more quick actions
minimize quick actions
Archive Data mgt
6
Posts
2
Users
0
Likes
38
Views
Topic starter
Hi all
What could be the good or bad effect if I increase the parameter MaxBufferedEvents (default=20000)
Is there a maximum value for this parameter
Thanks a lot
Posted : 31/05/2018 12:28 am
Hi Nico P,
In my opinion there is no really bad effect except you are consuming more memory.
Note increasing this value would solve an issue rarely but it worth to try.
Try 100,000. If you still reach this value you can consider your issue not solved
For troubleshooting I suggest the following:
1. Create a proprietary archive unit 'SYSTEM'
2. Set the variable SYSTEM.HDS.PendingRecords in trend to the SYSTEM archive unit. Period 5s.
3. Create a mimic with a trendviewer with this variable. Stay in realtime and set the period to 30mn
4. Restart PcVue and check the TrendViewer
What do you see?
Nico K
Posted : 31/05/2018 10:39 am
Topic starter
Hi Kiko K
I Have no trouble Today. The question is one from a customer who wants to know how long Pcvue can stay disconnected from the database
Posted : 31/05/2018 11:46 am
Topic starter
Hi Nico K
When you said "consuming more memory" does it mean more space disk or do you really speak of RAM ?
Posted : 31/05/2018 11:38 pm
Hi Nico P,
I Have no trouble Today. The question is one from a customer who wants to know how long Pcvue can stay disconnected from the database
Answering this question is not really possible. Look, if I have 1 point to store every 1mn I can keep the DB disconnected for 1 century!
Everything is depending on the recording flow.
By the way I think you are mixing-up the problems. The PendingRecords doesn't really affect the behaviour when you lose the DB connection. Actually when connection is lost HDS is switching in downgraded mode and store the points directly in a file. Now, according the recording flow, this file can grow quite quickly.
By experience reaching 2Gb will trigger problems when the connection will come back...
The problem is: HDS doesn't manage a limit and continue storing until...no more disk space!
That's definitively a feature we must integrate in the next version I guess.
When you said "consuming more memory" does it mean more space disk or do you really speak of RAM ?
My reply was done regarding the PendingRecords. PendingRecord is just a buffer that means if you increase the buffer size then you increase the RAM consumed and not the HD.
But, just few words about PendingRecords role (to avoid mix-up): Originally this buffer has been done to absorb a sudden peak of records (alarms peak, devices communication restored,..). Suddenly the recording flow is too high therefore HDS must store in the buffer the pending records and free it step by step the buffer.
So, it is possible for a project to have peaks generating bigger that 20,000 records in one shot. In this case you can increase this limit.
Nico K
Posted : 01/06/2018 9:19 am
Topic starter
Hi Nico K
Ok much more clear for me
Thanks a lot
See you next week
Posted : 01/06/2018 12:09 pm