname -X NumCPU = 128 psrinfo -pv The physical processor has 64 virtual processors (0-63) UltraSPARC-T2+ (cpuid 0 clock 1415 MHz) The physical processor has 64 virtual processors (64-127) UltraSPARC-T2+ (cpuid 64 clock 1415 MHz) Now I am loading a table with HPL in deluxe mode (the table has a blobspace blob) into a non-logged database. The table has ~3.5 million rows and the rowsize is 165 bytes. When the load begins it is inserting ~1,000 rows per second. After ~1.5 million rows are loaded the load rate begins to crawl, down to 100 rows per second. No other activity is occuring on the server; I even began a second load of a larger table (express mode) and it ran very fast - (while the second load was running the first load's insert rate was not affected, it stayed at 100 rows per second). When I look at active threads I rarely see any kaio threads running - instead there is a pl_wkr thread that is almost constantly running - for what that's worth. Checkpoint freq. and duration are fine, every 5 minutes, 1 second long. What are the onstats I can look at to figure out what is causing this mid-load slowdown? Signed, Performance Challenged in KC
Get off of the T2 (Niagra2) processors. They don't work well with IDS. That's all I'm willing to say here. Art Art S. Kagel Oninit (www.oninit.com) IIUG Board of Directors (art@iiug.org) Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Oninit, the IIUG, nor any other organization with which I am associated either explicitly or implicitly. Neither do those opinions reflect those of other individuals affiliated with any entity with which I am affiliated nor those of the entities themselves. On Thu, Jul 9, 2009 at 11:09 AM, MIKE MAGIE <jmmagie@yahoo.com> wrote: > IDS 10.00.UC8 > Solaris 10 > > I am playing with our new Solaris server - I just increased the VPCLASS > parameter to VPCLASS cpu,num=32,aff=64-95,noage from > cpu,num=5,aff=13-21,noage. Our new server has 132 virtual processors > available: > > uname -X > > NumCPU = 128 > > psrinfo -pv > The physical processor has 64 virtual processors (0-63) > UltraSPARC-T2+ (cpuid 0 clock 1415 MHz) > The physical processor has 64 virtual processors (64-127) > UltraSPARC-T2+ (cpuid 64 clock 1415 MHz) > > Now I am loading a table with HPL in deluxe mode (the table has a blobspace > blob) into a non-logged database. The table has ~3.5 million rows and the > rowsize is 165 bytes. When the load begins it is inserting ~1,000 rows per > second. After ~1.5 million rows are loaded the load rate begins to crawl, > down > to 100 rows per second. No other activity is occuring on the server; I even > began a second load of a larger table (express mode) and it ran very fast - > (while the second load was running the first load's insert rate was not > affected, it stayed at 100 rows per second). When I look at active threads > I > rarely see any kaio threads running - instead there is a pl_wkr thread that > is > almost constantly running - for what that's worth. Checkpoint freq. and > duration are fine, every 5 minutes, 1 second long. What are the onstats I > can > look at to figure out what is causing this mid-load slowdown? > > Signed, > > Performance Challenged in KC > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --001636c5adf4d5e7a2046e481dc2
I wish it were that easy :-) Contractors for the FED have little say in regards to purchasing decisions... Thanks for the feedback though, as always. Mike