7:51 a.m.
于 2011年09月07日 20:37, Peter Krempa 写道:
On 09/07/2011 01:45 PM, Osier Yang wrote:
> 于 2011年09月06日 21:20, Peter Krempa 写道:
>> diff --git a/tools/virsh.c b/tools/virsh.c
>> index 629233f..458252b 100644
>> --- a/tools/virsh.c
>> +++ b/tools/virsh.c
>> @@ -1054,16 +1054,46 @@ cleanup:
>> */
>> static const vshCmdInfo info_domblkstat[] = {
>> {"help", N_("get device block stats for a domain")},
>> - {"desc", N_("Get device block stats for a running
domain.")},
>> + {"desc", N_("Get device block stats for a running
domain.\n\n"
>> + " Explanation of fields:\n"
>> + " rd_req - count of read operations
>> (old format)\n"
>> + " rd_operations - count of read operations
>> (new format)\n"
>> + " rd_bytes - count of read bytes\n"
>> + " rd_total_times - total time read
>> operations took\n"
>
> Should also document the units IMHO, it's nano-seconds.
I agree, that documenting also the unit is important. I'm afraid that
the unit is hypervisor specific
(by now, the code is implemented only for QEmu) and we would create a
precedent so that other
hypervisors using other values will have to convert it before. I think
we should chose this unit wisely.
On the other hand, I think that nano-seconds are okay.
>
>> + " wr_req - count of write
>> operations (old format)\n"
>> + " wr_operations - count of write
>> operations (new format)\n"
>> + " wr_bytes - count of written bytes\n"
>
> Missed docs for wr_total_times.
Will add it in v4
>
>> + " flush_operations - count of flush
>> operations\n"
>> + " flush_total_times - total time flush
>> operations took\n"
>
> likewise
>
>> + " errs - error count")},
>> {NULL,NULL}
>> };
>
> Think it would be better if converting "rd_operations",
"wr_operations",
> and "flush_operations" to the old names. They have the same meaning,
> and we used "rd_req", "wr_req" in virsh for a long time. After
> conversion,
> we can have less and not duplicate documention.
I am aware, that the meanings of the two fields are the same. I was
surprised
to see new naming of this fields. If the new API has to stick with
these new names,
I'll have to convert them, but I don't really like approach.
There is no field of the old API, virsh just printed same words of the
struct
members. I intended to keep same with the QEMU returned keys' name.
>
>> static const vshCmdOptDef opts_domblkstat[] = {
>> {"domain", VSH_OT_DATA, VSH_OFLAG_REQ, N_("domain name, id
or
>> uuid")},
>> {"device", VSH_OT_DATA, VSH_OFLAG_REQ, N_("block
device")},
>> + {"human", VSH_OT_BOOL, 0, N_("print a more human readable
>> output")},
>> {NULL, 0, 0, NULL}
>> };
>>
>> +struct _domblkstat_human_readable {
>> + const char *field;
>> + const char *human;
>> +};
>> +
>> +static const struct _domblkstat_human_readable
>> domblkstat_translate[] = {
>> + { VIR_DOMAIN_BLOCK_STATS_READ_BYTES, N_("number of read
>> bytes: ") }, /* 0 */
>> + { VIR_DOMAIN_BLOCK_STATS_READ_REQ, N_("number of read
>> operations: ") }, /* 1 */
>> + { VIR_DOMAIN_BLOCK_STATS_READ_TOTAL_TIMES, N_("total duration
>> of reads: ") }, /* 2 */
>> + { VIR_DOMAIN_BLOCK_STATS_WRITE_BYTES, N_("number of bytes
>> written: ") }, /* 3 */
>> + { VIR_DOMAIN_BLOCK_STATS_WRITE_REQ, N_("number of write
>> operations:") }, /* 4 */
>> + { VIR_DOMAIN_BLOCK_STATS_WRITE_TOTAL_TIMES, N_("total duration
>> of writes: ") }, /* 5 */
>> + { VIR_DOMAIN_BLOCK_STATS_FLUSH_REQ, N_("number of flush
>> operations:") }, /* 6 */
>> + { VIR_DOMAIN_BLOCK_STATS_FLUSH_TOTAL_TIMES, N_("total duration
>> of flushes: ") }, /* 7 */
>
> Not sure if we should keep the same docs for these fields defined in
> struct
> info_domblkstat and virsh.pod? it looks a bit long though. :-)
Yes this information is tripple-redundant :/ but I don't really see
other options. Maybe just drop it from
help command as there is the --human option, or drop the human option.
>
>> + { VIR_DOMAIN_BLOCK_STATS_ERRS, N_("error
>> count: ") }, /* 8 */
>> + { NULL, NULL }
>> +};
>> diff --git a/tools/virsh.pod b/tools/virsh.pod
>> index d826997..c8d88c9 100644
>> --- a/tools/virsh.pod
>> +++ b/tools/virsh.pod
>> @@ -501,13 +501,27 @@ be lost once the guest stops running, but the
>> snapshot contents still
>> exist, and a new domain with the same name and UUID can restore the
>> snapshot metadata with B<snapshot-create>.
>>
>> -=item B<domblkstat> I<domain> I<block-device>
>> +=item B<domblkstat> I<domain> I<block-device>
[I<--human>]
>>
>> Get device block stats for a running domain. A I<block-device>
>> corresponds
>> to a unique target name (<target dev='name'/>) or source file
>> (<source
>> file='name'/>) for one of the disk devices attached to
I<domain>
>> (see
>> also B<domblklist> for listing these names).
>>
>> +Use I<--human> for a more human readable output.
>> +
>> +B<Explanation of fields:>
>> + rd_req - count of read operations (old format)
>> + rd_operations - count of read operations (new format)
>> + rd_bytes - count of read bytes
>> + rd_total_times - total time read operations took
>
> Units
Fix in v4.
>
>> + wr_req - count of write operations (old format)
>> + wr_operations - count of write operations (new format)
>> + wr_bytes - count of written bytes
>
> Missed wr_total_times
Fix in v4.
>
>> + flush_operations - count of flush operations
>> + flush_total_times - total time flush operations took
>
> Units.
Fix in v4.
>
>> + errs - error count
>> +
>> =item B<domifstat> I<domain> I<interface-device>
>>
>> Get network interface stats for a running domain.
>
> Others look good to me.
>
> Regards
> Osier
Thanks for finding the small glitches.
Peter