Hot answers tagged

9

There is loss, you're just not seeing the impact. The display is only showing 155706 input errors out of 46982598919 packets in the previous three days and four hours. This is .0003% packet loss, which is why it's so difficult to see first hand during testing. If you're not seeing any operational impact, odds are that this can be safely ignored. Such a ...


8

It's true that you have to know it in advance in order to calculate the CRC. The correct polynomial depends on the application of the CRC. For Ethernet, for example, the CRC-32 polynomial is part of the IEEE 802.3 standard. For you trivia buffs, it's 0x82608EDB.


7

Do a 'show proc mem sorted' and check for biggest memory users. If that's something that is not related to configuration, it may be bug in the code. If that's something that's related to configuration, you may consider re/deconfigure unnecessary services. As IOS can't defragment memory and you can't reclaim it unless some big allocating process frees memory,...


6

It can't be PC ethernet card, as DES-1024D is not 'cut-through', that means, if PC sent broken frame, DES-1024D dropped it and did not forward it. You've already done good debugging by moving the wire to another switch, this allows us in high degree of confidence to exclude 3COM as culprit for the errors. Options that are left is room 3 D-Link or wiring ...


6

Just to provide an answer for anyone else experiencing similar issues, this was a cosmetic bug that was fixed from JUNOS 12.3 onwards. Actual update from TAC "There is an issue with the stats collection command where the HW counter was not getting cleared. Thus the stats reporting was continuously incrementing it's counter even though no actual EOPE ...


5

This will be a bit of a layman's answer, rather than quoting spec, but it should be accurate. The collision domain is the sum total of the network segment (hosts, hubs) affected by the collision, and forced to deal with it by a random backoff delay. Anytime a collision is detected on an Ethernet line, any and all interfaces on that line are required to ...


5

You need to open a case with Cisco TAC... obligatory Cisco Support Forums reference to a similar problem Without TAC support perhaps the best you can do is reseat the LC... replace the LC... Upgrade IOS and pray it was a bug... Be sure the screws on the LC are snugged on tight when you do maintenance... loose screws can cause more problems than you might ...


4

You're seeing CRC and framing errors and general input errors. If this happened while setting up the port this could be caused by people still fiddling with the fiber. If this happens during normal operation most of the time it indicates a low light level or some other error with the fiber(s) or optics. You can check the light levels with show interfaces ...


4

If a packet gets corrupted on the wire, then the Ethernet FCS should catch that. However, many switches and routers will strip off the FCS, route the packet, and then compute a new FCS. If the packet gets corrupted while being shuffled around inside the switch (which happens occasionally), then you will end up with a corrupt packet that has a valid Ethernet ...


4

Ethernet layer 2 doesn't correct errors, it only detects them by frame check sequence (FCS). The algorithm for FCS is CRC with the polynomial G(x) = x32 + x26 + x23 + x22 + x16 + x12 + x11 + x10 + x8 + x7 + x5 + x4 + x2 + x + 1. CRC cannot detect certain paired (even-number) bit error combinations - with a stable connection, these are very rare. A flaky ...


4

@PHLiGHT, hello there! your cisco 861, imho, has overloaded by ingress traffic. Lets look on your cisco iface counters: 5 minute output rate 198000 bits/sec, 56 packets/sec 157099395 packets input, 3610517494 bytes Received 1 broadcasts, 76 runts, 0 giants, 655 throttles 338 input errors, 0 CRC, 0 frame, 149 overrun, 189 ignored you have 655 ...


3

An interface running at 20 Gbit/s processes between 1.6 and 30 million frames per second or 100 to 1800 Mframes per minute. One FCS error/min corresponds to a bit error rate of .8x10-12 which is slightly below the standard 10-12 that 802.3 usually designs for. Two errors/min are a bit on the high side but nothing to really worry about if you've maxed out ...


3

For the specific case of 100BASE-TX and the 100Mbps MII interface: there is a receiver signal RX_ER. The Physical Coding Sublayer (PCS) specification says: 24.2.2.1.7 Invalid code-groups ...The PCS indicates the reception of an Invalid code-group on the MII through the use of the RX_ER signal. The Reconciliation Sublayer and Media Independent ...


3

According to this juniper document it is the sum of transmit drops. This Juniper forum post explains it a bit more: Basically, when a packet cannot be either put in the buffer or retrieved from the buffer (i.e. when packet is scattered across several buffers due to the pkt size and one of the buffers cannot be accessed) then this counter increments. I ...


3

How are acknowledgement sent at data link layer to the sender when crc of frame is incorrect? That depends on the Data-Link protocol, but most Data-Link protocols simply drop bad frames. For example, if an ethernet device receives a frame with a bad FCS, the hardware simply discards the frame, and it never reaches the Data-Link layer. Also, most Data-...


