Minimum system requirements

This section describes the minimum system requirements for this release.

• Intel® Pentium® III 550MHz processor
• 512MB disk drive or CompactFlash® card
• 256MB RAM
• Supported browsers: Microsoft® Internet Explorer 5.0 or 5.5; Netscape® Navigator 4.7x

---

Installing the PTF

Important:  If you are upgrading a Link Controller redundant system, both units must be upgraded. We do not support running different PTF versions on a Link Controller redundant system. Additionally, If you are updating the Link Controller module on a BIG-IP system, refer to the BIG-IP, version 4.5 PTF-04 note for instructions on installing the PTF.

Apply the PTF to the BIG-IP Link Controller, version 4.5 using the following process.  Note that the installation script saves your current configuration.

Note:  If you have installed prior PTFs, this installation does not overwrite any configuration changes that you made for prior PTFs.

Important:  If you are upgrading an IP Application Switch or a Link Controller unit that uses a CompactFlash® media drive, use the installation instructions here.

1. Change to the /var/tmp/ directory by typing:
   
   cd /var/tmp/
   
   
2. Connect to the F5 Networks FTP site (ftp.f5.com).
   
   
3. Use FTP in passive mode from the BIG-IP unit to download the file.  To place FTP in passive mode, type pass at the command line before transferring the file. 
   
   
4. Download the PTF file BIGIP_4.5PTF-04.im to the /var/tmp/ directory on the Link Controller.
   
   
5. To install this PTF, type the following command:
   
   im BIGIP_4.5PTF-04.im

   The Link Controller automatically reboots once it completes installation.

To upgrade an IP Application Switch or a CompactFlash media drive, use the following process.

1. Create a memory file system, by typing the following:
   
   mount_mfs -s 200000 /mnt
   
   
2. Type the following command:
   
   cd /mnt
   
   
3. Connect to the FTP site (ftp.f5.com).
   
   
4. Download the PTF file, BIGIP_4.5PTF-04.im, from the /crypto/bigip/ptfs/bigip45ptf4/ directory.
   
   
5. On the BIG-IP unit, run the im upgrade script:
   
   im /mnt/BIGIP_4.5PTF-04.im

   When the im script is finished, the Link Controller reboots automatically.

Note:  This procedure provides over 90MB of temporary space on /mnt.  The partition and the im package file are deleted upon rebooting.

---

Software enhancements and fixes

What’s new in this PTF

Enabling one-time automatic discovery in the Setup utility (CR24565)
The Setup utility now includes an option to enable automatic discovery of the local system's configuration, and its peer's configuration, if applicable, when you run the Setup utility for the first time. This option is most useful if you are running the Link Controller module on a BIG-IP system.

Synchronizing Link Controllers with 3-DNS Controllers (CR25753)
If your network includes both 3-DNS Controllers and Link Controllers, you can add the Link Controllers to the 3-DNS sync group, if you have one configured. For details on adding a Link Controller to a 3-DNS sync group, see the Adding a Link Controller to a 3-DNS sync group section of this PTF note.

New proxy ARP exclusion class (CR25801)
You can now create a proxy ARP exclusion class on the Link Controller, proxy_arp_exclude. Use this class to prevent the Link Controller from generating gratuitous ARP requests to its peer unit when you have a redundant system. To configure the proxy_arp_exclude class, in the navigation pane, click Classes, and then click the Add Class button. (For assistance with the settings, click the Help button.) You can also find information about the proxy_arp_exclude class in the BIG-IP Reference Guide , version 4.5.

If you use VLAN groups, you must configure a proxy ARP forwarding exclusion list. We recommend that you configure this feature if you use VLAN groups with a BIG-IP redundant systems. The reason is that both BIG-IP units need to communicate directly with their gateways and the back-end nodes. Creating a proxy ARP exclusion list prevents traffic from being proxied through the active BIG-IP due to proxy ARP. This traffic needs to be sent directly to the destination, not proxied.

If you do not configure a proxy ARP exclusion group for systems configured with VLAN groups, you may see problems such as:

• Nodes being marked down for a period of time after a failover.
• The inability to access resources through the active BIG-IP unit when there are multiple physical or logical connections to the same VLAN group (especially likely to be noticed when there are multiple connections between the active and standby BIG-IP units).

