[c-nsp] ASR1001 tracelogs

Klaus Kastens kiste at bawue.de
Sun Mar 23 17:43:09 EDT 2014


Hi Peter,

> I got a ASR1001 with a constantly growing "tracelogs" folder. The space on
> disk is not an issue, its more of annoying when i get my rancid updates.
> 
> Do anyone of you have any solution about this?

We had the same problem with RANCID starting somewhere in IOS-XE 3.6.
You can avoid writing tracelogs by setting the log level to "emergency",
unfortunately this has to be done for every process:

    platform trace runtime slot F0 bay 0 process chassis-manager module all-modules level emergency
    platform trace runtime slot F0 bay 0 process emd module all-modules level emergency
    platform trace runtime slot F0 bay 0 process forwarding-manager module all-modules level emergency
    |
    platform trace runtime slot R0 bay 0 process pluggable-services module all-modules level emergency
    platform trace runtime slot R0 bay 0 process shell-manager module all-modules level emergency
    platform trace runtime slot R0 bay 0 process virt-manager module all-modules level emergency

There was a bug in earlier IOS-XE versions, platform trace level
setting was ignored. IIRC this was fixed somewhere in 3.9 (CSCuc89971).

Before this fix we used a filter in the rancid script's "ShowFlash"
parser, also filtering the changing size output:

    next if (/ \/bootflash\/$/);
    next if (/ \/bootflash\/tracelogs$/);
    next if (/ \/bootflash\/tracelogs\//);
    next if (/\d+ bytes total \(\d+ bytes free\)$/);


Best regards,
 Klaus


More information about the cisco-nsp mailing list