Chapter 4 - LAN, T1, and ISDN Interface Information and Statistics
This chapter describes configuration information and statistics that VNSM displays for a server's LAN, T1, and ISDN card interfaces. A LAN interface is the point at which a server accesses a LAN, such as an Ethernet card. A T1 interface is the point at which a server accesses a T1 wide-area communications link. An ISDN interface is the point at which the server accesses an ISDN interface.
This chapter discusses the following topics:
Accessing network interface information and statistics Viewing summary statistics for LAN, T1, and ISDN interfaces Viewing detailed statistics for LAN, T1, and ISDN interfaces
See your LAN documentation for more information on LAN interfaces. See the Banyan T1 Server-to-Server Guide for more information on setting up T1. See Managing Communications for more information on ISDN.
Accessing Network Interface Information and Statistics
To access network interface configuration information and statistics, choose SHOW communications statistics from the VINES Network Summary menu. The Communication Statistics menu appears, as follows.
This menu displays the LAN interfaces, T1 interfaces, and ISDN interfaces in the server; summary statistics for each interface; and the amount of time, since the last reboot of the server, that VNSM statistics have been accumulating. The sample menu above shows a server with a LAN card and a T1 interface.
The menu displays one entry for each LAN interface. This entry includes a description of the interface, the slot number of the interface's LAN card, and summary statistics.
For T1 interfaces, the menu displays an entry for the T1 card, such as Promptus T1 Access Module™, and for each Link Access Protocol D (LAPD) connection to other servers that the card supports. The T1 card entry also includes the slot number of the card and summary statistics.
Each LAPD entry includes a description (LAPD-[T1-E1]), a connection identifier in the Slot field, and summary statistics. The connection identifier in the Slot field consists of the slot number of the corresponding Promptus T1 card plus -n, where n is a number that identifies a T1 connection. Because the server supports only one connection per card at this time, the field displays -1. For example, if the LAPD interface's T1 card is installed in slot 4, the Slot field displays 4-1.
Each entry for an ISDN B channel interface includes a description, such as SCII ISDN BRI; a connection identifier in the Slot field; and summary statistics. The connection identifier in the Slot field consists of the slot number of the corresponding ISDN card plus -n, where n is a number that identifies an ISDN B channel. For example, if the ISDN card is installed in slot 3 and the card supports two ISDN B channels, the menu displays an entry for each channel. The Slot field in one entry displays 3-1. The Slot field in the other entry displays 3-2.
The menu also lets you perform the following actions:
Select LAN, T1 and ISDN interfaces to view detailed statistics. See "Viewing Detailed Interface Statistics" later in this chapter for more information. View serial line statistics. See Chapter 5 for more information. View protocol family statistics. See chapters 10, 11, 12, and 13 for more information. View configuration information and statistics on communication resource usage. See Chapter 6 for more information.
Viewing Summary Interface Statistics
The Communication Statistics Menu displays summary statistics for each LAN interface, LAPD interface, and ISDN interface in the server.
In some cases, summary totals do not apply to a particular interface. In these cases, "--" appears in the statistic field that does not apply.
The statistics are as follows:
Totin - Total number of frames received
Totout - Total number of frames sent
Ierrs - Total number of received frames that contained errors
Oerrs - Total number of sent frames that contained errors
Viewing Detailed Interface Statistics
You can view detailed performance data about interfaces on the selected server. To view this data, perform the following steps:
1. Choose SELECT for interface statistics from the Communication Statistics menu.
2. Choose the interface about which you want to view detailed information. The Interface Details screen for the specific interface appears. These statistics are described in the sections that follow.
Remember that you can press F10 to request an immediate update at a statistics screen.
Ethernet and StarLAN Interface Details
The fields listed in this section describe Ethernet and StarLAN™ traffic in detail.
Note: Whether all the fields described in this section are displayed depends on the specific Ethernet or StarLAN card. All the fields are displayed for most cards.
Total In and Total Out
Total number of Ethernet or StarLAN frames sent (Total Out) and received (Total In).
Input Errors and Output Errors
These statistics indicate the total number of Ethernet or StarLAN frames with errors sent (output errors) and received (input errors).
Some examples of input errors include:
Frame Check Sequence Errors Alignment Errors Overflows/Dribbles Errors
Some examples of output errors include:
Underflows Shorted Media
Collisions
The number of transmit collisions. You can expect to have a number of these occur, especially on a busy LAN. A large value for this field indicates poor performance for all members of the LAN.
To determine if an excessive number of collisions are occurring, use the following formula in conjunction with your understanding of traffic patterns on your LANs:
collisions/(total in + total out)
In almost all cases, a result of 0.02 or less means that an excessive number of collisions are not occurring. Depending on the traffic patterns on your LANs, any result above 0.02 could indicate that an excessive number of collisions are occurring.
If the interface is operating properly and the number of collisions is large, you should off load the LAN. Maintenance or any other disturbance of the LAN can cause a temporary flurry of these errors. A node attempting to transmit while the Ethernet or StarLAN network is not terminated properly or broken will cause hundreds of collisions within a few seconds.
