[rancid] Reasoning behind not running a "show vlan" if specific VTP options are set?

Daniel Kerse daniel.kerse at gmail.com
Tue Jan 24 22:58:17 UTC 2017


Can someone please look at iosxr.pm and see if a similar change is also
required?

We get a lot of unwanted diffs there for growing files.

Here:
!Flash: harddiska: 3093 -rw- 32 Thu Jul 28 15:47:35 2016 env_hist
!Flash: harddiska: 3094 -rw- 2996 Tue Oct 11 23:18:49 2016 env_cont
!Flash: harddiska: 3095 -rw- 32 Thu Jul 28 15:47:43 2016 genstr_hist
-!Flash: harddiska: 3096 -rw- 977158 Mon Jan 23 16:01:55 2017 genstr_cont
+!Flash: harddiska: 3096 -rw- 980538 Mon Jan 23 16:56:16 2017 genstr_cont
!Flash: harddiska: 3097 -rw- 32 Thu Jul 28 15:47:39 2016 errmsg_hist
!Flash: harddiska: 3098 -rw- 15531 Wed Oct 12 01:12:37 2016 errmsg_cont
!Flash: harddiska: 3099 -rw- temp_static_data
Here:
!Flash: harddiska: 3176 -rw- 1892 Thu Oct 20 03:50:48 2016 temp_hist
!Flash: harddiska: 3177 -rw- temp_cont
!Flash: harddiska: 3178 -rw- 1124 Wed Sep 30 00:54:09 2015 volt_static_data
-!Flash: harddiska: 3179 -rw- 7800 Mon Nov 14 03:39:04 2016 volt_hist
+!Flash: harddiska: 3179 -rw- 8260 Tue Jan 24 02:48:19 2017 volt_hist
!Flash: harddiska: 3180 -rw- volt_cont
!Flash: harddiska: 3181 -rw- 514 Sun Jan 15 22:25:01 2017 errmsg_hist
!Flash: harddiska: 3182 -rw- 75802 Tue Jan 24 00:18:13 2017 errmsg_cont
And here:
!Flash: harddiska: 3179 -rw- 7800 Mon Nov 14 03:39:04 2016 volt_hist
!Flash: harddiska: 3180 -rw- volt_cont
!Flash: harddiska: 3181 -rw- 514 Sun Jan 15 22:25:01 2017 errmsg_hist
-!Flash: harddiska: 3182 -rw- 71876 Mon Jan 23 15:43:30 2017 errmsg_cont
+!Flash: harddiska: 3182 -rw- 73084 Mon Jan 23 16:59:46 2017 errmsg_cont
!Flash: harddiska: 3183 -rw- 24 Tue Jun 4 08:13:53 2013 diag_hist
!Flash: harddiska: 3184 -rw- 24 Tue Jun 4 08:13:55 2013 diag_cont

In iosxr.pm sub DirSlotN I see:
        # filter frequently changing files from IOX bootflash, hardiska,
        # and nvram
        if ($dev =~ /(bootflash|harddisk|nvram)/) {
            if
(/(ce_switch.log|temp_cont|temp_static_data|uptime_cont|volt_cont)\s*$/) {
                # change
                # 57          -rw-  23100       <no date>
 volt_cont
                # 614788      drwx  4096        Fri Aug 20 12:06:25 2010
 temp_cont
                # to
                # 57          -rw-              <no date>
 volt_cont
                # 614788      drwx
 temp_cont
So perhaps that second if statement needs to catch a few other filenames
please. We are running rancid-3.2-2 from here:
http://pkgs.fedoraproject.org/cgit/rpms/rancid.git/?h=epel7

Thanks,
Dan

On Tue, Jan 24, 2017 at 10:07 AM Piegorsch, Weylin William <weylin at bu.edu>
wrote:

