[c-nsp] NPE-G2 Gigabit Ignored Errors

Rodney Dunn rodunn at cisco.com
Fri Sep 12 15:07:46 EDT 2008


  ring sizes: RX = 128, TX = 256                                                
  rx_particle_size: 512                                                         
  rx_pak = 0x0444F908                                                           
  rx_head = 122                                                                 
  rx_discard = FALSE                                                            
  tx_head = 4, tx_count = 0                                                     
  chip_state = 2, ds->tx_limited = 0                                            
  throttled = 0, enabled = 0, disabled = 11                                     
  reset=5(init=1, restart=4), auto_restart=6                                    
  tx_underflow = 0, tx_overflow = 0,  tx_end_count = 51049471                   
  rx_nobuffer = 0, rx_overrun = 0                                               
  rx_no_descriptors = 2115,  rx_interrupt_count = 41996968                      
  rx_crc_error = 0, rx_too_big = 0, rx_resource_error = 4629    


128 rx ring depth.

rx_resource and rx_no_descriptors errors.

   Whenever there is lack of descriptor to store the incoming packet in         
the RX ring an ignored/resource error will be reported for the native           
gige interfaces. Resource error indicates the number of instances when          
the DMA ran out of Rx descriptors. Ignored indicates the number of              
packets when the DMA ran out of Rx descriptors.   


