[c-nsp] high CPU usage when coyping to flash

Xu Hu jstuxuhu0816 at gmail.com
Mon Mar 12 02:06:21 EDT 2012


So the problem cannot solved by TAC?

2012/2/28 James Slepicka (c-nsp) <cisco-nsp at slepicka.net>

> I previously ran into this issue w/ 4500 + Sup6E running 12.2(52)SG.  OSPF
> adjacencies would drop during file transfers.  TAC pointed me to CSCsw84727:
>
> Writing to bootflash creates instabilities when using low timers
> Symptom:
> Writing to bootflash creates instabilities when using low timers (< 10 sec)
>
> Conditions:
> - low timers for control protocols are in use (e.g. HSRP, OSPF, BGP)
> - supervisor/switch is one of: Sup6-E, Sup6L-E, 4900M, 4948-E
>
> Workaround:
> Avoid lengthy bootflash operations, like copying really large files in IOS.
> Upgrade to 12.2(50)SG06, 12.2(53)SG02, 12.2(54)SG or above.
>
> I'm running 12.2(54)SG now.  It's been a while, but I seem to recall still
> seeing high CPU utilization during flash operations, but OSPF would remain
> stable.
>
> -----Original Message-----
> From: cisco-nsp-bounces at puck.nether.net [mailto:
> cisco-nsp-bounces at puck.nether.net] On Behalf Of Jiri Prochazka
> Sent: Sunday, February 26, 2012 2:43 AM
> To: cisco-nsp at puck.nether.net
> Subject: [c-nsp] high CPU usage when coyping to flash
>
> Hi,
>
> I've came to a strange behaviour on 4900M. When trying to copy a new IOS
> to flash, CPU utilization rises from usual 15% to 100%, even ssh stops
> answering, some ports are being disabled due to UDLD error, some were
> unbundled from etherchannels because of LACP-PDU not being generated.
>
> The switch is just absolutely overloaded which cause a lot of issues. I
> did not want to believe simple 'copy tftp:// bootflash:' could cause this
> so I tried it once more, with the same result.
>
>
> I am now trying to figure it in a lab, but it seems I am in a dead end.
>
>
> I tried in on 4948E, 4900M, different IOSes (12.2(54)SG1, 15.0(2)SG1,
> 15.0(2)SG3).
>
> I tried to use tftp, scp, http, all with the same result of 100% CPU usage.
>
>
> I am not concerned about the high cpu usage itself, but I do absoltely do
> not understand why it effectively stops the rest of the switch to work
> properly..
>
>
>
> Thank you for your advice.
>
> --
>
> Jiri Prochazka
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
>
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/
>


More information about the cisco-nsp mailing list