Cisco Notification Alert -3750X Series Switch-01-Jul-2015 16:48 GMT

Feedback Sign Up Unsubscribe         Known Bugs - Catalyst 3750-X Series Switches Bug Id: CSCun51532 Title: Sometimes C8945 with port synchronisation lose SW link on PC link Description: Symptom: The link is up but there is speed mismatch between both ends of the link. This does not allow traffic to flow. Conditions: This can be seen when a large no of ports (~20) observe link change event at the same time. Workaround: Do a "shut" and "no shut" of the port which is affected. Further Problem Description: "shut" and "no shut" will toggle the link and clear the speed mismatch on either side of the link. Status: Fixed Severity: 2 Severe Last Modified: 05-JUN-2015 Known Affected Releases: 15.0(2)SE2 Known Fixed Releases: 15.0(2)EX8, 15.0(2)SE7, 15.2(1.41)PSR, 15.2(2)E, 15.2(2.2.32)EA, 15.2(2b)E, 3.6(0)E Bug Id: CSCui65252 Title: ARP packets processing stops when DAI is enabled over Port-Channel Description: Symptom: Enabling DAI with Port-Channel breaks ARP processing DHCP snooping on port-channel breaks. Conditions: 1. Configure the required for DAI, DHCP snooping 2. Setup an DHCP server 3. Setup an Cross-Stack Port-Channel between DHCP and Access switch 4. Enable arp inspection for the client vlan 5. Bring up the Client interface 6. Client would be able to get an ip address, binding added to snooping table Run a continous ping from client 7. From the client side clear the arp binding (arp -d) was used 8. Once the ARP cache was cleared from client; client traffic drops 9. Binding table still holds the entry for the client 10.Tried Release / Renew, works fine client can renew ip address still traffic drops 11.Disable "ip arp inspection" for the client vlan - traffic resumes 12.Enable ip arp inspection again ; traffic drops Workaround: When a direct link is used between Access and DHCP issue is not seen. Further Problem Description: The issue is seen only with port-channels. Status: Fixed Severity: 2 Severe Last Modified: 08-JUN-2015 Known Affected Releases: 12.2(55)SE8, 15.0(2)SE4, 15.0(2)SE5, 15.2(1)E Known Fixed Releases: 12.2(55)SE9, 15.0(2)SE5, 15.0(2)SE6, 15.2(1)E1, 15.2(1)E2, 15.2(1)E3, 15.2(2)E, 15.2(2b)E Bug Id: CSCut84437 Title: Ping failed after added "permit host XXX any" at MAC ACL Description: Symptom: Ping failed after added "permit host XXX any" at MAC ACL. Conditions: Adding "permit host XXX any" when "permit host XXX any 0x806 0x0" already exists. Workaround: For every vlan filter create a seperate acl filled with all arp permits that are required: Create one permit arp ACE for every permit ACE in the other ACLs on the vlan fliter. Attach the acl with arp ACEs on to the vlan filter. e.g., if permit host XXX any is an 'ACE' in the vlan filter Create a mac extended ACL Arp_Fix_Acl with permit host XXX any 0x806 0x0 attach the Arp_Fix_Acl in the vlan filter. Further Problem Description: Status: Terminated Severity: 2 Severe Last Modified: 08-JUN-2015 Known Affected Releases: 15.0(2)SE6 Known Fixed Releases: Bug Id: CSCus56320 Title: ping packet loss occur intermittently when a new member add into stack. Description: Symptom: intermittent ping packet loss occurred. Conditions: a new switch(C3750X) add into stack. Workaround: none. Further Problem Description: Status: Terminated Severity: 2 Severe Last Modified: 08-JUN-2015 Known Affected Releases: 15.0(2)SE6 Known Fixed Releases: Bug Id: CSCur17365 Title: 15.2(2)E:CSCup68355 verification errdisabled pagp-flap b/w 3750x & 2960s Description: Symptom: ports in port channel go to error disabled state due to pagp flap. Conditions: during formation of lacp portchannel bundling the ports go to the error disabled state when switch is having more number of stack members say 5. Workaround: do shut and no shut on the physical interfaces which