sleek citrus
Hello guys
We have some strange IO numbers in the volume dashboards, where a volume has over 350k IOPs which cannot be true, compare to the system manager. Interesting is that after the upgrade it looks fine and after one hour there is that increase. Current harvest version is 24.11 and it looks the same with 25.02. Interesting is, it is only one trident lun pool and not all of them. I will generate the logs tomorrow and will send them
twin anchor
@sleek citrus That is strange. Could you please upload Harvest logs @https://upload.nabox.org/metu-saro-puxa
How do other volume metrics such as volume_read_ops, volume_write_ops, and volume_other_ops behave in this instance?
Do the output values of these metrics match the results obtained from the ONTAP CLI in diagnostic mode? Replace VOLUME_NAME as needed and use the commands below:
set d
statistics show-periodic -object volume -counter read_ops|write_ops|other_ops|total_ops -instance VOLUME_NAME -interval 60 -iterations 50
Hi @twin anchor very strange the read,write looks fine but other iops is very high
twin anchor
Okay. Let's monitor the other_ops metric using the ONTAP CLI command that was shared earlier.
sleek citrus
other total write Complete Number of
ops read_ops ops ops Aggregation Constituents
Minimums:
383764 34 384173 303 - -
Averages for 3 samples:
385096 54 385467 309 - -
Maximums:
386054 67 386369 313 - -
similar I will open a case
twin anchor
Yes, seems like ONTAP related.
sleek citrus
Case is open 2010361754
steep jolt
what ontap version are you running?
sleek citrus
9.15.1 P8 now
sleek citrus
It is a bug in 9.15.1 combined with the Vino platform
plain canyon
Yeah was gonna say, is this an A70 or something?
sleek citrus
Fix should come with P11 around June.
cursive haven
I mean if support referred to it, looks like the information has been hidden since I last looked at it.
cursive haven
^ The bug mentioned has been unlocked now and again shows some information.