btscanner running v/s ready

Hi, Where could I find info on "btscanner 0" seen on the "onstat -g act" and "onstat -g rea" output. I would like to know more info and the difference, correlation as to when it becomes ready and when it becomes running. === IBM Informix Dynamic Server Version 9.40.FC3 -- On-Line -- Up 4 days 00:51:03 -- 1285976 Kbytes Ready threads: <<<<tid tcb rstcb prty status vp-class name 133 c00000003ddbfdb0 c00000003c41bdb8 1 ready 1cpu btscanner 0 ===== IBM Informix Dynamic Server Version 9.40.FC3 -- On-Line -- Up 4 days 16:20:59 -- 1285976 Kbytes Running threads: <<<tid tcb rstcb prty status vp-class name 133 c00000003ddbfdb0 c00000003c41bdb8 1 running 3cpu btscanner 0 === This question may be considered general for any IDS version, but my specific version is IDS 9.40.FC3 running on HP-UX 11.11 on HP9000. Thank you, Yash

The onstat -C report and its options report on the btscanner threads and the table/index hot lists that it processes. As to when it runs, it runs when there is sufficient work in the hot lists for it to process. When the thread finds no more work it sleeps and wakes periodically to check the lists again. Sessions place indexes on the hot list when they encounter deleted keys in an index. When any thread in Informix wakes the thread scheduler places the thread in the ready queue (onstat -g rea) and as soon as a CPU VP is available it is moved to the active queue (onstat -g act) where it remains until it finishes what it has to do, has to wait for a resource (lock, memory latch, critical section condition, etc.) at which point it is put to sleep for a period of time. Art Art S. Kagel, Principal Consultant Advanced DataTools (www.advancedatatools.com) Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. 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 Mon, Oct 7, 2013 at 2:02 PM, YASH DAVE <ydave@shoppersdrugmart.ca> wrote: > Hi, > > Where could I find info on "btscanner 0" seen on the "onstat -g act" and > "onstat -g rea" output. > > I would like to know more info and the difference, correlation as to when > it > becomes ready and when it becomes running. > > === > IBM Informix Dynamic Server Version 9.40.FC3 -- On-Line -- Up 4 days > 00:51:03 > -- 1285976 Kbytes > > Ready threads: <<<<> tid tcb rstcb prty status vp-class name > 133 c00000003ddbfdb0 c00000003c41bdb8 1 ready 1cpu btscanner 0 > ===== > IBM Informix Dynamic Server Version 9.40.FC3 -- On-Line -- Up 4 days > 16:20:59 > -- 1285976 Kbytes > > Running threads: <<<> tid tcb rstcb prty status vp-class name > 133 c00000003ddbfdb0 c00000003c41bdb8 1 running 3cpu btscanner 0 > === > > This question may be considered general for any IDS version, but my > specific > version is IDS 9.40.FC3 running on HP-UX 11.11 on HP9000. > > Thank you, > > Yash > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --001a11c3eee4d6350704e82a9197

Thank you Art for the quick response and great answer. You are a great help as always. I can understand this output better now. I did check the "onstat -C hot", "onstat -C clean" as you had mentioned. I have two more related questions? [1] If I am capturing "onstat -g rea" and "onstat -g act" every five minutes for monitoring and seeing in different 5 min. time captures that there are "ready" threads, but I don't get as many "running" threads, Would it be safe assumption that those "ready" become "running" and those "running" complete quickly before a capture gets to capture it? [2] If 7 days ago, I captured less "Ready" and "Running" than today, would it be safe to assume that my app. performed more deletion that btscanner has to do more cleaning? Thank you, Yash

Yash: onstat -g rea and -g act are VERY dynamic. If sessions are staying ineither queue for more than a couple of seconds at a time there is usually a problem (OK except during the processing of very complex queries). When I want to monitor how well a server is moving data out to users and get a quick sense of whether there are any bottlenecks in the CPU VPs, I will run onstat -g rea -r 1 and I want to see that no session stays in the readyqueue for more than two iterations meaning that it was waiting less than 2 seconds before being granted a CPU Vp to run on. Similarly, I will also run onstat -g act -r 1 |grep sqlexec to make sure that I'm not seeing any sessions that are spinning CPU for more than a few seconds (in an OLTP environment anyway). If your btscanner threads are always active, then you may want to schedule the indexes that are always hottest for manual rebuilds periodically with an appropriate FILLFACTOR set just to free up the btscanners so they can optimizer the less active indexes better. Art Art S. Kagel, Principal Consultant Advanced DataTools (www.advancedatatools.com) Blog: http://informix-myview.blogspot.com/ Disclaimer: Please keep in mind that my own opinions are my own opinions and do not reflect on my employer, Advanced DataTools, the IIUG, nor any other organization with which I am associated either explicitly, implicitly, or by inference. 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 Mon, Oct 7, 2013 at 2:29 PM, YASH DAVE <ydave@shoppersdrugmart.ca> wrote: > Thank you Art for the quick response and great answer. You are a great > help as > always. I can understand this output better now. > > I did check the "onstat -C hot", "onstat -C clean" as you had mentioned. > > I have two more related questions? > > [1] > > If I am capturing "onstat -g rea" and > > "onstat -g act" every five minutes for monitoring > > and seeing in different 5 min. time captures that there are "ready" > threads, > > but I don't get as many "running" threads, > > Would it be safe assumption that those "ready" become "running" and > > those "running" complete quickly before a capture gets to capture it? > > [2] If 7 days ago, I captured less "Ready" and "Running" than today, would > it > be safe to assume that my app. performed more deletion that btscanner has > to > do more cleaning? > > Thank you, > > Yash > > > > ******************************************************************************* > Forum Note: Use "Reply" to post a response in the discussion forum. > > --001a11c34e6cf4036204e82bb8f9

Thank you Art. I believe the active (running) threads also don't run too long, they just get captured by the monitoring as a coincidence. I will monitor for the behaviour that you explained to ensure the active threads of btscanner don't run too long.