3

Well, I see your question as "What is a collision domain". In simple terms, a collision domain, like the name suggests, is an area where collision can occur. Think of it as a narrow tunnel where space for only one car to pass through exists. If two cars pass through, they collide. Hence, the tunnel is known as a collision domain. Now, let's take this to ...


3

A collision domain is all devices connected by shared medium. Originally a collision domain was a physical cable, then repeaters and hubs were added. The shared medium ended at a router port and was temporally bounded (512 bit times). With the advent of switches the concept of collision domain shrank to the port on the switch, and as was noted collisions ...


3

Worst case scenario, your HW has gone bad Best case scenario, it's cosmetic failure due to software defect, luckily you are in SRE which will be supported until 2015, so maybe upgrade it to latest rebuild. There are two bugID which will cause this error in very benign way. CSCsk03373, due to large packets, fixed in SXH CSCsw32280, due to CRC errors, fixed ...


3

If you are getting memory allocation errors, a reload should be in order to clear up whatever is going on, at least temporarily. It looks like you are running out of memory (Cause: Not enough free memory). I would look at perhaps upgrading the memory, which is the cheapest option, or if this router is important enough, even upgrading to a new platform that ...


2

Updated Answer Originally I thought it was upgrading the firmware of my load-balancers that fixed my issues. That turns out to be coincidence. The actual problem involved a setting on the load-balancer itself. The Web servers in our cluster have to know the internal IP of the remote site in order to know what location is trying to access the page, then ...


2

Your NAT Exemptions on the Spokes are messed up. First, you have the Spoke 2 address as the source on Spoke 1, and you're missing Spoke 2 altogether: ip access-list extended NONAT deny ip 192.168.210.64 0.0.0.31 172.16.0.0 0.0.255.255 permit ip 192.168.210.64 0.0.0.31 any You need to have: ip access-list extended NONAT deny ip 192.168.210.224 0.0.0....


2

You didn't specify what protocol or media TCPs running on in your case, but assuming IP, the checksums calculated by the sender and checked by the receiver only verify the packet headers, not the pay load. As pointed out by others TCP can run on other protocols and media that may not provide data integrity, but TCP in accordance with RFC793, must guarantees ...


2

Remember that each layer is independent of every other layer, and the original design of each protocol in a layer is that my layer cannot depend on your layer doing any error detection, or the error detection which my layer requires. Also, if you have a datagram which is destined for a particular layer (e.g. ARP for layer-2, or ICMP for layer-3), the layer ...


2

UDP itself has no mechanisms for neither flow control, nor congestion control, and no error correction. If the application's datastream needs any of these, then they must be implemented within the application. However, UDP may have error detection: The UDP header has a 16bit checksum field, but it's use (with UDP-on-IPv4) is not mandatory (however it is ...


2

Ethernet's link layer uses CRC32 in the Frame Check Sequence field, trailing the payload data. If FCS fails the frame is dropped. Other reasons for dropping a frame include link congestion, QoS, ACL filtering. In any case, there's no recovery and it's up to higher level protocols to cope with it (or not). 10 gigabit and faster speed PHYs may use additional ...


2

A CRC in general isn't well suited for error correction, just detection. Any redundancy you add to an original data block is limited in the amount of errors it can detect (or correct). Any practical transmission channel has a statistical error rate greater than zero. Cleverly constructed redundant data can decrease the probability of an error by several ...


1

Have you got errors on any other ports? The Nexus switches in your diagram operate in cut-through mode by default, and packets larger than ~768 bytes can be forwarded even if they have errors. This is configurable as documented by Cisco, here. This means you could be getting bad packets from one connected device/server and they're just showing up on the ...


1

302323 input errors, 302323 CRC, 0 frame, 0 overrun, 0 ignored I think this is the key information. Apparently, the fiber from the 2960 transmitter to the 3650 receiver is very bad. If you've already ruled out transceivers and patch cables, double check for fiber type mismatches (SMF vs MMF and SPC vs APC) and what's left is the dark fibre and its ...


1

Have you noticed a degradation in performance since doing the upgrade? If not, the issue could be that you have been experiencing the issue all along but your previous IOS version wasn't reporting the errors. let's say you are experiencing a degradation in performance though. As you're seeing output drops this typically means that the links are experiencing ...


1

Take a look at scapy: http://www.secdev.org/projects/scapy/ and tcpdump/wireshark. With the former you can generate any ICMP packet, with the later you can see what your generated packets look like. Your router has certainly a default route. Remove it and it should generate the ICMP messages you are looking for. Alternatively, try this command on a host ...


Only top voted, non community-wiki answers of a minimum length are eligible