are error disabled so that It will recover and come to up state and will be part of port channel Further Problem Description: During formation of lacp portchannel bundling the ports go to the error disabled state when switch is having more number of stack members. Status: Fixed Severity: 2 Severe Last Modified: 09-JUN-2015 Known Affected Releases: 15.2(2.4.54)E1 Known Fixed Releases: 15.2(2)E2, 15.2(3)E1, 3.6(2)E, 3.7(1)E Bug Id: CSCus23125 Title: 3750x: MAC not learnt after removal of auth config Description: Symptom: IP Phone is connected to a port configured for mab & dot1x. IP Phone authenticates through mab and gets an IP in voice vlan fine. User defaults interface config and adds access vlan + voice vlan config. Switch is unable to dynamically learn the mac address of the phone in voice vlan. Keeps on adding and deleting mac entry in data vlan. ARP for phone remains incomplete on switch unless 3750X is reloaded. Conditions: NA Workaround: Remove and reconfigure voice vlan Further Problem Description: Status: Fixed Severity: 2 Severe Last Modified: 12-JUN-2015 Known Affected Releases: 15.2(1.1), 15.2(1.2), 15.2(1.3), 15.2(2.0.0), 15.2(2.1), 15.2(22.22), 15.2(3)E, 15.2(3.7.1), 15.2(3.7.2), 15.2(4.0.1) Known Fixed Releases: 15.2(3)E1, 3.7(1)E Bug Id: CSCud55483 Title: TS:REP ring lost multicast traffic if ring is disrupted:Cat3750X& IE2000 Description: Symptom: REP ring multicast traffic convergence failed when the REP segment is disrupted in Cat3750X. Conditions: Intermittent issue Affected software release: 15.0(2)SE, 15.0(2)SE1, 15.0(02)SE2. Workaround: none Status: Fixed Severity: 2 Severe Last Modified: 12-JUN-2015 Known Affected Releases: 15.0(2)SE Known Fixed Releases: 15.0(2)EJ, 15.0(2)EJ1, 15.0(2)SE3, 15.0(2)SE4, 15.0(2)SE5, 15.0(2)SE6, 15.2(1)E, 15.2(1)E1, 15.2(1)E2, 15.2(1)E3 Bug Id: CSCuu67162 Title: 1000M port link up with 100M after OIR Description: Symptom: The customer did OIR of the interface, after inserted back, the ports link up with 100M. Conditions: OIR is exexuted. Workaround: do OIR again may make the link up with 1000M again. Further Problem Description: Status: Open Severity: 2 Severe Last Modified: 13-JUN-2015 Known Affected Releases: 15.0(2)SE5 Known Fixed Releases: Bug Id: CSCuu56067 Title: 3750 Dual SDM template incorrectly allocates QoS TCAM Description: Symptom: 3750x running 12.2(55)SE5 can fail to apply a policy-map though there is room in TCAM . The following error message is generated on this failure: Master SVI send all failed for interface Vlan100, cmd = 38 Conditions: 3750x running 12.2(55)SE5 configured with "desktop IPv4 and IPv6 vlan" SDM template. This is seen when using policy-maps with a large amounts of access-lists Workaround: Use the default SDM template (for ipv4 only based qos policies) Limit the size of the policy-map and access-lists Further Problem Description: Status: Open Severity: 2 Severe Last Modified: 18-JUN-2015 Known Affected Releases: 12.2(55)SE5 Known Fixed Releases: Bug Id: CSCuu47539 Title: Stack port becomes root port after reload on 3750X Description: Symptom: Stack port shows up as root port for some or all vlans after a reload. Reload could be member or whole stack Logs similar to below are see for the actual root port Mar 30 04:51:47.128: %DTP-5-TRUNKPORTON: Port Gi4/1/1-Gi4/1/1 has become dot1q trunk (3750-stack-4) Mar 30 04:51:47.439: %DTP-5-NONTRUNKPORTON: Port Gi4/1/1 has become non-trunk (3750-stack-4) Mar 30 04:51:50.576: %LINK-3-UPDOWN: Interface GigabitEthernet4/1/1, changed state to up Conditions: The uplinks should be fiber connections configured as trunk "switchport nonegotiate" should be configured on it Workaround: remove "switchport nonegotiate Further Problem Description: Issue is same as reported in CSCue03558. But that bug was only fixed for 3750V2 The issue is being