D-Link DGS-3100-48, Руководство пользователя

Страница: 43 / 293

DGS-3100 Series Gigabit Stackable Managed Switch User Manual
27
Unit ID 1, and serving as the stack Backup master, while the already running stacking member with Unit ID 2
remains the active Stack Master.
•
The Stack Master performs Unit ID Allocation and Conflict Resolution process.
– If the new stacking member is in the
Factory Default
mode, the new stacking member is assigned the lowest
available Unit ID by the Stack Master. It is strongly recommended that
Auto Assign
mode is used to assign the
Unit ID. The
Auto Assign
mode provides better stack resiliency.
– If the new stacking member already has an assigned Unit ID, and that Unit ID is unused in the current stack,
the incoming stacking member is assigned Unit ID. The Stack Master applies any device configuration to the
new stacking member.
– If the new stacking member already has an assigned Unit ID, and that UnitID conflicts with an existing Unit
ID the Stack Master reallocates a new Unit ID to the new stacking member. The lowest available Unit ID is
applied to the new stacking member. This occurs only if new stacking member does not have a manually
assigned Unit ID, which the Stack Master cannot change.
– If the new stacking member cannot be assigned an available Unit ID, then the new stacking member is
effectively shut down and cannot join stack. For example, if stacking member replacement that can only occur
if the new stacking member has a manually assigned Unit ID.
•
The Stack Master performs Unit
and Stacking Port Configuration
for the new stacking member.
•
Any configuration information the Stack Master retains that is relevant to the Unit ID of the new stacking member
is applied. If the new stacking member was assigned the same Unit ID of the replaced stacking member, then the
new stacking member receives the same switch configuration as the failed stacking member, described in
Replacing Failed Stacking-Members in a Running
Stack
.
Dividing Stacks
This example assumes that a working stack is divided into two groups. The stack is divided either by a failed stacking link
connected to two stacking members in the stack or by a failed stacking members in a chain topology which causes
disconnection between two units in the stack. In this case we should consider each sub-group as an independent running
stack configuration. For each sub-group we should consider three sub options:
•
Both the Stack Master and the Backup Master are part of the sub-group.
•
Either the Stack Master or the Backup Master is part of the sub-group.
•
Neither the Stack Master nor the Backup Master is part of the sub-group.
When a stack is split into two parts, the following occurs in each partial stack according to the following scenario:
Both the Stack Master and the Backup Master are part of the sub-group.
Nothing changes, except the Stack Master recognizes the missing stacking members as removed stacking members and
routes traffic around them.
Since both the Stack Master and Backup Master are in this stacking section, this section is operating and the other section
cannot operate.
The following occurs when the stack is divided and both the Master and Backup master are in the sub-group:
•
The Master Discovery, Master Election and Unit ID Allocation & Duplicate Unit ID Conflict Resolution processes
are performed, resulting in the following:
– Any configuration information stored by Stack Master, which remained in the group, that is relevant to the
stacking members remains unchanged.
– Topology information, inter-stacking member forwarding information for transmitting traffic to any other
stacking member, managed by the Stack Master includes only stacking members that remain connected after
the stack is divided.
–
The divided stack continues to operates normally, the only difference is there are less stacking members than
prior to the stack division.
–
No Unit ID changes are performed in each divided stack.
–
The Stack Master notifies the system administrator using SYSLOG messages and SNMP traps of the removed
stacking members. In addition the Stack Master also notifies the system administrator whichd ports belong to
unreachable stacking members and are reported as
Not Present
.
•
Either the Stack Master or the Backup Master remain in the divided group