thiell
thiell
I agree with this statement: "I think default verbosity should be set to warning or error. And we should consider using syslog instead of flat file by default."
In 1.8, the `NodeSet` class does only recognize the "," separator for node/nodeset that also implictly means "union set". We should probably support other separators for convenience, like space, tab...
``` [routes] sh-hn01: @service @sherlock:1&@service: @sherlock:1&@compute @sherlock:2&@service: @sherlock:2&@compute ``` results in: ``` clush: Unknown group source: "service" ``` Quotes don't help: ``` [routes] "sh-hn01": "@service" "@sherlock:1&@service": "@sherlock:1&@compute" "@sherlock:2&@service": "@sherlock:2&@compute" ```...
Currently, there is a mismatch between the API (eg. local/distant worker selection) and the options offered by clush. It's not great. The API could be improved for a better worker...
support password-based authentication for initial connection to nodes
It will be very interesting to set environment variables which could be set by task or worker. ``` python # Set environment variable vars = { 'PATH': "...", 'LD_LIBRARY_PATH': "..."...
For cluster of random equipments, like switches, [netmiko](https://github.com/ktbyers/netmiko) may be useful, so perhaps we should write a Worker for it.
Develop `/etc/bash_completion.d/clush` and `/etc/bash_completion.d/nodeset` to provide intelligent completion for these tools.
When using clush -u _timeout_ -b, clush doesn't display the number of nodes associated with the nodeset that timed out.
Low priority, but possible improvement when there are many group sources (we tend to have more and more group sources, currently 16 on Sherlock, 12 in yaml file + 4...