The default Isilon protocol audit logs that are collected when protocol auditing is enabled is "create, delete, rename, set_security, close". We have a requirement to enable the rest of the audit options available. The full available list of options is : close | create | delete | get_security | logoff | logon | read | rename | set_security | tree_connect | write | all
The Isilon documentation does not provide the implications (if any), when full protocol auditing is enabled. Does anybody know if the performance and storage capacity of the Isilon cluster is affected by enabling full protocol auditing?
In the immortal words of Robert Heinlein, TANSTAAFL (there ain't no such thing as a free lunch). Or in other words, because of the nature of auditing (by definition, it operates inline with client-initiated operations and therefore generates additional latency), every additional operation that is audited has a performance impact.
It is very hard to give hard and fast guidelines because the impact is highly dependent on the nature of the workflow and the capabilities of the cluster. For example, enabling full auditing on an F800 cluster with a light load will generate a much lower impact than attempting to do the same on heavily-loaded A2000 nodes.
Now, all that said, some of the audit types that are not enabled by default will be very expensive. Auditing every read, write or close will generally have significant impact and caution is advised. It will have a noticeable performance impact and it will generate significantly more audit data. It is supported, it does work, but it is not free.
DELL-Sam L
Community Manager
•
8101 Posts
•
34565 Points
912
0
Posted September 14th, 2020 16:00
Hello ronanb,
Here is a link to a KB that may help. https://dell.to/3kiTeTK