Thanks, already rewriting my import client to support multiple threads.
With ITSM 7 I can see why BMC added this feature :-) -- Jarl On 2/8/07, Ben Chernys <[EMAIL PROTECTED]> wrote:
Server threads. Though if your client is single-threaded, the server will only allocate one thread to that client session. So yes, both would have to be multi-threaded. You can emulate this by spawning multiple instances of a single threaded client against a multi-threaded server. Again, though the improvements comes when you are creating multiple records on different tables through workflow in an out-of-sequence manor simultaenously across multiple threads. Rick is correct in that when block IDs are used the locking on the arschema goes away. In fact, a single, separate, private server thread is used for next ids when next id blocking is enabled. So, really, the feature is very useful for preventing a "deadlock" issue on arschema which is possible with the example I have given below. Cheers Ben >----- ------- Original Message ------- ----- >From: Rick Cook <[EMAIL PROTECTED]> >To: [email protected] >Sent: Thu, 8 Feb 2007 06:53:20 > >As I recall, it allows mechanisms that use a large >number of Request IDs in >a short time (like an NMS) to only go to the DB >once to grab the defined >chunk of IDs and cache them for its own use, rather >than constantly go ask >for them one at a time. > >Rick > >On 2/8/07, Jarl Grøneng <[EMAIL PROTECTED]> >wrote: >> >> Thanks, >> >> Do you mean several threads on the server, or >several treads from the >> client? I has a import program doing this, but >will rewrite it to >> support multiple threads. >> >> -- >> Jarl >> >> On 2/8/07, Ben Chernys ><[EMAIL PROTECTED]> wrote: >> > This only has effect if you are running >multiple threads. Further, the >> real use is to not place a lock on arschema so >the effects will not be >> noticed until you have workflow creating multiple >records firing >> simultaenously across multiple threads (possibly >in different orders and >> timings). >> > >> > Example: >> > Thread 1 Thread 2 >> > cre XXX cre YYY >> > wait 1s cre XXX >> > cre YYY wait 1s >> > cre XXX cre XXX >> > >> > Cheers >> > Ben >> > >> > >----- ------- Original Message ------- ----- >> > >From: =?ISO-8859-1?Q?Jarl_Gr=F8neng?= >> > ><[EMAIL PROTECTED]> >> > >To: [email protected] >> > >Sent: Thu, 8 Feb 2007 12:33:03 >> > > >> > >Hi, >> > > >> > > >> > >Has anyone notice any incrase in performance >> > >setting >> > >Next-ID-Block-Size to a value higher than 2? >> > > >> > >Tried setting this up to 1000, but does not >notice >> > >any increase in >> > >create operation. >> > > >> > >-- >> > >Jarl >> > > >> > >>__________________________________________________ >_ >> > >____________________________ >> > >UNSUBSCRIBE or access ARSlist Archives at >> > >www.arslist.org ARSlist:"Where the Answers >Are" >> > >> > >> >___________________________________________________ >____________________________ >> > UNSUBSCRIBE or access ARSlist Archives at >www.arslist.org ARSlist:"Where >> the Answers Are" >> > >> >> >> >___________________________________________________ >____________________________ >> UNSUBSCRIBE or access ARSlist Archives at >www.arslist.org ARSlist:"Where >> the Answers Are" >> > > > >-- >Rick Cook >Cook Enterprises >253-278-4112 > >___________________________________________________ >____________________________ >UNSUBSCRIBE or access ARSlist Archives at >www.arslist.org ARSlist:"Where the Answers Are" >** ><div>As I recall, it allows mechanisms that use a >large number of Request IDs in a short time (like >an NMS) to only go to the DB once to grab the >defined chunk of IDs and cache them for its own >use, rather than constantly go ask for them one at >a time. ></div> ><div></div> ><div>Rick<br></div> ><div><span class="gmail_quote">On 2/8/07, <b >class="gmail_sendername">Jarl Grøneng</b> <<a >href="mailto:[EMAIL PROTECTED]">[EMAIL PROTECTED] >mail.com</a>> wrote:</span> ><blockquote class="gmail_quote" >style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px >0.8ex; BORDER-LEFT: #ccc 1px >solid">Thanks,<br><br>Do you mean several threads >on the server, or several treads from >the<br>client? I has a import program doing this, >but will rewrite it to ><br>support multiple >threads.<br><br>--<br>Jarl<br><br>On 2/8/07, Ben >Chernys <<a >href="mailto:[EMAIL PROTECTED]">ben >[EMAIL PROTECTED]</a>> >wrote:<br>> This only has effect if you are >running multiple threads.Further, the >real use is to not place a lock on arschema so the >effects will not be noticed until you have workflow >creating multiple records firing simultaenously >across multiple threads (possibly in different >orders and timings). ><br>><br>> Example:<br>> >Thread >1&n >bsp; >Thread 2<br>> cre >XXX >&nb >sp;cre YYY<br>> >wait >1s& >nbsp;&nbs >p;cre XXX<br>> cre >YYY >&nb >sp;wait 1s<br>> >cre >XXX >&nb >sp;cre XXX ><br>><br>> Cheers<br>> Ben<br>><br>> >>----- ------- Original Message ------- >-----<br>> >>From: > =?ISO-8859-1?Q?Jarl_Gr=F8neng?=<br>> >><<a >href="mailto:[EMAIL PROTECTED]">[EMAIL PROTECTED] >MAIL.COM ></a>><br>> >>To:&n >bsp; <a >href="mailto:[email protected]";>[EMAIL PROTECTED] >RG</a><br>> >Sent: Thu, 8 Feb 2007 >12:33:03<br>> ><br>> >Hi,<br>> >><br>> ><br>> >Has anyone notice any >incrase in performance ><br>> >setting<br>> >Next-ID-Block-Size >to a value higher than 2?<br>> ><br>> >>Tried setting this up to 1000, but does not >notice<br>> >any increase in<br>> >>create operation.<br>> > ><br>> >--<br>> >Jarl<br>> >><br>> >>_______________________________________________ >____<br>> >>____________________________<br>> >>UNSUBSCRIBE or access ARSlist Archives >at<br>> ><a href="http://www.arslist.org";> >www.arslist.org</a> ARSlist:"Where the Answers >Are"<br>><br>> >___________________________________________________ >____________________________<br>> UNSUBSCRIBE or >access ARSlist Archives at <a >href="http://www.arslist.org";> >www.arslist.org</a> ARSlist:"Where the Answers >Are"<br>><br><br>__________________________ >___________________________________________________ >__<br>UNSUBSCRIBE or access ARSlist Archives at <a >href="http://www.arslist.org";> >www.arslist.org</a> ARSlist:"Where the Answers >Are"<br></blockquote></div><br><br >clear="all"><br>-- <br>Rick Cook<br>Cook >Enterprises<br>253-278-4112 >__20060125_______________________This posting was >submitted with HTML in it___ _______________________________________________________________________________ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org ARSlist:"Where the Answers Are"
_______________________________________________________________________________ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org ARSlist:"Where the Answers Are"