The OpenSSL package has been upgraded (CR26518)
The OpenSSL package has been upgraded to version 0.9.7a. This upgrade addresses several recent security issues with OpenSSL. For more information on the resolved security issues, see the CERT web site at http://www.cert.org.

Understanding the system_check script
The system_check script is useful for displaying and logging hardware failures. For more information about the system_check script, please refer to the BIG-IP New Features Guide for version 4.5 PTF-04.

Configuring SYN Check
The new SYN Check ^TM feature mitigates a particular type of denial-of-service attack known as a SYN flood. A SYN flood is an attack against a system for the purpose of exhausting that system resources. For more information about configuring the SYN Check feature, please refer to the BIG-IP New Features Guide for version 4.5 PTF-04.

Script to set up core capture
We have added a new script to automate core capturing on a BIG-IP system, if the system has a hard drive. The script runs automatically after you install this PTF and reboot the system. It provides functionality to enable and disable core capture.

After you install this PTF, the script runs, and creates the /var/crash directory. In addition, if the swap partition on the primary drive is not sufficiently large to capture the core file, but another unused partition is found to be, that partition is used for core capture.

You can disable this functionality with the following command:
config_savecore -disable

You can re-enable the functionality with the following command:
config_savecore -enable

Important: As long as this functionality is enabled, you see the message savecore: no core dump during boot time.

What’s fixed in this PTF

SNMP trap sink  (CR19769, CR24111)
The SNMP trap source is now an IP address that is routable to the trap sink.

Rebooting the system and lost interrupt error message  (CR19813)
In certain circumstances when you reboot, you no longer receive the error message wd0: lost interrupt.

Changing the CORBA port number using the Configuration Utility (CR19780)
You can no longer change the CORBA port number using the Configuration Utility. The CORBA IIOP port should be set only to the default setting of 683.

SNMP checktrap  (CR21701)
When the port for the node that is being marked up or down is any, checktrap now correctly identifies the port.

Windows uploads  (CR22043)
Delayed-acks no longer throttle Windows uploads at 40K per second.

Default wildcard ports  (CR22191)
Default wildcard ports now can use ICMP monitoring.

Network virtual servers  (CR22203)
Creating more than 1024 network virtual servers no longer causes the BIG-IP system to become unstable.

Short-lived rapid connections from the same source IP  (CR22232)
When dealing with short-lived rapid connections from the same source IP address, the BIG-IP system no longer arbitrarily resets some packets.

SNMP traffic and a VLAN that has port lockdown enabled (CR22677)
A VLAN configured with port lockdown enabled no longer passes SNMP traffic unless you have explicitly enabled the SNMP port using the open_snmp_port global setting.

Connection and packet display statistics with the bigtop utility (CR22709)
Connection and packet statistics now display correctly when you run the bigtop utility.

Upgrading from version 4.3 to version 4.5 and duplex billing status (CR23053, CR24773)
When you upgrade your Link Controller from version 4.3 to version 4.5, the upgrade process no longer enables duplex billing support, by default, for the links in your configuration.

Using a VLAN group configuration in transparent or translucent mode (CR24409)
You can now configure the BIG-IP system to bridge between two VLANs in either transparent or translucent mode without creating duplicate packets.

Upgrades and process checking in the snmpd.conf file (CR24450)
When you upgrade the software, the process checking entries (proc) in the snmpd.conf files are no longer populated with incorrect values.

Remote LDAP authentication and login errors (CR24487)
If you mistype the login name, and you are using remote LDAP authentication rather than RADIUS authentication, you no longer see a RADIUS error message.

Audit logs now show the correct user name (CR24600)
The audit logs now show the correct user name when a user makes configuration changes.

SNMP virtualAddressEntry table and wildcard virtual servers (CR24647)
The SNMP virtualAddressEntry table can now handle wildcard virtual servers.

Setting prepaid segments all to 0 (zero) and Over Prepaid setting on the Link Statistics screen (CR24680)
On the Link Statistics screen, in the Configuration utility, the Over Prepaid statistic now properly displays as No, under the following conditions:

• You do not set any bandwidth limits when you create the link
• You set the Prepaid Segment, on the Link Weighting screen, to 0 (zero)

Configuring the default gateway pool (CR24717, CR24740)
If you add only one default route when you run the Setup utility, the utility does not create a default gateway pool. If you then add a second link (and route) using the Link Configuration screen in the Configuration utility, the utility creates a default gateway pool and adds the second router to the pool. This process now automatically adds the first default route to the newly-created default gateway pool.