seen in 3750X as well The test was done with 15.0(2)SE7 Status: Fixed Severity: 2 Severe Last Modified: 21-JUN-2015 Known Affected Releases: 15.0(2)SE6 Known Fixed Releases: 15.2(4.10.61)PI5 Bug Id: CSCuc78173 Title: 3750X-12S crashes with 15.0(2)SE CCO image Description: Symptom: 3750X-12S is crashing with 15.0(2)SE image at boot-up time. Currently, this issue is seen intermittently and it is not consistently observed. Issue is not seen on other 3750-X SKU's Many of this case is observed in the combination of C3750X-12S with 15.0(2)SE. But in a few cases, it is also observed in other C3750X or 15.0(1)SEx. Conditions: This issue occurs at boot-up and there is no specific configuration which cause this issue. Workaround: Right now, there is no work-around available. Status: Fixed Severity: 2 Severe Last Modified: 23-JUN-2015 Known Affected Releases: 15.0(2.0.99)SE Known Fixed Releases: 15.0(2)EJ, 15.0(2)EJ1, 15.0(2)EX, 15.0(2)EX1, 15.0(2)EX3, 15.0(2)EX4, 15.0(2)EX5, 15.0(2)EZ, 15.0(2)SE1, 15.0(2)SE2 Bug Id: CSCuc47185 Title: WS-C3750X-12S keeps crashing after IOS upgrade to 15.0(2)SE Description: Symptom: WS-C3750X-12S keeps crashing due to Macsec POST failure after IOS upgrade to 15.0(2)SE. Conditions: WS-C3750X-12S after IOS upgrade to 15.0(2)SE Workaround: Go back to previous IOS Note that this fix has been merged to CSCuc78173. Status: Fixed Severity: 2 Severe Last Modified: 23-JUN-2015 Known Affected Releases: 15.0(2)SE Known Fixed Releases: 15.0(2)EZ, 15.0(2)SE1, 15.0(2)SE2 Bug Id: CSCuu55421 Title: Error - The FRULink 10G Service Module (C3KX-SM-10G) is not responding Description: Symptom: %PLATFORM_SM10G-3-NO_RESPONSE: The FRULink 10G Service Module (C3KX-SM-10G) is not responding. Seen in the log after service policy is applied to a VLAN carried on an interface on the SM Conditions: C3KX-SM-10G installed in a 3750X Switch#sh switch service-module Switch/Stack supports service module CPU version: 03.00.78 Temperature CPU Switch# H/W Status (CPU/FPGA) CPU Link Version ----------------------------------------------------------------- 2 OK 67C/63C notconnected 03.00.78 Workaround: none Further Problem Description: Status: Open Severity: 2 Severe Last Modified: 24-JUN-2015 Known Affected Releases: 15.0(2)SE6 Known Fixed Releases: Bug Id: CSCum86316 Title: %ENTROPY-0-ENTROPY_ERROR: Unable to collect sufficient entropy on bootup Description: Symptom: c2960s OR c3750x will print Below errors some times. Mar 30 01:30:04.263: %ENTROPY-0-ENTROPY_ERROR: Unable to collect sufficient entropy Mar 30 01:31:04.275: %ENTROPY-0-ENTROPY_ERROR: Unable to collect sufficient entropy Mar 30 01:32:04.279: %ENTROPY-0-ENTROPY_ERROR: Unable to collect sufficient entropy Mar 30 01:33:04.283: %ENTROPY-0-ENTROPY_ERROR: Unable to collect sufficient entropy Mar 30 01:34:04.278: %ENTROPY-0-ENTROPY_ERROR: Unable to collect sufficient entropy Mar 30 01:35:04.290: %ENTROPY-0-ENTROPY_ERROR: Unable to collect sufficient entropy Conditions: Happens only in the c3750x and c2960s which are using ACT1 base Entropy. Not seen in other models. This does not happen all always and on all the units. Happens only in some units. Workaround: None Further Problem Description: Status: Fixed Severity: 2 Severe Last Modified: 25-JUN-2015 Known Affected Releases: 15.4(2.2)T Known Fixed Releases: 15.2(1.30)PSR, 15.2(2)E, 15.2(2.2.32)EA, 15.2(2b)E Bug Id: CSCur07371 Title: CPU queue wedge on 3750x Description: Symptom: A CPU queue wedge has been identified on the 3750X. Conditions: This defect has been seen soon after upgrading all 3750X stacks to 15.2(2)E. After an arbitrary amount of time, the switch begins to show the following syslog message %SUPQ-4-CPUHB_RECV_STARVE: Still seeing receive queue stuck after throttling (124A3701-3) The stack may become inaccessible. Workaround: The only known resolve is to