On Fri, Sep 12, 2008 at 03:02:07PM -0400, Clayton Zekelman wrote:
> 
> Here are the sh controller and sh proc results.
> 
> I'll send the config directly - too much to sanitize ...
> 
> Thanks!
> 
> Hardware is MV64460 Internal MAC (Revision MV64460-Ethernet)
>   network link is up
>   Config is 1Gbps, Full Duplex
>   Selected media-type is RJ45
>   GBIC is not present
>  Ethernet Unit Global Registers:
>   PHY Address            = 0x00000820    SMI (PHY Control)      = 0x0C001000
>   Default Address (Err)  = 0xFE200000    Default ID (Err)       = 0x000001D1
>   Interrupt Cause        = 0x00000210    Interrupt Mask         = 0x0000108E
>   Error Address          = 0xD7FC6000    Internal Addr Error    = 0x00000000
>   Port Pad Calibration   = 0x0013000B    Base Address 0         = 0xD7FC0002
>   Size (BA0)             = 0x00030000    Base Address 1         = 0x38002E00
>   Size (BA1)             = 0x03FF0000    Base Address 2         = 0x00000000
>   Size (BA2)             = 0x00000000    Base Address 3         = 0x00000000
>   Size (BA3)             = 0x00000000    Base Address 4         = 0x00000000
>   Size (BA4)             = 0x00000000    Base Address 5         = 0x00000000
>   Size (BA5)             = 0x00000000    Header Retarget Base   = 0x00000000
>   Header Retarget Ctrl   = 0x00000000    High Address Remap 0   = 0x00000000
>   High Address Remap 1   = 0x00000000    Base Address Enable    = 0x0000003C
>   Port Access Protect 0  = 0x0000000F    Port Access Protect 1  = 0x0000000F
>   Port Access Protect 2  = 0x0000000F
> 
>  MAC Specific Registers:
>   Port Cfg           PxC = 0x00000000    Port Cfg Extd     PxCX = 0x00000000
>   MII Serial Params      = 0x00218823    GMII Serial Params     = 0x00000006
>   VLAN EtherType  EVLANE = 0x00000000    MAC Addr Low     MACAL = 0x0000091B
>   MAC Addr High    MACAH = 0x001A6D30    SDMA Config        SDC = 0x01002005
>   PORT SERIAL CTRL   PSC = 0x00A2A60D    PORT STATUS         PS = 0x00001C16
>   TX Queue Cmd       TQC = 0x00000000    TxQ Fixed Prio   TQFPC = 0x00000001
>   MTU (Token-Bucket) MTU = 0x00000000    INTR CAUSE          IC = 0x80000007
>   INTR CAUSE EXT     ICE = 0x80000001    Intr Mask          PIM = 0x00080804
>   Extend Intr Mask  PEIM = 0x00000101    RXQ Desc Ptr 0   CRDP0 = 0xD7FC0F80
>   RX Queue Cmd       RQC = 0x0003FE01    TXQ Desc Ptr 0  TCQDP0 = 0xD7FC1100
>   TX Curr Desc     TCSDP = 0xD7FC1100
> 
>   MAC Serial Port is ENABLED
>   MAC Autoneg Capability: [Speed] [Duplex] [Flow Control]
>   MAC Status: Link is UP, Speed is 1000Mbps, Duplex is Full, Interface: GMII
> 
>   RX Queue [0] is ENABLED
> 
>  PHY Registers:
>   PHY is Marvell 88E1146C (Rev D0), address 0x0
>   Control                = 0x1000           Status                 = 0x796D
>   PHY ID 1               = 0x0141           PHY ID 2               = 
> 0x0CD4
>   Auto Neg Advertisement = 0x0000           Link Partner Ability   = 0xCDE1
>   Auto Neg Expansion     = 0x000D           Next Page Tx           = 0x2001
>   Link Partner Next Page = 0x4773           1000BaseT Control      = 0x0200
>   1000BaseT Status       = 0x3C00           Extended Status        = 0x3000
>   PHY Specific Control   = 0x0008           PHY Specific Status    = 0xAC00
>   Interrupt Enable       = 0x6C00           Interrupt Status       = 0x0000
>   Ext PHY Spec Control   = 0x0CE2           Receive Error Counter  = 0x0000
>   LED Control            = 0x4101
>   Ext PHY Spec Control 2 = 0x000A           Ext PHY Spec Status    = 0x800B
>   PHY says Link is UP, Speed 1000Mbps, Full-Duplex [AUTONEG Done]
> 
>   AUTONEG - Our ability is     1000M/FD
>   AUTONEG - Partner ability is 1000M/FD 1000M/HD 100M/FD 100M/HD 10M/FD 
>   10M/HD
> 
>  IDB Information:
>   lc_ip_turbo_fs = 0xAB7F4, ip_routecache = 0x11 (dfs = 0/mdfs = 0)
>   rx cache size = 1000, rx cache end = 849
>   max_mtu = 1528
>  Software MAC address filter(hash:length/addr/mask/hits):
>  need_af_check = 0
>   0x00:  0  ffff.ffff.ffff  0000.0000.0000         0
>   0x5B:  0  0100.5e00.0005  0000.0000.0000         0
>   0xC0:  0  0100.0ccc.cccc  0000.0000.0000         0
> 
>  Internal Driver Information:
>   RX Ring base: 0xD7FC0000
>   TX Ring base: 0xD7FC1000
>   Software RX Head: 0xD7FC0F40
>   Hardware RX Head: 0xD7FC0800
>   Software TX Head: 0xD7FC1080
>   Hardware TX Head: 0xD7FC2F00
>   ring sizes: RX = 128, TX = 256
>   rx_particle_size: 512
>   rx_pak = 0x0444F908
>   rx_head = 122
>   rx_discard = FALSE
>   tx_head = 4, tx_count = 0
>   chip_state = 2, ds->tx_limited = 0
>   throttled = 0, enabled = 0, disabled = 11
>   reset=5(init=1, restart=4), auto_restart=6