Name field on the Add VLAN Group and VLAN Group Properties pages (CR24719)
The maximum number of characters for a VLAN group name is 15 characters.

Monitor names in the Configuration utility and from the command line (CR24864)
Monitor names typed in the Configuration utility and the command line are no longer limited to 31 characters.

LDAP authentication and user names (CR24880)
If you use LDAP authentication, and you use the user name, user, the system no longer fails to update the configuration.

Audit logs and resetting statistics for services (CR24923)
The audit logs now correctly show the services when you reset statistics with the command b global stats reset.

Resetting statistics for node server (CR24924)
The audit logs now correctly show when you reset the statistics for a node server.

Gratuitous ARPs with MAC masquerading and VLAN failsafe configured (CR24925)
Gratuitous ARPs are now handled correctly in an active/standby redundant scenario with MAC masquerading and VLAN failsafe configured. When the active unit detects no traffic on the VLAN, such as when the cable is unplugged, or the unit is rebooted, the other unit becomes active. When the unit that was demoted to standby reboots, it now sends a gratuitous ARP for its self IP addresses.

SSH key generation now uses hardware random number generators when available (CR24955)
The BIG-IP system now uses hardware random number generators (when available) if you try to log in from an SSH client for which the BIG-IP system does not have a valid key. This increases the security of DSA host keys, and reduces the probability that the key can be guessed or that a random key collision could occur.

Reaper no longer sends RSTs for unaccepted, timed-out connection requests (CR24984)
We have corrected a problem that could be caused if a packet was sent from a client through a virtual server to a server, and the server did not answer before the connection timeout was reached, the reaper sent a reset packet (RST) in both directions.

TCP SYN packet to self IP that matches TIME_WAIT connection now handled correctly (CR24993)
If a TCP SYN packet is received for a self IP, and it matches an old connection that is in TIME_WAIT state (same source and destination address and port), the system no longer deletes the old connection and creates a new one.

Invalid OID for the shutdown trap in the SNMP MIB (CR25059)
The shutdown trap, in the SNMP MIB, now has the appropriate object identifier (OID) associated with it.

Connection table entry reaping for UDP packets with node address disabled (CR25186)
We have corrected a problem where in rare circumstances, connection table entries were not reaped for UDP packets when the node address was disabled.

MAC masquerade addresses and forcing a system to standby (CR25453)
When you purposefully change the state on a BIG-IP unit in a redundant system from active to standby, the first octet of the MAC address for any self IPs that you have configured no longer changes to 02. This happened only under certain conditions.

Turning off Total Traffic Limit after setting all limits (CR25466)
In the Configuration utility on the Link Configuration screen, you can now turn off the total traffic limit for a link once you have configured a limit for total traffic.

bigpipe interface show command returns data for interfaces (CR25470)
The bigpipe interface show command now returns data for interfaces that are passing traffic.

SNMP: enterprises.ucdavis.memory.* OID now returns valid information (CR25488)
The enterprises.ucdavis.memory.* now returns valid information.

Associating multiple monitors with the same service (CR25572)
You can now associate multiple monitors with the same service with the Configuration utility and not receive the message Error 132 - Monitor template not found.

Logging the forced down state for nodes (CR25614)
When you force a node to the DOWN state using the Configuration utility, or from the command line, the forced down state is now logged in the /var/log/bigd file.

Reboot of standby 2400 unit and connectivity with the active unit (CR26078)
We have corrected a problem where in certain cases, on the 2400 platform with network failover configured, rebooting the standby unit in an active/standby redundant configuration caused the active unit to lose existing connections. We recommend that if you require network failover, you configure the admin ports (port number 3.1) for failover.

Multiple VLAN SNATs when virtual servers are fully accelerated (CR26242)
On a platform with the Packet Velocity ASIC, when you have multiple VLAN SNATs configured, they are now partially accelerated when virtual servers are fully accelerated.

The b load command and connection limits (CR26451)
The b load command no longer causes connection limits to break.

