Excel Reporting Overloading Mango
-
Hi,
I am setting up an Excel Report (around 200 points) to run daily and give the one-minute average for these points.
When I run a test report, the fan on the mango increases in speed and I lose comms between the Web interface and the Mango unit. It also starts to run quite hot.
I've looked at the ma.log for clues, but this is where my expertise shows limitations. When I run the report, it says:
WARN 2018-03-13T00:00:08,235 (com.serotonin.m2m2.util.timeout.TaskRejectionHandler.rejectedTask:75) - Rejected task: NoSQL Point Value Mover because Task Queue Full
ERROR 2018-03-13T00:00:10,513 (com.infiniteautomation.nosql.MangoNoSqlBatchWriteBehindManager$StatusProvider.scheduleTimeout:727) - 1 BWB Task Failures, first is: Task Queue FullWhich is puzzling as I'm only polling about 200 points.
Thanks,
Regards,
Tom
-
Hi Tom,
Is your instance fully up to date? The last time I saw this, the user had configured their named ranges to be over a million cells long! We since got rid of the code that was blanking out those million cells inefficiently, but that was the cause. One need only define two cells to tell a named range where it starts and which direction it grows. Having a named range defined to the end of the sheet creates unnecessary overhead.
Have you attempted a simpler report and had better results? You may see an improvement if you allocate a few more MB of RAM in the /opt/mango/bin/ext-enabled/ memory script, if you have one.
-
Hi,
Thanks for your reply. I ended up rebuilding and it turns out the original Excel template was 2.3MB - the new one was 100kB!
Hence the poor mango trying to parse through 2.3MB of corrrupted Excel. It caught me out as the sites have steadily become bigger (in terms of IO) and just about handled it, so I didn't spot the corrupted file until now.
I've rebuilt the smaller sites and they are running much faster.
Kind regards,
Tom
-
Thanks for sharing the resolution!