> Thanks Lee.  I haven’t looked yet, and probably won’t until February at
> the earliest.  I can copy/paste that code, replacing “log” for “debug” on
> the pasted portion.  I’ll try it out and see how it works. If that works,
> is that something that can be forward ported into new versions of code?
> I’m a network engineer, my group doesn’t have software development as a
> core competency, but I’m willing to do this if there’s “vendor” support (to
> use the generic term my boss uses) in future versions.
> weylin
>
> On 1/22/17, 18:09, "Lee" <ler762 at gmail.com> wrote:
>
>     On 1/21/17, Piegorsch, Weylin William <weylin at bu.edu> wrote:
>     > Thanks Lee.  Is ios.pm called for NX-OS devices (“cisco-nx”)?  I
> tried
>     > looking through nxos.pm but didn’t see anything equivalent.  I’m
> using
>     > v3.4.1 if it makes a difference.
>
>     It looks like 3.4.1 already has code to blank out the file size &
>     timestamp.  Starting at line 509 of rancid-3.4.1/lib/rancid/nxos.pm
>
>     # This routine parses "dir /all ((disk|slot)N|bootflash|nvram):"
>     sub DirSlotN {
>
>     down a bit further it's got:
>             next if (/BufferMonitor-1HourData/);
>             if (/ log\/$/) {
>                 # change
>                 #         8192    Jan 08 14:05:05 2015  log/
>                 # to
>                 #                                       log/
>                 if (/(\s*\d+\s+)(\S+ \d+\s+\d+:\d+:\d+ \d+)(.*)/) {
>                     my($a, $dt, $rem) = ($1, $2, $3);
>                     my($dtl) = length($dt);
>                     my($fmt) = "%s%-". $dtl ."s%s\n";
>                     $_ = sprintf($fmt, $a, "", $rem);
>                 }
>             }
>
>     Regards,
>     Lee
>
>
>     > weylin
>     >
>     > On 1/21/17, 13:33, "Lee" <ler762 at gmail.com> wrote:
>     >
>     >     On 1/20/17, Piegorsch, Weylin William <weylin at bu.edu> wrote:
>     >     >> What we see on most of our Nexus infrastructure is every
> night:
>     >     >> ...
>     >     >> Which is quite annoying...
>     >     >
>     >     > If there’s a mechanism for removing these “debug” updates for
> NX-OS,
>     > that
>     >     > would be *awesome.*  I opened a TAC case about this, it’s not
>     > something I
>     >     > have any kind level of control over, whatsoever.
>     >
>     >     Take a look at rancid/lib/rancid/ios.pm - it has code to blank
> the
>     >     file size & timestamp so that updates to the file(s) don't cause
>     >     rancid to think the config has changed.  Look for
>     >      # filter frequently changing files (dhcp & vlan database)
>     >      # change from:
>     >      #    9  -rw-         660  Jan 15 2011 20:43:54 vlan.dat
>     >      #    9  -rw-         660  Jan 15 2011 20:43:54 +00:00  vlan.dat
>     >      # to:
>     >      #       -rw-                                   vlan.dat
>     >      #       -rw-                                           vlan.dat
>     >     and adapt as needed
>     >
>     >     Regards,
>     >     Lee
>     >
>     >     >
>     >     > weylin
>     >     >
>     >     > On 1/19/17, 15:55, "Mischa Diehm" <mischa.diehm at unibas.ch>
> wrote:
>     >     >
>     >     >     Hi,
>     >     >
>     >     >
>     >     >     On 13.01.17, 19:37, "heasley" <heas at shrubbery.net> wrote:
>     >     >
>     >     >
>     >     >     >It causes nevery-ending changes in the config as ports
> become
>     >     >     >active/inactive,
>     >     >     >eg: someone disconnects or powers-off equipment.  and in
> the case
>     > of
>     >     > VTP,
>     >     >     >it
>     >     >     >potentially is network wide.  if you wanted to monitor
> such
>     > things,
>     >     > its
>     >     >     >probably best done in a NMS, where you'd catch all such