Note: This statistic indicates the number of collisions that have been detected by the server. The number of collisions on the entire LAN will be higher.
Runts
Runts indicates the number of frames received that were truncated. Truncations can be caused by a LAN card with a failing transmit timer. Run diagnostics at each node on the LAN to obtain more information.
Unknowns
The total number of frames processed that have an unknown protocol type. In a network that includes only VINES servers, this value should be zero. Sometimes, a newly started Ethernet or StarLAN card produces one of these frames, causing a small number of unknowns to appear on other systems. Unknowns also appear if the server tries to process TCP/IP or AppleTalk packets, but does not have TCP/IP or AppleTalk installed.
In a mixed-vendor environment, unrecognized broadcast frames from the other equipment can arrive at the server frequently. A large value for unknowns can indicate that the other equipment is either addressing frames improperly, or sending a large number of broadcasts. Correcting this situation improves server performance.
ARP Packets In
Total number of TCP/IP Address Resolution Protocol (ARP) and AppleTalk Address Resolution Protocol (AARP) packets received.
For TCP/IP ARP, these packets are dropped if a VINES TCP/IP option is not installed, if this network interface does not have an IP address, or if ARP has been disabled for this interface.
For AARP, these packets are dropped if AppleTalk is stopped or if this AppleTalk port is disabled. Note that the terms port and interface are used interchangeably. Both terms refer to the server's physical connection to the network. Interface is commonly used in the TCP/IP community; port is the accepted term in the AppleTalk community.
ARP Packets Out
The number of TCP/IP ARP and AARP packets sent. No ARP packets are sent under the conditions described in the previous section.
The fields listed in this section describe Omninet™ traffic in detail.
Total In and Total Out
The total number of frames sent (Total Out) and received (Total In).
Input Errors and Output Errors
Input errors indicate that a card somewhere in the network is malfunctioning. Output errors can occur harmlessly if someone reboots or powers off a workstation while outstanding packets for that workstation have not yet been sent by the server.
Bad Result Codes
The number of times the card has returned an unrecognizable response to a command. This error should not occur unless the controller is malfunctioning.
Missed Interrupts
The total number of lost interrupts. Under heavy traffic, lost interrupts may occur and are self-correcting. With light or moderate traffic, rising values indicate a possible card malfunction.
Sends Refused
The number of frames sent that the destination card does not accept. When a sent frame is refused, it is resent multiple times before being discarded. This is not a fatal error. It simply reflects the level of activity in the network.
Drops
The number of frames not transmitted due to an excessive number of sends refused.
A large number of consecutive sends refused causes the VINES server to stop trying to transmit a frame, incrementing this counter by one. A significant value for this field may indicate a malfunctioning card somewhere else in the network (the one refusing the sends).
Output Q Overflow
The value that is incremented by one each time the maximum size of the output queue is exceeded. This can be due to excessive demand, but is more likely to be caused by a hung card. When the card is hung, you can gather this data only from the server console or a workstation on another interface on the same server.
Not Ready
Total number of times the card was not ready for the next command. While the card is receiving a frame off the wire, the card is not capable of accepting another command for a few milliseconds. If the card is still not ready after a significant period of time, the interface is restarted.
Unknowns
Total number of frames received with an unknown protocol type. This number tends to be larger in mixed-vendor networks than in networks with all VINES servers.
ARCNET and VistaLAN/PC Interface Details
The fields listed in this section describe ARCNET and VistaLAN™/PC traffic in detail.
Total In and Total Out
The total number of frames sent (Total Out) and received (Total In).
Sends Refused
The number of frames sent that are not accepted by the destination card. This is not a fatal error, but simply reflects the level of activity in the network.
Sends Dropped
The number of frames that are not transmitted due to an excessive number of sends refused.
A large number of consecutive sends refused causes the server to stop trying to transmit a frame, incrementing this counter by one. A significant value for this field may indicate a malfunctioning card.
Unknowns
Total number of frames received with an unknown protocol type.
Bad Size Packets
The total number of packets with an invalid frame size value in the frame header.
Range Finder Value
The value written to the timer chip on the card that is used by the RF range finder circuit on the card. This value indicates the number of milliseconds it takes to send a packet to the VistaLAN/PC hub and back. This value also indicates the time it takes for the card to transmit properly.
Recons Generated
The total number of reconfiguration bursts that the VistaLAN/PC card generates. The interface restarts if this number increases too quickly over a relatively short period of time.
IBM PC Network Interface Details
The fields listed in this section describe IBM PC Network traffic.
Total In and Total Out
The total number of frames sent (Total Out) and received (Total In).
Command Errors
The number of primary command errors.
Collisions
The number of transmit collisions. You can expect to have collisions, especially on a busy LAN. A large value in this field indicates poor overall LAN performance. If the interface is operating properly and the number of collisions is large, off load the LAN.
Re-transmits
The total number of frames retransmitted. You can expect occasional values in this field. Large values indicate degraded performance and a possible card malfunction.
Align Errors
The total number of alignment errors. Any significant value indicates a card malfunction.
