Jared, You're right. There's a cool diagram in the Server Concepts manual. So back to the original issue, scalabilty could be affected in a dedicated server configuration depending on how many files needed to be opened. I guess the problem could be mitigated by fewer/larger datafiles and/or MTS
Cheers On Fri, 29 Nov 2002, Jared Still wrote: > On Friday 29 November 2002 08:43, Jeff Herrick wrote: > > My understanding > > from the question was that he was wondering whether each > > user's process in a dedicated-server configuration opened > > all of the datafiles too > > Maybe not all of the data files, but the users dedicated server > process will open datafiles as needed to read data into the > block buffer. > > Now I don't know if I've helped any, or just added to the confusion. > > Jared > > > ....but I might have mis-understood the question. > > > > On Fri, 29 Nov 2002, Jeremiah Wilton wrote: > > > On Fri, 29 Nov 2002, Jeff Herrick wrote: > > > > None...only the oracle background processes (threads in Winblows) > > > > access the datafiles/logfiles etc. All other communication is > > > > done through the SGA. On some Unix variants you _can_ reach > > > > a file_open max kernel parameter because each process (in a > > > > dedicated server scenario) opens it's own stdin/stdout/stderr. > > > > I guess the same could be true of processes running under > > > > windows too. So in the limit...you could hit a wall but only > > > > due to the per-process overhead. > > > > > > Uh, I'm probably not going to be the only one to point out this isn't > > > true. I don't know about Win32 thread architecture, but in Unix and > > > unix-like operating systems, the shadow (server) processes each open > > > whatever files they need for write. It is true that they also open > > > the shared memory segments in order to write and read from the SGA, > > > but they do the reading from disk. Otherwise, which process do you > > > think is reading from the datafiles? > > > > [snip] > > > > > > On Fri, 29 Nov 2002, Grant Allen wrote: > > > > > Saw an interesting post in comp.databases.oracle.server postulating > > > > > that if a shadow thread needed an open file handle on all files in a > > > > > instance (or even some of them), the process handle limit in windows > > > > > could constrain user scalability (e.g. too many users would result in > > > > > ora-12500 unable to spawn errors and the like). (Let's ignore > > > > > MTS/shared server mode for the moment) > > > > > > > > > > Sounded interesting, but I thought I'd ask if anyone knows whether a > > > > > shadow thread (or process under unix) does open a handle on each file > > > > > (control, data, redo), some of them, or none of them? > -- > Please see the official ORACLE-L FAQ: http://www.orafaq.com > -- > Author: Jared Still > INET: [EMAIL PROTECTED] > > Fat City Network Services -- 858-538-5051 http://www.fatcity.com > San Diego, California -- Mailing list and web hosting services > --------------------------------------------------------------------- > To REMOVE yourself from this mailing list, send an E-Mail message > to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in > the message BODY, include a line containing: UNSUB ORACLE-L > (or the name of mailing list you want to be removed from). You may > also send the HELP command for other information (like subscribing). > -- Please see the official ORACLE-L FAQ: http://www.orafaq.com -- Author: Jeff Herrick INET: [EMAIL PROTECTED] Fat City Network Services -- 858-538-5051 http://www.fatcity.com San Diego, California -- Mailing list and web hosting services --------------------------------------------------------------------- To REMOVE yourself from this mailing list, send an E-Mail message to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in the message BODY, include a line containing: UNSUB ORACLE-L (or the name of mailing list you want to be removed from). You may also send the HELP command for other information (like subscribing).