The bigpipe command and values for ip_tos (CR26478)
The bigpipe command now limits the possible values for ip_tos to the correct value range (0 - 255).

Port translation default settings for the Configuration utility and command line (CR26543)
The following settings are the default port translation settings for both the Configuration utility and the command line:

Type of object	Port Translation
net:*	disabled
ip:*	disabled
vlan:*	disabled
*:*	disabled
ip:port	enabled
net:port	enabled
vlan:port	enabled
*:port	disabled

Losing connectivity during configuration of second unit in a redundant system (CR26705)
When you configure a unit from the command line Setup utility, we recommend that you reboot the unit after you complete the configuration. This activates the license and allows traffic to pass through the system. Also, before you reboot the system it is in the active mode and unlicensed. While the unit is in the active mode, the other unit in the redundant system is placed in standby mode. If left in this state, traffic can not pass through the system.

Log rotation for the ITCM.log file (CR26781)
The frequency of the log rotation for the ITCM.log file has been increased from once every 7 days to once every 24 hours. This improves the system efficiency if you are monitoring the controller with the iControl Services Manager.

Admin password for authentication and updating the configuration (CR26824)
The adminpw setting is now saved correctly when you load a configuration using the b config load command.

bge message on reboot (CR26827)
You no longer see the following message when you reboot the 1000 and 5100 series platforms:

bge0: bge_wait_bit_clr timeout: reg=0x468 mask=0x2

Network virtual server loading in a particular order with others on the same subnet (CR26988)
We have corrected a problem that was preventing network virtual servers on the same subnet from working if they were not ordered in the /conf/bigip.conf file in a particular order.

Transaction level on systems monitored by the iControl ^TMServices Manager (CR27192)
We have reduced the quantity of transactions generated on systems monitored by the iControl Services Manager.

Configuration Utility: display warning if product is licensed however the EULA has not been accepted (CR27215)
A warning is now displayed if the system is licensed but you have not accepted the EULA.

Enhancements and fixes released in prior PTFs

Version 4.5 PTF-03

There are no enhancements or fixes for Link Controller in version 4.5 PTF-03.

Version 4.5 PTF-02

Enhancements to inbound load balancing
This PTF adds a new load balancing method, fallback, and two new load balancing modes for the fallback method, drop_packet and explicit_ip. The fallback method and load balancing modes are applicable to inbound load balancing only. The Link Controller uses the fallback method when the preferred and alternate load balancing modes do not provide an available virtual server to return as an answer to a query. When you specify the drop_packet mode, the Link Controller does nothing with the packet, and simply drops the request. (Note that a typical LDNS server iteratively queries other authoritative name servers when it times out on a query.) When you specify the explicit_ip mode, the 3-DNS Controller returns the IP address that you specify as the fallback IP as an answer to the query. Note that the IP address that you specify is not monitored for availability before being returned as an answer. When you use the explicit_ip mode, you can specify a disaster recovery site to return when no load balancing mode returns an available virtual server.

You can configure the fallback method only from the command line. For information on configuring the fallback method and load balancing mode, see the Configuring the fallback method for inbound load balancing section of this PTF note.

UDP checksums and TFTP packets  (CR22113, CR25181)
In rare instances, the checksums for TFTP packets were incorrect. This issue has been resolved.

Resets (RSTs) with incorrect sequence numbers   (CR22219)
Resets (RSTs) from aging-out connections no longer cause some connections to hang due to incorrect sequence numbers for the resets.

Apache web server and the CERT Coordination Center vulnerability, VU#672683 (CR24689)
This PTF addresses the vulnerability in the Tomcat package for the Apache web server that is described in Vulnerability Note VU#672683 on the CERT® Coordination Center website. For more information on the vulnerability, see http://www.kb.cert.org/vuls/id/672683.

iControl BaseServer::get_interfaces function and the 3dnsd process (CR24912)
The following iControl function, ITCMGlobalLB::BaseServer::get_interfaces, no longer causes the 3dnsd process to stop running when you specify an invalid type within the function.

Root servers list for BIND (CR25064)
The root servers list file for BIND, root.hint, has been updated to include the most current list of root servers.