CRC Errors
The number of Cyclic Redundancy Check (CRC) errors on receive. Any significant value in this field indicates a card malfunction.
CRC is a method of error checking that involves a series of calculations. The transmitting node places the results of a CRC calculation in the frame. When the frame is received, the receiving node recalculates the CRC values and compares its results with the results received from the sending node. If the values do not match, the frame is dropped.
Aborted Transmits
The number of times that the card aborted a transmission. A transmission can be aborted because of excessive collisions, or some other problem that the card detects.
Receiver Exhausted
The number of times the card recognized a frame addressed to it, but had no empty buffer space for the frame.
Output Q Overflow
The output queue overflow count. This value increments by one each time the maximum size of the output queue is exceeded.
Unknowns
The total number of frames received with an unknown protocol type.
Token-Ring Adapter and Token-Ring Bridge Details
This section describes statistics maintained by IBM Token-Ring adapters and VINES Token-Ring drivers. Many of the statistics described in this section apply to Token-Ring adapters from other vendors, such as the 3Com TokenLink (3C603) Token-Ring card.
The statistics in this section fall into two categories:
Adapter Details - These statistics apply to the Token-Ring adapter in the server.
Token-Ring Bridge Details - These statistics apply to the VINES Token-Ring Bridge option. These statistics do not appear if the option is not installed. The statistics appear for each IBM Token-Ring card that is configured for Token-Ring bridging.
Adapter statistics appear first on the screen. Token-Ring bridge statistics are listed after the adapter statistics.
Note: If you are retrieving Token-Ring bridge statistics from a VINES 4.xx server, each Token-Ring adapter that is configured for Token-Ring bridging is listed twice: once as a LAN entry for viewing adapter statistics and once as a Token-Ring bridge entry for viewing Token-Ring Bridge statistics. You must select the LAN entry to view adapter statistics and the Token-Ring bridge entry to select Token-Ring Bridge statistics.
The sections that follow describe adapter statistics and Token-Ring bridge statistics.
Each IBM Token-Ring adapter maintains two types of error counters: soft and hard.
Soft errors are intermittent, and are usually tolerated by the ring's error recovery procedures. The following types of soft errors are described in this section:
Line Errors Internal Errors Burst Errors A/C Errors Abort Delimiters
These errors usually reflect a problem with the adapter's ability to process received frames.
Hard errors are permanent faults that interfere with the operation of the ring.
An adapter transmits a beacon frame when it detects a hard error at the receiver side of its attachment. It continues to do so until its input signal is restored, or until it removes itself from the ring.
In such cases, you can suspect the beaconing adapter, its upstream neighbor (NAUN), and the cable path between them to be the fault domain. The beaconing adapter and its upstream neighbor, in turn, run self-diagnostic tests and remove themselves from the ring upon detecting any errors. Manual intervention is required if the ring cannot recover after these automatic tests.
Each ring contains one active monitor adapter. This adapter monitors token operation on the ring, and acts as the master clock for data transmission.
The sections that follow explain each of the fields in detail.
Total In and Total Out
The total number of frames sent (Total Out) and received (Total In).
Unknowns
The number of frames received with an unknown protocol type.
ARP Packets In
Total number of TCP/IP Address Resolution Protocol (ARP) and AppleTalk Address Resolution Protocol (AARP) packets received.
For TCP/IP ARP, these packets are dropped if a VINES TCP/IP Option is not installed, if this network interface does not have an IP address, or ARP has been disabled for this interface.
For AARP, these packets are dropped if AppleTalk is stopped or if this AppleTalk port is disabled. Note that the terms port and interface are used interchangeably. Both terms refer to the server's physical connection to the network. Interface is commonly used in the TCP/IP community. Port is the accepted term in the AppleTalk community.
ARP Packets Out
The number of TCP/IP ARP and AARP packets sent. No ARP packets are sent under the conditions described in the previous section.
Line Errors
The combined total of Frame Check Sequence (FCS) errors and code violations detected between the starting and ending delimiters of a frame or token. The FCS is a field in the frame that is used for error checking.
Internal Errors
The number of internal recoverable errors that the adapter detects. A non-zero count indicates an adapter in marginal operating condition.
Burst Errors
The number of times the adapter detected the absence of line signal transitions for five half-bit times between the starting and ending delimiters of a frame or token.
A/C Errors
The number of address-recognized (A) and frame-copied (C) errors that the adapter detects during the Neighbor Notification process (NAUN). This error indicates that the NAUN adapter is unable to set the A and C bits in a frame that it has copied.
Abort Delimiters
The number of times the adapter transmitted an abort delimiter. This occurs when the adapter detects an internal transient or hard error, or when it receives a token frame that is too long.
Lost Frames
The number of times a transmitted frame failed to return.
Congestion Count
The number of times the adapter recognized a frame addressed to it, but had no buffer space available for the frame. If this count is continually increasing and an IBM Token-Ring Adapter is being used, the additional memory on an IBM Token-Ring II Adapter or an IBM Token-Ring 16/4 Adapter may alleviate the problem.
