more quick actions
minimize quick actions
Topic starter
Ciao a tutti!
I'm analyzing a new performance problem over a big SQL Server HDS DB (about 500 GB) that has runs fine for about 1 year. In this DB hundreds of events/seconds are recorded. I need infos about the meaning of 2 kind of rows I found many times in the HDS.log:
----------------------------------------------------------------------------
1) TB_SCRL is the name of a table, and I found these rows:
2015/05/12,13:59:03.796,GPMSCADA,3GPM_SCADA,HDS,E,0,0,CSchTable::UpdateBatch __ Call TreatComError ID = 'TB_SCRL' iTest = 2
2) I have this message, but I don't see a DEFRAG action configured in the Maintenance Plan... ?? Which project files I need to check to exclude a dis-alignment between the PcVue GUI and HDS?
2015/05/12,14:00:00.772,GPMSCADA,3GPM_SCADA,HDS,I,0,0,Execute Scheduler Task HDS_DB_EVENT_IDXDEFRAG on Database DB_EVENT
2015/05/12,14:00:00.773,GPMSCADA,3GPM_SCADA,HDS,E,0,0,CHisSrvExploit::ExecuteScheduledTask __ Execution of scheduled maintenance task is obsolet. Upgrade your project and use only scheduled maintenance plan.
----------------------------------------------------------------------------
My suspect is that the IT department who is managing this SQL Server machine (separated from the PcVue one) is making something that is slowing down our HDS recordings (infact I have many messages like 2015/05/15,12:14:15.763,GPMSCADA,3GPM_SCADA,HDS,E,0,0,OPCCallback::OnDataChange __ HDS is no more saturated.
2015/05/15,12:14:15.764,GPMSCADA,3GPM_SCADA,HDS,E,0,0,OPCCallback::OnDataChange __ HDS can't treat all the events. Try to reduce the number of information recorded by your application.), but I want to be sure also about the previos two points. I already played a lot with HDS.INI by increasing DefaultMaxBufferedEvents to 300'000 and after the system has runs fine for about 1 year.
PcVue version is 11.1 Release [11.1.5002.11930 - 2015/02/05]
Thankyou!
Posted : 15/05/2015 7:16 pm
Ciao Filippo,
I also suspect something is bloking access to tables. UpdateBatch failed means probably we can't write the value in the table.
Regarding your second point and maintenance plan you will find all exeplanation in kb article 435.
Good luck with this enormous database 😛
Posted : 16/05/2015 12:39 am
Hi Filippo,
For the MAintenance Plan issue just remove the file Scheduler.xml
About the UpdateBatch error I have no idea.
About the Saturated HDS issue this is unfortunately a classic one after long-run. DB size is increasing constantly and the index fragmentation too. Did you (or the IT) defragmenting sometimes?
For analyzing the issue I recommend configuring the System.HDS.PendingRecords in trend in a proprietary archive unit. Keep a filtering period of about 10s otherwise you will have too many points.
Then after few days you can display the trend and see if the issue occured always on the same time or randomly. If it's always on the same time then you could ask the IT what operation they are doing at that time.
That's the first troubleshooting step I recommend.
Nico
PS: For Ludo, 500GB is not enormous actually. It's more the standard size we will have. Several project here in Asia have more than 2TB.... :whistle:
Posted : 18/05/2015 7:53 am