On Wed, Jul 26, 2017 at 16:46:23 -0500, Eric Blake wrote:
On 07/26/2017 05:00 AM, Peter Krempa wrote:
Add the blockstats data and fix the expected output. --- ...emumonitorjson-nodename-gluster-blockstats.json | 111 +++++++++++++++++++++ ...mumonitorjson-nodename-gluster-named-nodes.json | 90 ++++++++--------- .../qemumonitorjson-nodename-gluster.result | 15 ++- tests/qemumonitorjsontest.c | 2 +- 4 files changed, 167 insertions(+), 51 deletions(-) create mode 100644 tests/qemumonitorjsondata/qemumonitorjson-nodename-gluster-blockstats.json
Reviewed-by: Eric Blake <eblake@redhat.com>
diff --git a/tests/qemumonitorjsondata/qemumonitorjson-nodename-gluster-blockstats.json b/tests/qemumonitorjsondata/qemumonitorjson-nodename-gluster-blockstats.json new file mode 100644 index 000000000..665a6b05b --- /dev/null +++ b/tests/qemumonitorjsondata/qemumonitorjson-nodename-gluster-blockstats.json @@ -0,0 +1,111 @@ +[ + { + "device": "drive-virtio-disk0", + "parent": { + "stats": { + "flush_total_time_ns": 0,
This output was obviously generated by loading a simple <domain> XML and then getting query-blockstats output. I'm wondering, if for reproducibility (modulo the change of random nodename suffixes), you should enhance the commit message here (and in similar other patches) to call out the steps you used to create the domain (did you have to use qemu-img, and what <disk> spec did you use). It's not a hard request, as I've already given reviews, but it may be something we appreciate in the future when doing git archaeology.
Hmm, yeah that sounds like a good idea. The partial problem may be that I had most of the files already laying around from various other testing, so I'll need to recreate the steps.