Frame Copy Errors
The number of times the adapter copied a frame addressed to it, and another adapter on the ring had already copied the frame. This usually indicates either a line hit or a duplicate address.
Frequency Errors
The number of times the adapter detected a frequency error. A frequency error occurs when the ring clock maintained by the active monitor adapter differs from the clock frequency of the adapter by an excessive amount.
Token Errors
The number of times the adapter detected the need to transmit a new token. Only the active monitor adapter performs this function. (There is only one active monitor adapter per ring.) This count increases only if the IBM Token-Ring adapter in the server is the active monitor.
Packets Refused
The total number of frames that the destination adapter does not accept. This number can grow on the IBM Token-Ring interface without concern.
Packets Dropped
The total number of frames that are not transmitted because of excessive sends refused. Drops occur if a workstation is rebooted or powered down, and the server had outstanding frames addressed to it.
Ring Status Count
The total number of times the status of the ring has changed. This is a natural occurrence of the ring. Check the ring status statistic for the latest status that the adapter detects.
Last Beacon Type
The last beacon frame type that the adapter receives or transmits.The hexadecimal type code is interpreted as shown in Table 4-1.
Streaming refers to an adapter that is transmitting data continuously, or is receiving data without recognizing the ring protocol. Data received appears to be streaming data.
The monitor contention process selects an adapter to act as the new active monitor. This process starts when an adapter detects signal loss, a frequency error, the lack of an active monitor, or a malfunction in the active monitor.
Beaconing NAUN
The NAUN for the adapter that was last sending beacon frames. An adapter sends a beacon frame when it detects a serious network problem, such as a broken cable or malfunctioning adapter. By knowing the NAUN of the beaconing adapter, the source of the problem can be localized.
Ring Status
The ring status code is reported in hexadecimal. The bits are numbered 15 through 0 from left to right. If a bit is on (value 1), the condition described for that bit in Table 4-2 is true.
Other bits are reserved for future use. For more information, consult the Technical Reference Manual for the IBM Token-Ring Network PC Adapter.
Active Upstream NAU (NAUN)
Nearest Active Upstream Neighbor address. This is the address of the adapter that is sending frames or tokens directly to the adapter in the server.
This section describes the status and traffic statistics that are maintained for Token-Ring adapters that run the VINES Token-Ring Bridge option. These statistics appear for each Token-Ring card that is configured for Token-Ring bridging.
The Token-Ring Bridge option routes frames between the ring network connected to one adapter in the server and the ring network connected to the other adapter in the server. This lets the rings merge and act as a single Token-Ring network.
The source ring sends a frame to the target ring. The target ring receives a frame from the source ring. A bridge routes frames between the two rings. Both adapters in a bridge serve as a target ring and a source ring, depending on whether the adapter is sending or receiving a frame. For more information on Token-Ring bridges, see the VINES Token-Ring Bridge Option Guide.
The sections that follow describe each of the Token-Ring Bridge Interface fields in detail.
Single Route Broadcast
Tells whether the adapter is configured to forward single-route broadcast requests or not. If this field displays a value of 1, then single-route broadcasts are enabled. If the value is 0, they are disabled.
There are three types of Token-Ring broadcasts: single-ring, all routes, and single-route. Frames can broadcast within a ring, to all connected rings, or to all connected rings that have the single-route broadcast feature enabled.
When an adapter in the server has this option enabled, it transmits any single-route broadcasts that it receives. When an adapter in the server does not have this option enabled, it does not transmit single-route broadcasts.
Slot Number
Displays the slot number in which the Token-Ring adapter is installed in the server.
Bridge Number
Displays the bridge number that is configured for this adapter. Each server that acts as a Token-Ring bridge has a bridge number that is assigned to the two Token-Ring adapters in that machine. Both adapters in a single server must use the same bridge number.
Source Ring
Displays the source ring number assigned to this adapter.
The two Token-Ring adapters in a Token-Ring bridge are assigned a source ring number and a target ring number. The source ring number for one adapter must match the target ring number for the other adapter.
Target Ring
Displays the target ring number assigned to this adapter.
For the two Token-Ring adapters in a bridge, the target ring number for one adapter must match the source ring number of the other adapter.
Hop Count Limit
Displays the network hop count limit that is configured for this adapter. The hop count limit determines the number of consecutive Token-Ring bridges that a non-VINES, all-routes broadcast frame will travel. This adapter is included as one hop in the count.
Broadcast Frames Forwarded
Displays the total number of broadcast frames that this adapter received and then transmitted to a target ring. This value counts all-routes broadcasts and single-route broadcasts.
Broadcast Bytes Forwarded
Displays the total number of broadcast bytes that this adapter transmitted to a target ring.
Non-broadcast Frames Forwarded
Displays the total number of non-broadcast frames that this adapter received and then transmitted to a target ring. A non-broadcast frame is a frame that is addressed to a specific destination node.
Non-broadcast Bytes Forwarded
Displays the total number of non-broadcast bytes that this adapter transmitted to a target ring.
