Listing In Progress Multipart Uploads

List the initiated but uncompleted multipart uploads. These can be cancelled if they are old or not in progress, see https://caringo.atlassian.net/wiki/spaces/public/pages/2443822287/Canceling+a+Multipart+Write. Or, verify health.parallelWriteTimeout in Settings Reference is set to its new default 2592000 (one month) and waits for it to expire.

Here is how:

curl -i -u ADMIN 'http://DOMAIN/?stype=all&castor-system-partnumber=0&fields=context,name,etag,tmborn,content-length,castor-system-uploadid,castor-system-partnumber,Castor-System-StreamPath&format=json&sort=tmborn:asc&size=10000'

To list the parts themselves and see how much space they are using in the response header Castor-System-Bytes-Used, use the below command:

curl -i -u ADMIN 'http://DOMAIN/?stype=all&castor-system-partnumber=*&fields=context,name,etag,tmborn,content-length,castor-system-uploadid,castor-system-partnumber,Castor-System-StreamPath&format=json&sort=tmborn:asc&du=withoutreps&size=10000'

This default for this parameter was changed from 10 to 5: health.parallelWriteGCMissCount. This needs to be configured on the CSN or SCS:

CSN: set in cluster.cfg
cat "health.parallelWriteGCMissCount = 5" >> /var/opt/caringo/netboot/content/cluster.cfg

SCS: scsctl storage config set -d "health.parallelWriteGCMissCount=5"

This also needs to be applied to all nodes actively as it is a node level config with swarmctl:

swarmctl -C health.parallelWriteGCMissCount -V 5 -a -d [Swarm Node IP]