Invalid metrics statistics and graphs for down remote links (CR25146)
The Link Statistics screen, in the Configuration utility, no longer displays very large, invalid values for remote links that are down (red ball). The link statistics graphs now accurately display the data for both the link that is down, and any available links.

Using a serial terminal as a console (CR25183)
This PTF fixes the serial terminal as the console functionality, as described in the 3-DNS Reference Guide, Chapter 6, Monitoring and Administration, so that it works with all 2U controller platforms.

Version 4.5 PTF-01

CA-2002-31, Multiple Vulnerabilities in BIND
This PTF addresses the security vulnerabilities that are listed in CERT® advisory, CA-2002-31, Multiple Vulnerabilities in BIND. This PTF upgrades the BIND package to version 8.3.4. For more information on the CERT advisory, see http://www.cert.org/advisories/CA-2002-31.html.

Support for the 2400 platform
This release includes enhanced support for the F5 Networks 2400 platform.

Viewing licensing error log files from the Configuration utility (CR25055)
You can now view the log files for errors that occur during the licensing process using the Configuration utility. A View Log File button appears on the licensing screen when the licensing process generates errors.

---

Optional configuration changes

The following section provides information about optional configuration changes.

Adding a Link Controller to a 3-DNS sync group

If you have both 3-DNS Controllers and one or more Link Controllers in your network, you can add the Link Controllers to the 3-DNS Controllers' sync group, in a few simple steps. There are three tasks to adding a Link Controller to a 3-DNS sync group:

• Run the merge_configs script on the sync group's principal controller.
  
  
• Add the Link Controller to the sync group using the principal controller's Configuration utility.
  
  
• Run the 3dns_add script on the Link Controller.

The following sections explain the specific steps for each of the previous tasks. You must perform these tasks in the order they are listed.

Note: Before you add the Link Controller to the 3-DNS sync group, we recommend that you back up both the 3-DNS configuration and the Link Controller configuration.

To run the merge_configs script

From the command line on the principal 3-DNS Controller, run the merge_configs script by typing the following command, where <ip_address>is the IP address of the Link Controller that you want to add to the sync group.

/usr/local/bin/merge_configs -peer <ip_address>

To make the sync group aware of the Link Controller

Using the Configuration utility on the principal 3-DNS Controller, add the Link Controller to the sync group.

1. In the navigation pane, click 3-DNS Sync.
   
   The Synchronization screen opens.
   
   
2. On the toolbar, click Add to Group.
   
   The Add a 3-DNS to a Sync Group screen opens.
   
   
3. Check the box next to the controller that you want to add to the sync group, and click Add.

To add the Link Controller to the sync group and start synchronization

The final step in adding the Link Controller to a 3-DNS sync group is to run the 3dns_add script on the Link Controller. The script moves the synchronized configuration to the Link Controller, and finalizes the sync group setup.

• From the command line of the Link Controller, run the 3dns_add script.
  
  3dns_add
  
  The script runs, and finalizes the setup of the sync group.

---

Configuring the fallback method for inbound load balancing

You can configure the fallback method using the new load balancing modes either by using the Configuration utility, or by editing the wideip.conf file from the command line. You can specify either the Drop Packet load balancing mode, or the Explicit IP load balancing mode. Note that if you specify the Explicit IP mode, you also specify a fallback IP address.

To configure the fallback method with the Drop Packet mode using the Configuration utility

1. In the navigation pane, expand the Link Configuration item, and then click Inbound LB.
   
   The Wide IP list screen opens.
   
   
2. In the Wide IP column, click the name of the wide IP that you want to modify.
   
   The Wide IP Properties screen opens.
   
   
3. Click the Wide IP Load Balancing tab.
   
   The Load Balancing Options screen opens.
   
   
4. In the Fallback box, select Drop Packet.
   
   
5. Click Apply.
   
   The Configuration utility updates the configuration with the changes.

To configure the fallback method using the drop_packet mode from the command line

1. To ensure that the configuration files contain the same information as the memory cache, type the following command:
   
   3ndc dumpdb
   
   
2. Open the /etc/wideip.conf file in a text editor (either vi or pico).
   
   
3. Use the syntax highlighted in the sample below to configure the fallback method with the drop_packet mode.
   
   
4. Save and close the file.
   
   
5. Commit the changes to the configuration by typing:
   
   3ndc reload
   
   

Syntax example for the Drop Packet (drop_packet) load balancing mode

wideip { ...    pool {       name     "Pool"       dynamic_ratio     yes       preferred     qos       alternate     rr       fallback     drop_packet       address     <vs_ip_address>       address     <vs_ip_address>

To configure the fallback method with the Explicit IP mode using the Configuration utility

1. In the navigation pane, expand the Link Configuration item, and then click Inbound LB.
   
   The Wide IP list screen opens.
   