Frames not Forwarded
The total number of frames that this adapter did not successfully transmit to a target ring due to a ring problem, an adapter problem, or an expired Hop Count Limit that occurred between this adapter and the target ring. The ring reports to the adapter that is configured as the Source Ring whether or not a frame has been forwarded to the Target Ring.
VINES IP Intercepted
The total number of VINES IP packets that this adapter intercepted for transmission using VINES IP instead of transmission using the Token-Ring bridge.
When a server equipped with the Token-Ring Bridge option receives a VINES IP packet on one of the Token-Ring LANs that the server bridges, it intercepts it. By intercepting the packet, it prevents the bridge software from transmitting the packet, so that the packet is transmitted to the Target Ring only once by VINES IP.
The fields listed in this section describe ProNET-10 interface details.
Total In and Total Out
The total number of frames received (in) and sent (out).
Ierrs and Oerrs
Stands for input errors and output errors. For a ProNET-10 card, output errors are affected by stations entering and leaving the ring. They can become very large in a dynamic environment. Output errors do not include format errors.
Generally, output errors should not exceed 1.0 percent of the sum of the values in the Total In and Total Out fields. A higher rate of output errors indicates a faulty cable, connector, or LAN card. Run diagnostics to isolate the fault.
Input errors should not exceed 2.5 percent of the sum of the values in the Total In and Total Out fields. A higher rate of input errors indicates a faulty cable, connector, or LAN card. Run diagnostics to isolate the fault.
Parity Errors
The total number of parity errors on received frames. Parity errors should not exceed 1.0 percent of the sum of the values in the Total In and Total Out fields. A higher rate indicates a faulty cable, connector, or LAN card. Run diagnostics to isolate the fault.
Overruns
The total number of overrun errors on received frames (frame too long). This value increments each time the server's LAN card reads in a frame that does not have the required End-Of-Message marker. This can occur if the ring is initialized while the frame is in transit or if another card in the ring is failing.
These errors should not exceed 1.0 percent of the sum of the values in the Total In and Total Out fields. A higher rate indicates a faulty cable, connector, or LAN card. Run diagnostics to isolate the fault.
Format Errors
The total number of format errors on receive and transmit. This value is affected by nodes entering and leaving the ring and can become very large in a dynamic environment.
Each frame on the ring must have a well-defined format that contains:
Begin Message marker Source and destination address Data End Message marker Parity bit Refused bit
The format error value increments whenever a frame contains a field that the interface cannot identify, on either receive or transmit. A format error usually indicates that either signal noise or one of the LAN cards on the ring has corrupted the frame contents.
Format errors should not exceed 1.0 percent of the sum of the values in the Total In and Total Out fields. A higher rate indicates a faulty cable, connector, or LAN card. Run diagnostics to isolate the fault.
Sends Refused
The number of frames that the destination node does not acknowledge due to its receive buffer being busy. When a sent frame is refused, the server resends it multiple times before discarding the frame.
Sends Refused is not a fatal error, but simply reflects the level of activity in the network. It indicates transmission efficiency on the cable as compared with the Total Out value.
The ratio between the Sends Refused and the Total Out value increases if there is a timing mismatch between the server and the workstations. For example, if the server uses a high-speed bus such as those found in PS/2 and 386-based workstations, but the workstations do not use high-speed buses, the server will have a high number of retransmissions, as indicated by this field.
Sends Dropped
The number of frames that the server discards due to an excessive number of sends refused on those frames. A large number of consecutive sends refused for a single frame causes the server to discard the frame, incrementing the counter by one. A significant value for this field may indicate a malfunctioning card.
The Sends Dropped value is incremented under these conditions:
If the same frame was refused by its intended recipient 16 consecutive times. This can happen if a workstation hangs or loses power while data is being sent to it. If the server sees two consecutive ring errors, such as output errors, while attempting to send one frame to another node.
The Sends Dropped value does not have a 16-to-1 relationship to the Sends Refused value. The Sends Refused value increments every time any frame is refused, but the Sends Dropped value increments only when a single frame is refused 16 consecutive times.
If the Sends Dropped value becomes unusually high, take the following action:
Monitor workstation usage to see if users are powering off workstations without logging out or if users are rebooting workstations unnecessarily. If workstations are not being powered off or rebooted, check for an error on the ring. To determine the problem, run diagnostics and evaluate the Ring Pauses and Ring Breaks fields.
Ring Pauses
The number of times the server stopped monitoring the ring because of an excessive amount of input errors.
Excessive network errors can put a heavy load on the server - the server attempts to process the large amounts of corrupted data from the ring. If the server sees more than 100 input errors within a 200-millisecond period, it pauses and increments this value. A pause means the server stops receiving and transmitting frames for 400 milliseconds. The server is effectively ignoring the ring during this period, waiting for errors to abate. As a general rule, on small rings, pauses tend to be caused by workstations booting (rather than cable noise).
Assuming that a ProNET-10 ring contains n nodes, a pause value of n/2 or less per day is acceptable. A pause value of more than n/2 indicates that you should run diagnostics to isolate a faulty cable, connector, or LAN card.
Ring Breaks
The number of times the server had to break the ring attempting to correct the cause of several consecutive pauses.