>   tx_underflow = 0, tx_overflow = 0,  tx_end_count = 51049471
>   rx_nobuffer = 0, rx_overrun = 0
>   rx_no_descriptors = 2115,  rx_interrupt_count = 41996968
>   rx_crc_error = 0, rx_too_big = 0, rx_resource_error = 4629
>   rx_sop_eop_error = 0
>   tqc = 0xF1002448, cause = 0xF1002460, cause_ext = 0xF1002464
>  Address Filter:
>   Promiscuous mode OFF
>   (All other entries are empty)
> 
>  Statistics:
> 
>   Receive and Transmit Statistics:
>     RX Good Octets         45879122062      TX Good Octets         
>     22423886055
>     RX Good Packets           58699012      TX Good Packets           
>     52765087
>     RX Bad Octets                    0
>     RX Bad Packets                   0
>     RX Broadcast Packets           289      TX Broadcast Packets            
>     25
>     RX Multicast Packets          3255      TX Multicast Packets           
>     336
> 
> 
>   Error Statistics:
>     TX MAC Error                     0
>     TX Excess Collisions             0
>     RX Bad MAC Ctrl Frame            0
>     RX Good Flow Control             0
>     RX Bad Flow Control              0
>     RX Undersize Paks                0
>     RX Oversize Paks                 0
>     RX Fragments                     0
>     RX Jabber Paks                   0
>     RX MAC Error Events              0
>     RX Bad CRC Events                0
>     RX Collisions                    0
>     RX Late Collisions               0
>     RX MAC Discard Frame        283215
>     RX MAC Overrun Frame             0
> 
>   Total Packets (RX and TX), including errors:
>        0 ->   64              30696290
>       65 ->  127              22210681
>      128 ->  255               9294449
>      256 ->  511               4627758
>      512 -> 1023               6428675
>     1024 -> 
> max               38947325
> 
> 
> 
> 
> 
> 
> 
> 
> lns3-Windsor>sh proc cpu sort | excl 0.00
> CPU utilization for five seconds: 57%/51%; one minute: 59%; five minutes: 
> 58%
>  PID Runtime(ms)   Invoked      uSecs   5Sec   1Min   5Min TTY Process
>  191    943061842068783944         45  1.80%  1.77%  1.79%   0 PPP Events
>  150   133733268  50856299       2629  1.06%  1.26%  1.21%   0 AAA 
> SEND STOP EV
>   58   3289317521961441484        167  0.65%  0.66%  0.65%   0 ATM OAM Input
>   84   167482244 113323165       1477  0.57%  0.62%  0.64%   0 CEF process
>  181     9657704-1355730504          0  0.32%  0.37%  0.35%   0 L2X 
> Data Daemon
> 
>   80    40515732 130839022        309  0.24%  0.24%  0.24%   0 L2TP 
> mgmt daemon
>  203    35772716  92421692        387  0.24%  0.11%  0.12%   0 OSPF-1 Router
>  199    388871201904317329         20  0.16%  0.16%  0.16%   0 RADIUS
>   79    20652720 155060775        133  0.16%  0.12%  0.13%   0 L2X SSS 
>   manager
>    2      355544  10795823         32  0.08%  0.03%  0.02%   0 Load Meter
>  190     18684041721110574          1  0.08%  0.09%  0.08%   0 PPP manager
>   66   1206876561422016980         84  0.08%  0.22%  0.19%   0 IP Input
>  120     7097564  43773086        162  0.08%  0.03%  0.02%   0 PPP Bind
>   43     4238636  55478504         76  0.08%  0.08%  0.08%   0 Per-Second 
>   Jobs
>   16   198394500 979392997        202  0.08%  0.29%  0.14%   0 EnvMon
>   82     9579436  68673081        139  0.08%  0.02%  0.01%   0 IP Background
>  177      328704 541870926          0  0.08%  0.03%  0.02%   0 CCPROXY_CT
>  173    244061721459348569         16  0.08%  0.04%  0.06%   0 Net Input
>   73    21365328 131389120        162  0.08%  0.08%  0.08%   0 SSS Manager
>   62    12653484  48028054        263  0.08%  0.04%  0.06%   0 AAA 
> Server
>  186    19462048 591135940         32  0.08%  0.06%  0.09%   0 PPPoE 
> Background
> 
> 
> At 02:43 PM 9/12/2008, Rodney Dunn wrote:
> >On Fri, Sep 12, 2008 at 02:40:04PM -0400, Clayton Zekelman wrote:
> >>
> >> No luck... didn't fix it. Is it fixed in a subsequent release?  Are
> >> there any other parameters I can tune?
> >
> >Not really because you can't tune the rx ring depth.
> >
> >Check 'sh controller'.
> >
> >What does 'sh proc cpu sort | excl 0.00' say?
> >
> >Can you post the configuration..I'm curious what your features
> >look like because the more you have the less pps you get through
> >this box..it's all done in software and can't do all features at line
> >rate during a microburst.
> >
> >sh int stat
> >
> >
> >Rodney
> >
> >
> >>
> >> GigabitEthernet0/1 is up, line protocol is up
> >>   Hardware is MV64460 Internal MAC, address is 001a.6d30.091b (bia
> >> 001a.6d30.091b)
> >>   Description: to gig-fastiron Ethernet11
> >>   MTU 1500 bytes, BW 1000000 Kbit, DLY 10 usec,
