Hi, Daniel Thank you for your comment.
> My experience with systematic logging at command line level is that > you end up polluting your file system with (hidden) log files, and while it > may be fine for a developper it's not something you want to inflict to > your customers. Also the fixed name for the log file makes it easilly > useless, a random run in the same directory will just wipe out the > data you tried to collect in a separate process. > I think having permanent systematic logging to a fixed file is not > proper, I would rather not use that. > But since the patch is relatively simple based on existing virsh logging > code, I think this could go as a command line option for virsh, for example > --log filename > where the detailed logs can then be saved if needed when a problem occurs. > I think this would avoid the main drawbacks of your proposed patch. I agree about a command line option. So, I remaked the patch which --log option is added for virsh. How about this one? Signed-off-by: Nobuhiro Itou <[EMAIL PROTECTED]> Thanks, Nobuhiro Itou.
virsh_log.patch
Description: Binary data
-- Libvir-list mailing list Libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