A break occurs when the server has to pause after each of the last four 200-millisecond periods. In this case, the server forces its ProNET-10 card to leave and re-enter the ring and wait 1 second before attempting to use the ring again.
When the card leaves and re-enters the ring, the card generates a new token. If the server is forced to pause again in the first 200-millisecond interval after the 1-second waiting period, the server breaks again for a longer period.
On an excessively noisy ring, this process can continue until the break periods reach 2 minutes, causing many "File Volume Name@Group@Organization not available" and "File not found" messages to appear.
Breaks indicate a serious, persistent network-level problem that cannot be effectively handled by software retries and normal error correction techniques. The ratio of pauses to breaks indicates the persistence of the errors.
If the ratio of pauses to breaks is four or less, then errors are severe. This means that the server is in a condition in which it breaks the ring more than once after each group of four pauses. Too much effort is required to make the errors subside. Run diagnostics to isolate a faulty cable, connector, or LAN card.
No action is required if the ratio of pauses to breaks is greater than 12. If the ratio of pauses to breaks is between 4 and 12, then errors are moderate. You may want to run diagnostics to determine the cause of the errors.
Circulating Packets
The total number of times that a frame was detected as constantly circulating the ring without being removed by its sender. The server transmits a frame to clear the ring.
Ring Not OK
The total number of times the controller indicated that the ring was not operating properly after trying to send a frame. Ring Not OK usually indicates that a certain amount of time elapsed without detecting a token. This number reports only those cases where a lost token caused the transmission of a frame to fail.
Output Timeouts
The total number of times the card had to wait too long to send a frame. This error is usually accompanied by a Ring Not OK error.
Unknowns
The total number of frames received with an unknown protocol type.
ARP Packets In
The number of TCP/IP Address Resolution Protocol (ARP) packets received. No ARP packets are received under the following conditions:
VINES TCP/IP option is not installed on the selected server. ARP is disabled for the selected interface. No ARP address is assigned to the selected interface.
ARP Packets Out
The number of TCP/IP ARP packets sent. No ARP packets are sent under the conditions described in the previous section.
Each Token-Ring Adapter maintains two types of error counters: soft and hard.
Soft errors are intermittent, and are usually tolerated by the ring's error recovery procedures. The following soft errors are described in this section:
Line Errors Burst Errors ARI/FCI Errors Abort Delimiters
Soft errors usually reflect a problem with the adapter's ability to process received frames.
Hard errors are permanent faults that interfere with the operation of the ring.
An adapter transmits a beacon frame when it detects a hard error at the receiver side of its attachment. It continues to transmit beacon frames until its input signal is restored, or until it removes itself from the ring.
In such cases, you can suspect the beaconing adapter, its upstream neighbor (NAUN), and the cable path between them to be the cause of the fault. The beaconing adapter and its upstream neighbor, in turn, run self-diagnostic tests and remove themselves from the ring upon detecting any errors. Manual intervention is required if the ring cannot recover after these automatic tests.
Each ring contains one active monitor adapter. This adapter monitors token operation on the ring, and acts as the master clock for data transmission.
The sections that follow explain each of the fields in detail.
Total In and Total Out
The total number of frames received (in) and sent (out).
Ring Status
The ring status code is reported in hexadecimal. The bits are numbered 15 through 0 from left to right. If a bit is on (value 1), the condition described for that bit in Table 4-3 is true. Other bits are reserved for future use.
Unknowns
The number of frames received with an unknown protocol type.
Line Errors
The combined total of Frame Check Sequence (FCS) errors and code violations detected between the starting and ending delimiters of a frame or token. The FCS is a field in the frame that is used for error checking.
Burst Errors
The number of times the adapter detected the absence of line signal transitions for five half-bit times between the starting and ending delimiters of a frame or token.
ARI/FCI Errors
The number of address-recognized (ARI) and frame-copied (FCI) errors that the adapter detects during the Neighbor Notification (NAUN) process. This error indicates that the NAUN adapter is unable to set the A and C bits in a frame that it has copied.
Lost Frames
The number of times a transmitted frame failed to return.
Congestion Count
The number of times the adapter recognized a frame addressed to it, but had no buffer space available for the frame.
Frame Copy Errors
The number of times the adapter copied a frame addressed to it, and another adapter on the ring had already copied the frame. This usually indicates either a duplicate address or, more likely, a line hit.
Token Errors
The number of times the adapter detected the need to transmit a new token. Only the active monitor adapter performs this function. There is only one active monitor adapter per ring. This count increases only if the ProNET-4 card in the server is the active monitor. This is usually the case unless VINES is not running.
Sends Refused
The total number of frames that the destination card does not accept.
Ring Status Count
The total number of times the status of the ring has changed. This is a natural occurrence of the ring. Check the Ring Status statistic for the latest status that the adapter detects.
ARP Packets In
The number of TCP/IP Address Resolution Protocol (ARP) packets received. No ARP packets are received under the following conditions:
A VINES TCP/IP option is not installed on the selected server. ARP is disabled for the selected interface. No ARP address is assigned to the selected interface.
