[ClusterLabs] Antw: [EXT] Re: Configuring millisecond timestamps in pacemaker.log.

[Klaus Wenninger]

On 1/12/21 8:42 AM, Ulrich Windl wrote:
>>>> Ken Gaillot <kgaillot at redhat.com> schrieb am 11.01.2021 um 21:16 in
> Nachricht
> <cca457ea7a3885c044ba4ead8fd43b4eb2c8ee15.camel at redhat.com>:
>> Pacemaker doesn't currently support it, sorry. It should be pretty easy
>> to add though (when built with libqb 2), so hopefully we can get it in
>> 2.1.0.
> Some time ago I wrote my own set of logging routines, and actually I uses
> nanoseconds (as muc as I could get) as timestamp. Just to shorten output, I
> just truncate the extra digits off.
Actually there is no need to implement high-res timer handling
in pacemaker.
On one hand it seems to be available (at least for logging) via
current libqb.
And pacemaker already has built in high-res timer handling
already - currentlybeing used by the alerts code (that is at
least what I had implemented it for -maybe used in other places
meanwhile as well ;-) ).
Formatting/trimming can be done as in the 'date' command using '%N'.

That said alerts are one more way to get certain events logged at high
resolution.
Cool part about it is that the stamps should be quite accurate as the
stamp is being taken in the context of the event that is 'alerted' and
queued to execd for actual alerting.
Not sure if timestamps are being taken that close to the call via the
other means suggested.

Klaus
> The (somewhat paranoid) code looks like this:
>
>                         if ( (Log.format & LTF_FRAC_MASK) != 0 )
>                         {
>                                 ssize_t nsz = 10;
>
>                                 rval = snprintf(fp + 1, nsz, "%09ld",
>                                                 ts.tv_nsec);
>                                 assert(rval < nsz);
>                         }
>                         switch ( Log.format & LTF_FRAC_MASK )
>                         {
>                         case LTF_FRAC_LONG:
>                                 fp[7] = '\0';           /* .hhtuuu */
>                                 break;
>                         case LTF_FRAC_MEDIUM:
>                                 fp[4] = '\0';           /* .hht */
>                                 break;
>                         case LTF_FRAC_SHORT:
>                                 fp[3] = '\0';           /* .hh */
>                                 break;
>                         default:
>                                 fp[0] = '\0';   /* '\0' */
>                         }
>
>> Of course Pacemaker has always supported logging via syslog, and syslog
>> can be configured to use high‑res timestamps, so that's a workaround.
>>
>> Does anyone have a strong opinion regarding using high‑res timestamps
>> in the Pacemaker detail log whenever supported, vs adding a new
>> sysconfig option for it? I feel like we have a ridiculous number of
>> options already, and the detail log is expected to be verbose.
> Yes options may be a problem. For the routines above I offered these options
> (that may be an overkill):
>
>               D      Selects the date component.  The possible formats are (Y
> is year, M is month, D is day):
>
>                      L      Long format: YYYYMMDD
>
>                      M      Medium format: YYMMDD
>
>                      S      Short format: MMDD
>
>                      N      None
>
>               T      Selects the time component.  The possible formats are (H
> is hour, M is minute, S is second):
>
>                      L      Long format: HHMMSS
>
>                      M      Medium format: MMSS
>
>                      S      Short format: SS
>
>                      N      None
>
>               F      Selects the fractional seconds component.  The possible
> formats are (h are hundreds, t are thousands, u are microseconds):
>
>                      L      Long format: .hhtuuu
>
>                      M      Medium format: .hht
>
>                      S      Short format: .hh
>
>                      N      None
>
> Regards,
> Ulrich
>
>> On Mon, 2021‑01‑11 at 19:12 +0000, Gerry R Sommerville wrote:
>>> Hello,
>>>
>>> I am wondering if it is possible to configure high resolution
>>> timestamps (including milliseconds) in the pacemaker.log? I was able
>>> to get hi‑res timestamps in the corosync.log by adding 'timestamp:
>>> hires' under the logging directive in corosync.conf. I was hoping
>>> Pacemaker has something similar but I don't see anything in
>>> '/etc/sysconfig/pacemaker' or the Pacemaker documentation regarding
>>> hi‑res timestamps.
>>>  
>>> Gerry Sommerville
>>> Db2 Development, pureScale Domain
>>> E‑mail: gerry at ca.ibm.com 
>> ‑‑ 
>> Ken Gaillot <kgaillot at redhat.com>
>>
>> _______________________________________________
>> Manage your subscription:
>> https://lists.clusterlabs.org/mailman/listinfo/users 
>>
>> ClusterLabs home: https://www.clusterlabs.org/ 
>
>
> _______________________________________________
> Manage your subscription:
> https://lists.clusterlabs.org/mailman/listinfo/users
>
> ClusterLabs home: https://www.clusterlabs.org/