I'm working on the device tracking (to support disk failure prediction
etc), and have a question about how to approach the CLI commands. We have
several commands,
device info <devid>
device ls
device ls-by-daemon <who>
device ls-by-host <host>
info will dump metadata about a single device, e.g.
$ bin/ceph device info
INTEL_SSDPEDMD400G4_CVFT520200G7400BGN
device INTEL_SSDPEDMD400G4_CVFT520200G7400BGN
attachment gnit:nvme0n1
daemons osd.0,osd.1
expected_failure 2019-10-20 21:12:12.000000 (as of 2018-06-06 09:35:21.876620)
$ bin/ceph device info INTEL_SSDPEDMD400G4_CVFT520200G7400BGN -f json-pretty
{
"devid": "INTEL_SSDPEDMD400G4_CVFT520200G7400BGN",
"location": [
{
"host": "gnit",
"dev": "nvme0n1"
}
],
"daemons": [
"osd.0",
"osd.1"
],
"expected_failure": "2019-10-20 21:12:12.000000",
"expected_failure_stamp": "2018-06-06 09:35:21.876620"
}
I'm not sure about the ls commands, though. Originally, ls would just
list the device ids, like so:
$ bin/ceph device ls
INTEL_SSDPEDMD400G4_CVFT520200G7400BGN
but it is nice to get a table with useful information like so:
$ bin/ceph device ls
DEVICE HOST DAEMONS EXPECTED FAILURE
INTEL_SSDPEDMD400G4_CVFT520200G7400BGN gnit:nvme0n1 osd.0 osd.1 2019-10-20 21:12:12.000000
Same for the ls-by-* variants, which skip (or change) the HOST and DAEMONS
columns.
Similarly, the original JSON commands just output a JSON list, but I
changed them to dump everything:
$ bin/ceph device ls-by-daemon osd.0 -f json-pretty
[
{
"devid": "INTEL_SSDPEDMD400G4_CVFT520200G7400BGN",
"location": [
{
"host": "gnit",
"dev": "nvme0n1"
}
],
"daemons": [
"osd.0",
"osd.1"
],
"expected_failure": "2019-10-20 21:12:12.000000",
"expected_failure_stamp": "2018-06-06 09:35:21.876620"
}
]
I think this is all fine and good except that lazy bash graybeards like
myself who haven't mastered jq can do a simple for f in `ceph device ls`
loop. We could only dump the table if you add an optional 'detail'
argument like we've done elsewhere, but I kind of hate that.
Any suggestions on how we should approach this type of thing in the CLI?
The current version of the PR is doing the verbose thing:
https://github.com/ceph/ceph/pull/22423
sage
--
To unsubscribe from this list: send the line "unsubscribe ceph-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
