Table of Contents
Advertisement
JUNOS OS 10.4 Release Notes
62
Services Applications
In scaled environments, the thread in the Multiservices PIC or DPC for cflow might run
too long. This causes the PIC or DPC to crash. [PR/494457: This issue has been
resolved.]
On Multiservices 500 PICs with graceful Routing Engine switchover, wrong record
values are seen for the IPv4 netflow export packets. This error occurs when the route
records does not get installed. [PR/545422: This issue has been resolved.]
The MS400 PIC crashes due to a memory allocation failure when the PIC tries to
respond to a Routing Engine CLI request. [PR/558237: This issue has been resolved.]
The MultiServices PIC might crash when traffic is received on a Layer 2 Tunneling
Protocol (L2TP) session (MLPPP bundle), and a teardown request is also received at
the same time. [PR/561039: This issue has been resolved.]
If Bidirectional Forwarding Detection protocol (BFD) protection for BGP sessions is
configured on a BGP session in a nonmaster routing instance, the BFD might start for
that session before the kernel ID of the routing instance is set. This might cause the
BFD session to freeze. As a workaround, if the BFD session has the routing table value
of 4294967295, use the clear bfd session command to start a new session that will
address the issue as long as the routing instance's kernel table has been allocated.
[PR/563161]
If a class-of-service rule is applied to a service set , the inactive timeout under the
user-configured application does not take effect. As a workaround, match the
application in the class-of-service rule. [PR/571304: This issue has been resolved.]
User Interface and Configuration
Under certain circumstances, a nested Junos OS configuration group with a wildcard
match might not have the desired effect. [PR/556379: This issue has been resolved.]
When a "validate" RPC is executed using a NETCONF session, some essential
information about the session is not populated in the configuration database.
[PR/570778]
VPNs
In MVPN routing-instances with local-receivers, a flood nexthop is created for each
S,G entry for multicast traffic received from the CE. Once the local-receivers are joined
or pruned, a new flood nh is created. However, old flood nexthops are not deleted. This
leads to a memory leak within the routing protocol process. Once this routing protocol
process reaches a size of 2GB, it will trigger an assertion and restart. [PR569621: This
issue has been resolved.]
In local-switched l2circuit scenario, the control and forwarding plane might not be
properly updated by the routing protocol process when one of the logical interfaces
forming an l2ckt is down. [PR/572780: This issue has been resolved.]
Copyright © 2011, Juniper Networks, Inc.
Advertisement
Table of Contents
