Cloud Tiering Appliance VNX Atmos - Data corruption

Hi,

Newbie here on this board and hoping someone may have had similar issues.

We have a setup of a VNX5500 and a Cloud Tiering appliance that is archiving CIFS aged data from the VNX to Atmos.

We have an intermittent problem where when the CTA is under load, when data is recalled it intermittently comes back as corrupt. This mainly happens with JPGs, PDF and PPT files and initially the data looks OK but if you open the JPG for example there would be a big strip across or down the image where the image is not viewable. Like someone tore a strip across it.

We have done a number of tests and only happens when CTA is under load. 

Interestingly when we monitor the VNX logs, at the same we see the corruption happen, we also see the following in the VNX logs:

2020-06-13 06:13:37: DHSM: 4:[VDM_02] HTTP: Warning - http server responded with more than required amount of data

2020-06-13 06:13:37: DHSM: 4:[VDM_02] Ignoring the extra data

2020-06-13 06:13:37: DHSM: 4:[VDM_02] HTTP: Warning - http server responded with more than required amount of data

2020-06-13 06:13:37: DHSM: 4:[VDM_02] Ignoring the extra data

We also see the following repeated a number of times:

2020-06-13 06:13:32: DHSM: 4:[VDM_02] Offline I/O limit of 255 exceeded, returning File_Frozen

When the CTA is not under load, recalls work fine and these messages don't come up, even when data is being recalled.

Has any one experienced similar issues, seen this problem before?

I searched in regards to these errors but cannot find anything on the web around this issue.

Thanks in advance for any assistance.