> changes.
>     >     >     >
>     >     >     >rancid 3.5 added for some modules:
>     >     >     >
>     >     >     ># FILTER_OSC determines if oscillating data such as keys,
>     > passwords,
>     >     > etc
>     >     >     >are
>     >     >     ># filtered from configs by the value set (NO | YES).
> FILTER_PWDS
>     > may
>     >     >     >override
>     >     >     ># this.  see rancid.conf(5).
>     >     >     >#FILTER_OSC=YES; export FILTER_OSC
>     >     >     >
>     >     >     >such a thing could be dependent upon this configuration,
> but
>     > again I
>     >     > do
>     >     >     >not
>     >     >     >see the point.
>     >     >
>     >     >     Thanks for the clarification. This really kind of depends
> on
>     > where
>     >     > rancid
>     >     >     is run. In a campus/client env. I see your point but we
> also have
>     > more
>     >     >     static parts where this is usesful information. There it
> actually
>     > is
>     >     > also
>     >     >     useful to see what ports actually have changed. But you
> are right
>     > maybe
>     >     > no
>     >     >     worth it overall.
>     >     >     But thanks for pointing out the OSC part. What we see on
> most of
>     > our
>     >     > Nexus
>     >     >     infrastructure is every night:
>     >     >
>     >     >     - !Flash: debug:      124053    Jan 18 05:38:00 2017
>     > bootvar_debug.log
>     >     >     - !Flash: debug:           6    Jan 18 05:38:00 2017
>     >     > bootvar_debug.meta
>     >     >     + !Flash: debug:      124872    Jan 19 05:38:05 2017
>     > bootvar_debug.log
>     >     >     + !Flash: debug:           6    Jan 19 05:38:05 2017
>     >     > bootvar_debug.meta
>     >     >
>     >     >
>     >     >     And this:
>     >     >     - !Flash: logflash:       87092    Jan 17 05:43:42 2017
>     > accounting_log
>     >     >     + !Flash: logflash:      136367    Jan 18 05:43:35 2017
>     > accounting_log
>     >     >
>     >     >
>     >     >
>     >     >     Which is quite annoying...
>     >     >
>     >     >     >if you're not using VTP and its not including the vlan
> stuff, i
>     > do want
>     >     > to
>     >     >     >fix that.
>     >     >
>     >     >     Well then you have to fix it for this as said:
>     >     >     XXX-cc# show vtp status
>     >     >     Service not enabled
>     >     >
>     >     >
>     >     >     Cheers,
>     >     >     Mischa
>     >     >
>     >     >     >
>     >     >     >> Yes.  Which is yet another reason why it's so nice
> having the
>     > source
>     >     > -
>     >     >     >> it's easy to make trivial changes like that.  Or adding
>     > commands
>     >     > like
>     >     >     >> 'show vlan all-ports'
>     >     >     >>
>     >     >     >> Regards,
>     >     >     >> Lee
>     >     >     >>
>     >     >     >> _______________________________________________
>     >     >     >> Rancid-discuss mailing list
>     >     >     >> Rancid-discuss at shrubbery.net
>     >     >     >>
> http://www.shrubbery.net/mailman/listinfo/rancid-discuss
>     >     >
>     >     >
>     >     > _______________________________________________
>     >     > Rancid-discuss mailing list
>     >     > Rancid-discuss at shrubbery.net
>     >     > http://www.shrubbery.net/mailman/listinfo/rancid-discuss
>     >
>     >
>     >
>     >
>     >
>
>
>
>
> _______________________________________________
> Rancid-discuss mailing list
> Rancid-discuss at shrubbery.net
> http://www.shrubbery.net/mailman/listinfo/rancid-discuss
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.shrubbery.net/pipermail/rancid-discuss/attachments/20170124/c2c7c12b/attachment.html>


More information about the Rancid-discuss mailing list