   
2. In the Wide IP column, click the name of the wide IP that you want to modify.
   
   The Wide IP Properties screen opens.
   
   
3. Click the Wide IP Load Balancing tab.
   
   The Load Balancing Options screen opens.
   
   
4. In the Fallback box, select Explicit IP.
   
   
5. In the Fallback IP box, type the IP address for the server or host to which you want the Link Controller to forward the packet.
   
   
6. Click Apply.
   
   The Configuration utility updates the configuration with the changes.

To configure the fallback method with the explicit_ip mode from the command line

1. To ensure that the configuration files contain the same information as the memory cache, type the following command:
   
   3ndc dumpdb
   
   
2. Open the /etc/wideip.conf file in a text editor (either vi or pico).
   
   
3. Use the syntax highlighted in the sample below to configure the fallback method with the explicit_ip mode.
   
   
4. Save and close the file.
   
   
5. Commit the changes to the configuration by typing:
   
   3ndc reload

Syntax example for the Explicit IP (explicit_ip) load balancing mode

wideip { ...    pool {       name     "Pool"       dynamic_ratio     yes       preferred     qos       alternate     rr       fallback     explicit_ip       fallback_ip     <ip_address>       address     <vs_ip_address>       address     <vs_ip_address>

Known issues

The following items are the known issues identified since the release of BIG-IP Link Controller, version 4.5.

Receiver Link Controllers and monitoring single routers (CR27756)
Receiver Link Controllers in a sync group incorrectly display the state of the router that is in their own data center as unknown (blue ball). When the controller also monitors virtual servers in the same data center, the virtual servers inherit the state of the router (unknown), and are unavailable for load balancing. Note that this occurs when the data center contains only one router.

Global Availability or Ratio load balancing within a wide IP pool (CR13112)
When you create a pool for a new or for an existing wide IP, and you use the Global Availability or Ratio load balancing method, you may experience problems under the following circumstances:

• You are using Internet Explorer 5.0 or 5.5.
• You select Global Availability or Ratio in the Load Balancing Modes, Preferred list or in the Load Balancing Modes, Alternate list on the Configure Load Balancing for New Pool screen.
• You have a large quantity of virtual servers in your configuration.

If you want to use the Global Availability or Ratio load balancing method, and you meet the previous criteria, please see the Using the Global Availability or Ratio load balancing mode within a wide IP pool work-around following this section.

Setting active-active mode using the web-based Configuration utility  (CR19794)
With network failover enabled, you cannot use the Configuration utility to configure active-active mode. When you have network failover enabled, use the command line interface to set active-active mode.

Values for Link Limits  (CR20744)
On the Modify Link screen in the Configuration utility, when you type values for bandwidth limits, and you type a number that is not divisible by 8, the Configuration utility rounds the value to the next lowest number that is divisible by 8.

Manually deleting connections handled by the Packet Velocity ASIC (CR22494)
Manually deleting connections that are handled by the Packet Velocity ^TM ASIC does not generate a TCP reset.

Using the MGMT interface on units that include the Packet Velocity ASIC (CR22599)
It is important that you use the MGMT interface (3.1), on units that include the Packet Velocity ASIC, for administration only. We recommend that you do not use the MGMT interface on a VLAN you plan to use for load balancing traffic.

Changing active-active failback values (CR22715)
In active-active configurations, we recommend that you do not change the default failback value of 60 seconds. If you change this value, failback may not work as designed.

Default route missing after gated fails while you are creating node pools (CR23668)
In certain rare cases, the default route may be removed if you create a node pool at the same time gated fails. If this happens, run the Setup utility and add the default route back to the configuration. You can run the Setup utility from the command line by typing setup. You can access the Web-based Setup utility from the welcome page of the Web-based Configuration utility.

