One of my past systems shared the Oracle Server with the HR system (and one other thing that I can't remember). Every other Monday was a painful day for our Service Desk because of payroll. Payroll was more critical than Remedy :)
Jason On Mon, Oct 1, 2012 at 9:49 AM, Jose Manuel Huerta Guillén < arsl...@theremedyforit.com> wrote: > ** Problem discovered! > > The root cause was at the Oracle Server. The Oracle server is shared and > one application went mad. So the Oracle got stuck at sometimes, blocking > all SQL queries, but only at some moments. The result: All threads blocked, > and the rest of users waiting for a thread to be released, that means, > Oracle responded to the queries. > > The next question is Why an application can have all Oracle resources? The > answer: that application is critical, and Remedy isn't. Both system are > running for a healthcare organization, and the critical app is a HIS. > > Thank you all, > > Jose Manuel Huerta > http://theremedyforit.com/ > > > > > On Mon, Oct 1, 2012 at 5:31 PM, Misi Mladoniczky <m...@rrr.se> wrote: > >> Hi, >> >> The API-timestamps is always correct, and in this case the API-call and >> SQL-call executes very quickly, so the problem is not that the API/SQL >> executes slowly. >> >> What is your count on List threads? To seems extraordinary that you need >> to wait 30 seconds before a List thread is available. You should be >> waiting in the TC-thread and then get routed to the first availabel >> List-thread. This would mean that ALL users were tied up for 30+ seconds >> if it is your List-thread count that is too low. >> >> Best Regards - Misi, RRR AB, http://rrr.se >> >> > Jose, >> > Have you tried to run the SQL query yourself ? >> > Does it return a large number of records ? Do you perhaps have many >> > records but few distinct values ? >> > Don't trust the timestamps to much. Sometimes the log is written after >> > the command was executed. >> > >> > On Mon, Oct 1, 2012 at 3:32 PM, Longwing, LJ CTR MDA/IC >> > <lj.longwing....@mda.mil> wrote: >> >> Jose, >> >> I would agree that you were waiting on a List thread. Have you run an >> >> arloganalyzer on a larger sample of the log to determine if you are >> >> thread bound? >> >> >> >> -----Original Message----- >> >> From: Action Request System discussion list(ARSList) >> >> [mailto:arslist@ARSLIST.ORG] On Behalf Of Jose Manuel Huerta Guillén >> >> Sent: Monday, October 01, 2012 4:45 AM >> >> To: arslist@ARSLIST.ORG >> >> Subject: Analizing a client log >> >> >> >> ** Hi listers, >> >> >> >> When opening a menu we see a big delay. The log retrieved at the client >> >> is: >> >> >> >> <CLAT> /* lun oct 01 2012 12:38:23.861 */+API call, Procedure #55 >> >> <API > <TID: 0000004740> <RPC ID: 0001075159> <Queue: List > >> >> <Client-RPC: 390620 > <USER: S013802 >> >> > /* lun oct 01 2012 12:38:59.8080 */+EXPQRY ARExpandQueryMenu from >> >> Mid-tier (protocol 14) at IP address XXXXXXXXXXXXXX >> >> <SQL > <TID: 0000004740> <RPC ID: 0001075159> <Queue: List > >> >> <Client-RPC: 390620 > <USER: S013802 >> >> > /* lun oct 01 2012 12:38:59.8080 */SELECT >> >> T856.C1,C1000000009,C1000000008 FROM T856 WHERE ((T856.C1000000007 = >> >> 'Resolution Method') AND (T856.C7 = 1)) ORDER BY C1000000007 >> >> ASC,C1000000012 ASC,C1000000008 ASC, 1 ASC >> >> <SQL > <TID: 0000004740> <RPC ID: 0001075159> <Queue: List > >> >> <Client-RPC: 390620 > <USER: S013802 >> >> > /* lun oct 01 2012 12:38:59.8080 */OK >> >> <API > <TID: 0000004740> <RPC ID: 0001075159> <Queue: List > >> >> <Client-RPC: 390620 > <USER: S013802 >> >> > /* lun oct 01 2012 12:38:59.8080 */-EXPQRY OK >> >> <CLAT> /* lun oct 01 2012 12:39:00.109 */-API call, Procedure #55 >> >> >> >> >> >> What I see looking at the log is that the API call is made at 12:38:29 >> >> but the API started to process at 12:38:59, that is 30 seconds later. >> My >> >> deduction from this log is that the API call was waiting to the thread >> >> to be available. Since the CPU usage currently is normally beyond 10% >> >> and the database engine can still accept more consumption, that means >> >> that I must increase the number of LIST threads. >> >> >> >> Is this correct or am I wrong? >> >> >> >> Thanks, >> >> >> >> Jose Manuel Huerta >> >> http://theremedyforit.com/ >> >> >> >> >> >> _attend WWRUG12 www.wwrug.com ARSlist: "Where the Answers Are"_ >> >> >> >> >> _______________________________________________________________________________ >> >> UNSUBSCRIBE or access ARSlist Archives at www.arslist.org >> >> attend wwrug12 www.wwrug12.com ARSList: "Where the Answers Are" >> > >> > >> _______________________________________________________________________________ >> > UNSUBSCRIBE or access ARSlist Archives at www.arslist.org >> > attend wwrug12 www.wwrug12.com ARSList: "Where the Answers Are" >> > >> >> >> _______________________________________________________________________________ >> UNSUBSCRIBE or access ARSlist Archives at www.arslist.org >> attend wwrug12 www.wwrug12.com ARSList: "Where the Answers Are" >> > > _attend WWRUG12 www.wwrug.com ARSlist: "Where the Answers Are"_ > _______________________________________________________________________________ UNSUBSCRIBE or access ARSlist Archives at www.arslist.org attend wwrug12 www.wwrug12.com ARSList: "Where the Answers Are"