ARP Packets Out
The number of TCP/IP ARP packets sent. No ARP packets are sent under the conditions described in the previous section.
This section describes statistics for the NT LANSTAR™ PC card.
Total In and Total Out
The total number of NT LANSTAR PC frames received (in) and sent(out).
CRC Errors
The number of Cyclic Redundancy Check (CRC) errors on receive. CRC is a method of error checking that involves a series of calculations. The transmitting node places the results of a CRC calculation in the packet. When the packet is received, the receiving node recalculates the CRC values and compares its results with those received from the sending node. If the values do not match, a request for retransmission is sent to the sending node.
Any significant value in this field indicates a malfunction.
Runts
The total number of frames received that were too short to be recognizable.
Unknowns
The number of frames received with an unknown protocol type or destination class.
Overruns
An overrun occurs when data arrives faster than the NT LANSTAR PC receiver can process it. Any significant value in this field indicates a malfunction.
Unknown Bcasts
The number of broadcast frames received that the interface does not recognize. The only broadcast frames that the interface recognizes are NT LANSTAR PC broadcast packets and VINES IP broadcast packets.
If MS-NET is running on the same LAN, the number in this field will be high. The NT LANSTAR PC interface does not recognize MS-NET broadcasts.
Output Q Overflow
An output queue overflow count. This value increments by one each time the maximum size of the output queue is exceeded. A significant number of output queue overflows indicates a malfunction.
Drops
The total number of incoming packets that were dropped due to insufficient available communications buffer space.
Disconnects
The total number of times that this server was disconnected from the PTE.
The fields listed in this section describe LocalTalk™ traffic through the DaynaTALK™ card in detail. The statistics screen for this card displays the physical address of the card in the Node Address field. This address is a 1-byte address.
The statistics that the screen displays are described in the sections that follow.
Total In
The total number of LocalTalk frames received.
Input Errors
The total number of received LocalTalk frames that contained errors.
Total Out
The total number of LocalTalk frames that the card has sent.
Output Errors
The total number of LocalTalk frames that could not be sent because they contained errors.
T1 Card and LAPD Connection Details
This section describes statistics on network traffic through T1 cards, such as the Promptus T1 card, and LAPD statistics on T1 line activity. LAPD is the link-level protocol that servers use to communicate on T1 lines. VNSM keeps a set of card statistics for each T1 card in the server and a set of LAPD statistics for each LAPD server-to-server connection. At this time, a T1 card supports only one T1 connection to another server.
To access statistics for the T1 card, perform the following steps:
1. Choose SELECT for interface statistics from the Communication Statistics menu.
2. Choose the card interface, such as Promptus T1 Access Module.
At the next menu, you can view the card's NetId, which corresponds to the server's serial number. Ignore the Node Address field, which does not apply to T1 communications. You can also choose one of the following options to view other information:
Configurable Parameters - The T1 card's configuration parameters
Statistics - Statistics on T1 card activity that are generated by the T1 driver
Local ESF Statistics - Extended SuperFrame (ESF) statistics on line activity that are kept by the server
Telco ESF Statistics - ESF statistics on line activity that are kept by the T1 service provider
The sections that follow describe the configuration information and statistics that these options let you display.
Configurable Parameters
This section describes the T1 configuration parameters displayed by VNSM.
Network interface type can be DSX-1 or CSU. A DSX-1 interface is used when the server has an external channel service unit (CSU) or when a CSU is not required. CSUs are not required when the distance between servers is 1300 feet or less.
A CSU interface is used when the T1 card in the server has a built-in CSU.
Timing Mode
Timing mode can be either Master or Slave. When the T1 card is in master mode, the server is providing clocking to the server on the other end of the T1 line. The server on the other end of the line must be in slave mode. Master mode should be configured only in private networks.
If you have contracted with a service provider, both of the servers that are connected by the T1 line get their clocking from the provider. In this case, both servers should be in slave mode.
Line Coding
Line coding can be either AMI (Alternate Mark Inversion) or B8ZS (Bipolar 8 Zero Substitution). AMI supports 56 Kbps per channel. B8ZS supports 64 Kbps per channel. The line coding depends on what the T1 service provider offers.
Framing
Framing can be either ESF (Extended SuperFrame) or D4 (SuperFrame), depending on what the service provider offers. ESF allows the server to exchange line performance information with the service provider.
Remote Loopback
This parameter, either enabled or disabled, determines whether remote loopback is inhibited. It applies only to T1 cards that have a built-in CSU. When remote loopback inhibition is disabled, the service provider can put the card into loopback mode, which allows the provider to run diagnostic tests on the card. If remote loopback inhibition is enabled, the service provider cannot run diagnostic tests on the card.
Line Buildout
This parameter is the line buildout switch setting on the CSU in the T1 card. The setting, in decibels, can be either 0.0, 7.5, or 15.0.
Line Buildout (DSX1/CEPT only)
This parameter is the line buildout setting for the DSX-1 interface. If the DSX-1 is connected to an external CSU, this parameter specifies the length, in feet, of the cable between the DSX-1 and the CSU. If the DSX-1 is connected directly to a DSX-1 on another server, this parameter specifies the length, in feet, of the cable between the two DSX-1 interfaces.