reload the switch stack. Further Problem Description: Status: Open Severity: 1 Catastrophic Last Modified: 26-JUN-2015 Known Affected Releases: 15.2(2)E Known Fixed Releases: Bug Id: CSCuv03089 Title: Switch crashes removing IPv6 raguard Description: Symptom: The switch crashes trying to free a bad chunk of memory. Additionally, this crash may lead to the switch hanging and not reloading. This is caused by the switch trying to write the crashinfo file. If "no exception crashinfo" is configured, the switch will dump a stack trace to the console, and then reload. Conditions: This occurs if a switchport has the following configured on it: ipv6 nd raguard And the switchport's configuration is defaulted (e.g., default int gi1/0/1) or the config for ipv6 nd raguard is removed. Workaround: Shutdown the port first prior to removing raguard. Further Problem Description: Status: Open Severity: 2 Severe Last Modified: 30-JUN-2015 Known Affected Releases: 15.0(2)SE Known Fixed Releases: Bug Id: CSCur74702 Title: Wrong SMI vStack group selected due to incorrect client MAC matched Description: Symptom: The SMI client loads the configuration from an incorrect vStack group, which breaks Zero Touch Deployment. Conditions: This is seen on SMI client switches that have MAC addresses that differ in the last byte only. Workaround: There is no workaround. Further Problem Description: Status: Fixed Severity: 2 Severe Last Modified: 30-JUN-2015 Known Affected Releases: 15.2(2.0.0) Known Fixed Releases: 15.2(2)E2, 15.2(3)E1, 3.6(2)E, 3.7(1)E Bug Id: CSCud01798 Title: "FRU Power Supply is not responding" was seen unexpectedly Description: Symptom: "FRU Power Supply is not responding" message seen on 3560X/3750X intermittently. Issue may happen with one or two Power Supply Unit. Conditions: 3560X/3750X is running IOS 15.0(2)SE. Workaround: The power supply is operating normally. Upgrade to 15.0(2)SE2 or later to resolve this issue. Status: Fixed Severity: 2 Severe Last Modified: 30-JUN-2015 Known Affected Releases: 15.0(2.0.80)SE, 15.0(9.8)EMP Known Fixed Releases: 12.2(55)SE10, 15.0(2)EJ, 15.0(2)EJ1, 15.0(2)EZ, 15.0(2)SE2, 15.0(2)SE3, 15.0(2)SE4, 15.0(2)SE5, 15.0(2)SE6, 15.2(1)E Bug Id: CSCum75450 Title: CPUHB_RECV_STARVE when acl log configured in stack Description: Symptom: With following message, switch performance may be slow, sometimes switch gets hang and not recovered until power cycling. %SUPQ-4-CPUHB_RECV_STARVE: Still seeing receive queue stuck after throttling Following messages may also be observed when problem happens. %PLATFORM_RPC-3-MSG_THROTTLED: RPC Msg Dropped by throttle mechanism %XDR-6-XDRIPCNOTIFY: Message not sent to slot X because of IPC error timeout. Disabling linecard. (Expected during linecard OIR) Conditions: Catalyst 3750X in stack configuration running 12.2(58)SE or later including 15.0SE releases and 15.2E releases. log keyword is configured in access-list and traffic go through the member switches hit access-list that log keyword is configured. Workaround: Configuring longer logging interval may be the workaround for this problem. For example, ip access-list logging interval Risk for this workaround is packet count that hit the access-list might be inaccurate because logging process is involved with every logging interval. To recover the switch when problem is happening, you need to power cycling the switch. Further Problem Description: Note that this problem does not happen in 12.2(55)SE releases. Status: Fixed Severity: 2 Severe Last Modified: 30-JUN-2015 Known Affected Releases: 15.0(2)SE2, 15.2(1)E Known Fixed Releases: 15.0(2)SE7, 15.2(1.30)PSR, 15.2(2)E, 15.2(2b)E, 3.6(0)E Find additional information in Bug Search index.   2013 Cisco and/or its affiliates. All rights reserved. Terms & Conditions | Privacy Statement | Cookie Policy | Trademarks