We have hooked into 3 datacore arrays.
STOR2RRD is showing latencies of over 1000ms, sometimes peaks of 1200ms. on the "volume response time" charts.
All 3 arrays have all-flash disk and have tested ultra-low latency using performance testing tools (sub 1ms sustained)
The vCenter graphs we pull show latency around 1-3ms sustained throughout the day which does not match the
STOR2RRD graphs.
We have additionally submitted performance reports to Datacore for analysis and we have all come to the joint
conclusion that STOR2RRD is just reporting/graphing incorrectly.
Please advise on how we can help STOR2RRD correct this.
Comments
I apologize for the delay. Can you try this debug script?
https://download.stor2rrd.com/patch/datacoreperf.pl.gz
Gunzip it and copy to /home/stor2rrd/stor2rrd/bin (755, stor2rrd owner)
-rwxr-xr-x 1 stor2rrd stor2rrd 133948 5. úno 13.59 datacoreperf.pl
If your web browser gunzips it automatically then just rename it: mv datacoreperf.pl.gz datacoreperf.pl
after copy script to bin directory then
cd /home/stor2rrd/stor2rrd # or where is your STOR2RRD working dir
echo "export DATACORE_DEBUG=1" >> etc/.magic
wait about 2 hours then send us logs
cd /home/stor2rrd/stor2rrd # or where is your STOR2RRD working dir
tar cvhf logs.tar logs tmp/*txt
gzip -9 logs.tar
Send us logs.tar.gz via https://upload.stor2rrd.com
You might even attach screenshots when it helps in understanding of the issue.
Note (after you send the logs.)
you remove line "export DATACORE_DEBUG=1" from etc/.magic
Thank you