> >>      reliability 255/255, txload 23/255, rxload 48/255
> >>   Encapsulation 802.1Q Virtual LAN, Vlan ID  1., loopback not set
> >>   Keepalive set (10 sec)
> >>   Full-duplex, 1000Mb/s, media type is RJ45
> >>   output flow-control is XON, input flow-control is XON
> >>   ARP type: ARPA, ARP Timeout 04:00:00
> >>   Last input 00:00:00, output 00:00:00, output hang never
> >>   Last clearing of "show interface" counters 00:10:09
> >>   Input queue: 0/4096/533/0 (size/max/drops/flushes); Total output 
> >drops: 0
> >>   Queueing strategy: fifo
> >>   Output queue: 0/40 (size/max)
> >>   30 second input rate 189692000 bits/sec, 30246 packets/sec
> >>   30 second output rate 91448000 bits/sec, 27197 packets/sec
> >>      18432915 packets input, 1555851456 bytes, 0 no buffer
> >>      Received 65 broadcasts, 0 runts, 0 giants, 0 throttles
> >>      1117 input errors, 0 CRC, 0 frame, 0 overrun, 1117 ignored
> >>      0 watchdog, 1034 multicast, 0 pause input
> >>      0 input packets with dribble condition detected
> >>
> >>
> >>
> >> At 02:16 PM 9/12/2008, Rodney Dunn wrote:
> >> >Can you bump up your input queue depth:
> >> >
> >> >hold-queue 4096 in
> >> >
> >> >and see if they stop.
> >> >
> >> >I don't suspect that is going to help because the ignores
> >> >are not increasing that would point to:
> >> >
> >> >CSCse05447
> >> >Externally found moderate defect: Resolved (R)
> >> >7200 ethernet interfaces should not throttle on input queue full drops
> >> >
> >> >Most likely you are seeing micro burst that are coming in faster
> >> >than the CPU can drain the rx ring.
> >> >
> >> >Rodney
> >> >
> >> >
> >> >
> >> >On Fri, Sep 12, 2008 at 12:50:29PM -0400, Clayton Zekelman wrote:
> >> >>
> >> >> I'm running a Cisco 7206/VXR with an NPE G2,  Version 12.4(4)XD4
> >> >> acting as an LNS.
> >> >>
> >> >> I'm getting input errors consistently incrementing on the Gig
> >> >> interface (ignored errors)
> >> >>
> >> >> Any way to fix this?  I saw some discussion a while back about this,
> >> >> and it seemed to have to do with buffers - but I can't find any
> >> >> definitive recommendations on what the settings should be.
> >> >>
> >> >>
> >> >>
> >> >> GigabitEthernet0/1 is up, line protocol is up
> >> >>   Hardware is MV64460 Internal MAC, address is 001a.6d30.091b (bia
> >> >> 001a.6d30.091b)
> >> >>   Description: to gig-fastiron Ethernet11
> >> >>   MTU 1500 bytes, BW 1000000 Kbit, DLY 10 usec,
> >> >>      reliability 255/255, txload 22/255, rxload 46/255
> >> >>   Encapsulation 802.1Q Virtual LAN, Vlan ID  1., loopback not set
> >> >>   Keepalive set (10 sec)
> >> >>   Full-duplex, 1000Mb/s, media type is RJ45
> >> >>   output flow-control is XON, input flow-control is XON
> >> >>   ARP type: ARPA, ARP Timeout 04:00:00
> >> >>   Last input 00:00:00, output 00:00:00, output hang never
> >> >>   Last clearing of "show interface" counters 00:22:17
> >> >>   Input queue: 0/75/1191/0 (size/max/drops/flushes); Total 
> >output drops:
> >> >0
> >> >>   Queueing strategy: fifo
> >> >>   Output queue: 0/40 (size/max)
> >> >>   30 second input rate 181384000 bits/sec, 29001 packets/sec
> >> >>   30 second output rate 86319000 bits/sec, 26045 packets/sec
> >> >>      38605963 packets input, 4274358612 bytes, 1 no buffer
> >> >>      Received 230 broadcasts, 0 runts, 0 giants, 0 throttles
> >> >>      2677 input errors, 0 CRC, 0 frame, 0 overrun, 2677 ignored
> >> >>      0 watchdog, 2196 multicast, 0 pause input
> >> >>      0 input packets with dribble condition detected
> >> >>      34556615 packets output, 1656923135 bytes, 0 underruns
> >> >>      0 output errors, 0 collisions, 0 interface resets
> >> >>      0 babbles, 0 late collision, 0 deferred
> >> >>      0 lost carrier, 0 no carrier, 0 pause output
> >> >>      0 output buffer failures, 0 output buffers swapped
> >> >> out
> >> >>
> >> >>
> >> >>
> >> >> ---
> >> >> Clayton Zekelman
> >> >> Managed Network Systems Inc. (MNSi)
> >> >> 344-300 Tecumseh Rd. E.
> >> >> Windsor, Ontario
> >> >> N8X 5E8
> >> >>
> >> >> tel. 519-985-8410
> >> >> fax. 519-985-8409
> >> >>
> >> >> _______________________________________________
> >> >> 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/
> >>
> >> ---
> >> Clayton Zekelman
> >> Managed Network Systems Inc. (MNSi)
> >> 344-300 Tecumseh Rd. E.
> >> Windsor, Ontario
> >> N8X 5E8
> >>
> >> tel. 519-985-8410
> >> fax. 519-985-8409
> 
> ---
> Clayton Zekelman
> Managed Network Systems Inc. (MNSi)
> 344-300 Tecumseh Rd. E.
> Windsor, Ontario
> N8X 5E8
> 
> tel. 519-985-8410
> fax. 519-985-8409        


More information about the cisco-nsp mailing list