Layer 2 (L2) forwarding two VLANs on one interface  (CR23460)
When a VLAN group is bridging across the internal and external VLANs with the same IP network on both sides of the BIG-IP system, and you configure only one interface, with VLAN tags for both internal and external VLANs, the network becomes unusable. In this type of configuration, you need to configure one interface for each VLAN in the VLAN group in order for the BIG-IP system to function correctly.

Inaccurate titles for Billing Estimate graphs (CR23770)
When you change the date or time range on the Billing Estimate screen in the Link Statistics, the titles on the graphs do not update to reflect the changes. If you are using Internet Explorer, you can update the titles by holding down the Control key, right-clicking in the screen, and then clicking Refresh. If you are using Netscape Navigator, you can update the titles by holding down the Shift key, right-clicking in the screen, and then clicking Refresh.

Changing IP addresses on VLANs does not change the administration web server settings (CR24468)
If you use the Setup utility to change the floating IP addresses on VLANs, the web server settings are not updated. To update the web server settings, choose the (W) Configure web server option.

Platforms using Broadcom 570x controllers (CR24388, CR25464)
On rare occasions, some platforms using Broadcom 570x controllers may experience short interruptions in network connectivity.

Deleting the Default Gateway Pool using the Setup utility   (CR24519)
If you define a default gateway pool using the Setup utility, and then define a virtual server or other network objects on the pool, you will not be able to delete the pool using the Setup utility as long as the pool is in use. In order to delete the pool using the Setup utility, you must first remove all IP addresses and network objects associated with the pool.

FTP data connections have incorrect TOS or QoS values set (CR24644)
FTP data connections have incorrect TOS or QoS values set. Both values are set to 0.

Viewing wide IPs created in the 3-DNS Controller module from the Link Controller module (CR24842)
Wide IPs that you create in the 3-DNS Controller module that contain more than one pool, display only the first pool of the wide IP in the Inbound LB screen in the Link Controller module. You may encounter this known issue only when you are running a BIG-IP system with both the 3-DNS Controller module and the Link Controller module.

iControl SOAPPortal: .NET serialization errors on several methods (CR24862)
The following methods do not serialize correctly under certain situations. This is due to a problem in the .NET frameworks serialization. For nested structures within arrays, the framework cannot support an empty array represented as a single XML element.

For example, this method does not serialize:
<return type='Array' ArrayType='tns:someType[0]/>

This method does serialize:
<return type='Array' ArrayType='tns:someType[0]></return>

The BIG-IP Link Controller Solutions Guide is not available from the home screen in the Configuration utility (CR24946)
The BIG-IP Link Controller Solutions Guide is not available from the web server Welcome screen. You can obtain this guide from the Software and Documentation CD by navigating to the /doc directory, and opening the lc_solutions.pdf file. You can also obtain the guide from the AskF5 web site (http://tech.f5.com).

SNAT automap and acceleration (CR24959)
On the 2400 platform, if you configure SNAT automap and do not associate the SNAT with a virtual server, the traffic is not accelerated by the Packet Velocity ^TM ASIC. Note that you can associate the SNAT with a wildcard virtual server to accelerate any SNAT automap traffic.

Changing the hardware acceleration mode and resetting connections (CR25009)
When you change the hardware acceleration mode for a pool, and there are current connections for the nodes in the pool, the connections do not reset when you use the b conn reset command. The connections do close when they reach their time-to-live (TTL).

b conn dump verbose command does not display correct values for packet counts or byte counts (CR25119)
The bigpipe command, b conn dump verbose, displays incorrect values for packet counts and byte counts.

Microsoft® Internet Explorer security settings and the Link Configuration screens (CR25444)
If you are using a browser session in Internet Explorer to view the Configuration utility, and you have changed the security level for the browser to a setting higher than Medium (the default), then the Link Configuration screens do not work properly. The errors in the Link Configuration screens occur because the Link Controller's web server uses cookies. To avoid this error, set the security level for the browser session to Medium or lower.