If the T1 card has a built-in CSU, this parameter should be set to 0-133 feet.
Loopback Mode
This parameter specifies the local loopback mode (enabled or disabled). When loopback is enabled, you can run diagnostic tests on your T1 card. When it is disabled, you cannot run diagnostic tests on your T1 card.
This section describes the T1 driver statistics displayed by VNSM.
Xmit Zeros (CSU Only)
This statistic only applies to T1 cards that have a built-in CSU and that use Alternate Mark Inversion (AMI) line coding. If the status is Active, the T1 card is transmitting bytes that contain all 0 bits, which results in timing problems. AMI requires that each transmitted byte have at least a single 1 bit to ensure proper timing. If the status is Inactive, the T1 card is not transmitting all-zero bytes.
A possible cause of the Active status error is the use of normal HDLC as the protocol on the T1 link. You should use inverted HDLC as the protocol on T1 links that support AMI.
If the T1 line uses Bipolar 8 with Zero Substitution (B8ZS), this field always displays Active because B8ZS allows all-zero bytes.
Ignore this field if the T1 card does not have a built-in CSU. See the description of the Network Interface Type field in the preceding section, "Configurable Parameters," for information on determining whether a T1 card has a built-in CSU.
Xmit Density Violation (CSU Only)
This statistic applies only to T1 cards that have a built-in CSU and that use Alternate Mark Inversion (AMI) line coding. If the status is Active, the T1 card is failing to meet ones-density requirements for AMI data links. If the status is Inactive, the T1 card is meeting these requirements.
A possible cause of the Active error is the use of normal HDLC as the protocol on the T1 link. You should use inverted HDLC as the protocol on T1 links that support AMI.
If the T1 line uses Bipolar 8 with Zero Substitution (B8ZS), this field always displays Active because B8ZS does not require ones-density.
Ignore this field if the T1 card does not have a built-in CSU. See the description of the Network Interface Type field in the preceding section, "Configurable Parameters," for information on determining whether a T1 card has a built-in CSU.
CSU Remote Loopback (CSU Only)
This statistic applies only to T1 cards that have a built-in CSU. If the status is Active, the service provider has put the T1 card into remote loopback mode in order to run diagnostic tests. Otherwise, the card is not in loopback mode.
If the Remote Loopback Inhibit configuration parameter is set to enabled, this statistic is never Active.
Ignore this field if the T1 card does not have a built-in CSU. See the description of Network Interface Type field in the preceding section, "Configurable Parameters," for information on determining whether a T1 card has a built-in CSU.
Red Alarm
If the status is Active, the card has lost frame synchronization continuously from the network for more than 2.5 seconds, or synchronization has been intermittent. Otherwise, the card has not lost synchronization.
Red alarms automatically generate yellow alarms.
Yellow Alarm Detected
If the status is Active, the other end of the T1 connection has lost synchronization. Otherwise, synchronization has not been lost.
Loss of Signal (CSU Only)
This statistic applies only to T1 cards that have a built-in CSU. If the status is Active, the card detected complete loss of carrier. If the status is Inactive, the card did not detect complete loss of carrier.
Local and Telco ESF Statistics
This section describes Extended SuperFrame (ESF) statistics on T1 line activity. ESF statistics are kept both by the server (local) and by the service provider (Telco). Both the server and the service provider keep the same statistics.
There is a set of ESF statistics for each 15-minute period in the previous 24 hours. For each period, the following statistics appear:
Err Secs
The number of seconds during the period in which a Cyclic Redundancy Check (CRC) error occurred. For example, if this field displays 5, there were five 1-second periods in which at least one CRC error occurred.
UA Secs
The number of seconds in the 15-minute period in which the network was down (a red or yellow alarm occurred).
Valid Secs
The number of seconds in the 15-minute period that have expired. For all time periods except the current one, this field displays 900 seconds (15 minutes).
This section describes statistics on LAPD activity on the T1 connection. LAPD is the link-level protocol that servers use to communicate over T1 lines. The server keeps a set of LAPD statistics for each T1 server-to-server connection. At this time, a T1 card supports only one connection to another server.
Transmitted Packets
The total number of packets that the server has transmitted over the T1 connection.
Received Packets
The total number of packets that the server has received over the T1 connection.
Bytes Transmitted
The total number of bytes that the server has transmitted over the T1 connection.
Bytes Received
The total number of bytes that the server has received over the T1 connection.
Link State
This field indicates whether the T1 connection is up (Circuit Up/LAP up) or down (Circuit Up/LAP down). You should never see Circuit Down.
The fields listed in this section describe ISDN traffic on ISDN B channels in detail. The statistics that the screen displays are described in the sections that follow.
Total In
The total number of ISDN frames received.
Input Errors
The total number of received ISDN frames that contained errors.
Total Out
The total number of ISDN frames that the card has sent.
Output Errors
The total number of ISDN frames that could not be sent because they contained errors.
