# Insights Cloud Edition 2.0.8

2021-12-01

## Object Store

[FIX] Creation of an inventory file will help speed up hdb load times.

aws --output json s3api list-objects --bucket kxtaq --prefix 'data' --query 'Contents[].{Key: Key, Size: Size}' | gzip > all.json.gz

Upload to the bucket itself via

aws s3 cp all.json.gz s3://kxtaq/_inventory/all.json.gz

User can control which file is used as inventory via env var

export KX_OBJSTR_INVENTORY_FILE=_inventory/myprivateinv.json.gz

The reading of the inventory file bypasses the cache, and to avoid cache invalidation issues, is not readable explicitly.

The file must be gzipped json, as an array of {Key:string,Size:int} objects.

[FIX] Renamed vendor neutral env vars using S3->OBJSTR.

KX_S3_CACHE_PATH->KX_OBJSTR_CACHE_PATH
KX_TRACE_S3->KX_TRACE_OBJSTR

[FIX] A memory corruption issue has been fixed in the kxreaper caching process. When the same cached file is accessed twice or more in a row, and then eventually when that file is squeezed out of the cache due to size constraints on the dir, it will double free/crash/abort. This scenario is more likely to be hit when the cache size is small(e.g. 10MB) and can only hold a few files.

## REST Client

[FIX] Azure metadata registration can return non json messages if there are no service accounts. This would cause an error as shown below

q).j.k "getting assigned identities for pod default/insights-discovery-proxy-79f67f865-c9zn5"
'illegal char g at 0
[0]  .j.k "getting assigned identities for pod default/insights-discovery-proxy-79f67f865-c9zn5"

The library has been updated to handle error conditions like this and log the errors as debug messages.

## Known Issues

• KX_S3_CACHE_SIZE is still active. This will also be renamed in an upcoming release.