Hi Phil, I have taken a few snapshots of the arserverd during the day. Looks like the memory usage is slowly creeping up. The slowdown is not too bad yet, but noticeably slower than immediately following the restart:
S UID PID PPID C PRI NI RSS SZ WCHAN STIME TTY TIME CMD S root 21007 21006 0 40 20 415296 423976 ? 05:00:39 ? 5:35 ./arserverd -s ars00srv -i /local/b S root 21007 21006 0 40 20 419392 428072 ? 05:00:39 ? 6:33 ./arserverd -s ars00srv -i /local/b S root 21007 21006 0 40 20 439880 448560 ? 05:00:39 ? 8:03 ./arserverd -s ars00srv -i /local/b S root 21007 21006 0 40 20 464456 473136 ? 05:00:39 ? 9:30 ./arserverd -s ars00srv -i /local/b S root 21007 21006 0 40 20 468552 477232 ? 05:00:39 ? 9:41 ./arserverd -s ars00srv -i /local/b S root 21007 21006 0 40 20 468552 477232 ? 05:00:39 ? 9:45 ./arserverd -s ars00srv -i /local/b S root 21007 21006 0 40 20 472648 481328 ? 05:00:39 ? 9:52 ./arserverd -s ars00srv -i /local/b Is this normal? Should the memory usage be increasing like this? The above is over a three hour period. Thanks. Larry On Fri, Dec 30, 2011 at 7:16 AM, Murnane, Phil <[email protected]> wrote: > Larry: > > How much memory is the server demon using before the slowdown and then during the slowdown? > > --Phil > > -----Original Message----- > From: Action Request System discussion list(ARSList) [mailto: [email protected]] On Behalf Of L G Robinson > Sent: Thursday, December 29, 2011 19:37 > To: [email protected] > Subject: ARS 7.6.03 performance problems > > Hi Folks, > > I am experiencing a performance problem on my 7.6.03 AR System. The problem gets worse over time. When the server is started, performance is fine. For example, it takes about one second to load a particular record in the Windows client and about two seconds to load the same record in the Mid Tier. However, if I load the same record later in the day, maybe after the server has been running for 12 hours, the time to load the same record is about twice as long. > > If I restart the server process, response time returns to the original timings. This behavior is consistent whether the system is heavily loaded or lightly loaded, but the degradation seems to happen faster if the system is heavily loaded > > The arserverd is running on a Sun SPARC-Enterprise-T5120 with 16 Gb of RAM. I have tried a variety of queue/thread configurations from minimal to way over-allocated. I have used Misi's fabulous RRR|Log tool and I see that there is absolutely no thread contention at all. > When the system is running, top shows: > > 90 processes: 86 sleeping, 3 stopped, 1 on cpu CPU states: 100% idle, 0.0% user, 0.0% kernel, 0.0% iowait, 0.0% swap > Memory: 16G real, 12G free, 509M swap in use, 11G swap free > > So it appears that I have plenty of hardware. > > I am pretty sure that it is not a database problem because I have analysed the SQL logs and the transaction times are consistent. The additional time between the short times and the long times appears to all be within the API time. That is to say, the time from beginning to end for any particular API call seems to increase while the embedded SQL transactions within the API calls stay relatively constant. > > Does anyone have any suggestions on what to look for? Thanks for any insights. > Larry > > ARS 7.6.03 > Mid-tier 7.6.04 > Solaris 10 > Oracle 11.2.0... > > Larry Robinson > Remedy Developer / Admin > NC State University _______________________________________________________________________________ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug12 www.wwrug12.com ARSList: "Where the Answers Are"
