NetApp’s Active IQ interfaces and tools are fantastic for gaining an
understanding of customer systems. However, sometimes you don’t have access to a
customer’s autosupport data or the customer has gone to a 3rd party support
situation. In that case I’ll request a full body.7z weekly autosupport from each
NetApp node in their enviroment. The script asup_parser.py will parse many of
the xml files within and produce a single Excel spreadsheet of their contents.
The spreadsheet consists of individual tabs with details extracted from their xml sources and are automatically placed into Excel tables with autofiltering enabled. Each table has, at a minimum, counts of rows currently visible. Several tables have sum calculations of visible fields as well. So in this way if you need to a summary of the used space of two or more customer data SVMs in the volumes list, you can easily calculate that. ONTAP values are in raw bytes so it’s fairly straightforward to add an additional calculation against the sums to get GB, GiB, etc.
Each generated tab also contains a list of the fields included in the XML file and the longer UI names for each field . I do not extract every single field in every xml.
This uses python3 . In addition to standard modules the following modules via
pip or easy_install
After obtaining a body.7z create a directory typically named for the node,
navigate into it, and 7z x ../body.yz to extract it into the current
directory. You can then run the script like so:
/path/to/asup_parser.py --source . --destination ../nodename -t /tmp If all is well this will exit back to the prompt. The generated file will be
one directory level up from the extracted 7zip and named nodename.xlsx .
Additionally the /tmp directory will contain various nodename_xmlname.csv
files and text files that are an interim step used to create the final Excel. I
do keep these intentionally
The script does attempt to handle missing files in the body.7z extraction gracefully. You will see warnings and the Excel will have tabs with empty tables. Fields that don’t exist in the version of ONTAP you are viewing will simply be blank in the output tables. It should be apparent at this point that if you have legacy 7mode data, this script won’t work at all.
Note that while the script is run against a single node’s autosupport payload, some of the items are cluster-scoped and some by their nature are node-scoped. Things like volumes and luns are only shown for the relevant node. While things like cifs shares, vservers, etc are cluster wide. Therefore when you want to look at two nodes or a larger cluster, you’ll need to hand-combine some things for now to get a comprehensive view in a single spreadsheet.
The spreadsheet tables are not automatically sized to match the field contents (eg ‘auto column width’) because the XlsxWriter python module doesn’t do that.
I do generate interim CSV files and I keep those. When I was first worked with this I was just doing CSVs and while I could refactor to skip this entirely, I still find having those rawer contents available handy for other potential automation tasks and the like.
The tabs list in the middle of the Excel is easy enough to understand. Each
tab maps to an XML file name in the asup payload. It also maps to a key in the
large tabdetails dict. To get the field contents from an xml, you would look
at an XMLs contents and update the dict or create a new dict entry accordingly.
The version of ONTAP I used to create this initial product is 9.5.
It should be possible to process multiple subdirectories containing extracted body.7z content. In that way the loop could combine node specific items such as volumes and luns into single tables and create a cluster wide xml in one go.
I deliberately do not have python extract 7z files via but that is technically
possible as well. The weekly autosupport emails, however, will typically have
them all coming from different nodes but with the same body.7z attachment name.