SSH access host restrictions are now configured in /etc/hosts.allow (CR25530)
In previous versions, the /etc/ssh2/sshd2_config and /etc/sshd_config files controlled SSH access. Upgrading to this version ignores previously configured SSH access restrictions configured in /etc/ssh2/sshd2_config and /etc/sshd_config. This upgrade reverts to an SSH access level that allows all hosts to connect. If you require restricted SSH access to certain networks/IP addresses, you need to reconfigure these restrictions once the upgrade has been completed. To do this, type the following command to start the Setup utility, and then press Enter:

setup

Choose option S (Configure SSH), and set the restrictions you prefer.

Adding support access after initial setup does not add the IP addresses to the /etc/hosts.allow file (CR25821)
If you add support access with the (Y) Set support access option in the Setup utility after you complete the initial setup of the system, the support IP addresses are not added to the hosts.allow file. To correct this situation, run the (S) Configure SSH option in the Setup utility to re-initialize the SSH information on the system.

VLAN names with "vlan" followed by any number of digits cause a syntax error (CR25890)
VLAN names that start with the text vlan, and are followed by any number of digits (for example, vlan123), cause a syntax error. We recommend that you do not use the text, vlan, as the initial portion of a VLAN name.

Creating invalid interface names (CR25950)
It is possible to create invalid interface names in your configuration by entering an invalid VLAN name from the command line. For more information about invalid VLAN names, see (CR25890).

Using 127.0.0.x as a pool member causes the system to lose network connectivity (CR26184)
If you add a node with an IP address of 127.0.0.x to a pool, the system loses connectivity to the network. The only way to reboot the system after this happens is to use the reboot switch. We recommend that you do not add nodes with this address range to a pool.

Changing iControl settings does not restart the CORBA portal (CR26384)
If you use the Configuration utility to change iControl settings, you must manually restart the CORBA portal. To restart the CORBA portal, type the following commands from the command line:

bigstart shutdown portal
bigstart startup

LDAP group name naming conventions (CR26418)
LDAP authentication for groups does not work properly when there are spaces in the group name. To avoid authentication issues with groups when you use LDAP authentication, do no use spaces in the group names.

Error message is invalid for ip_tos values (CR26566)
The valid ip_tos values are 0 - 255 or 65536, which returns ip_tos to a blank state. If you type an invalid value, you see the following incorrect error message: The requested IP TOS value is invalid. [0..65535].

Disabling the SNMP Auth Trap Enable setting using the Configuration utility (CR26610)
If you try to disable the Auth Trap Enable setting on the SNMP Administration screen in the Configuration utility, the SNMP configuration file, /etc/snmpd.conf , is modified with an incorrect setting of 0 (zero), and the following error is generated in the SNMP log:

"/etc/snmpd.conf: line ##: Error: authtrapenable must be 1 or 2

To correct this error and disable the Auth Trap Enable setting, you can edit the /etc/snmpd.conf file, and change the authtrapenable value to 2, disable.

Losing connectivity during configuration of second unit in a redundant system (CR26705)
When you configure a unit from the command line Setup utility, we recommend that you reboot the unit after you complete the configuration. This activates the license and allows traffic to pass through the system. Also, before you reboot the system, it is in the active mode and unlicensed. While the unit is in the active mode, the other unit in the redundant system is placed in standby mode. If left in this state, traffic cannot pass through the system.

Error messages in the Link Configuration screens in the Configuration utility (CR26851, CR27177)
The Link Controller may generate a JSP 500 error message when you are changing settings on the Link Configuration screens. If you see the JSP 500 error message, one or more utilities on the system may have stopped running. To correct the error, use the work-around described in the Correcting the JSP errors section of this PTF note.

Transparent monitors cannot be combined using an "and" rule (CR26915)
You can not combine transparent monitors using an and rule.

The Setup utility does not preserve MAC masquerade settings (CR26922)
The Setup utility, setup, does not preserve MAC masquerade settings. We recommend that you use the bigpipe utility or the web-based Configuration utility to make configuration changes after you have completed your initial setup. However, if you want to use the Setup utility to make changes to the configuration, and you want to preserve the MAC masquerade settings, then after you finish your configuration changes, recreate your MAC masquerade settings with bigpipe or the Configuration utility before you reboot the unit.

Regkey.license should not be synchronized or saved in the .ucs file (CR27020)
In a redundant system, the regkey.license file is synchronized, and it should not be, both when you perform a configuration synchronization and when you save a backup .ucs file. You can avoid this problem by adding the file to the list of files that are ignored when you generate a .ucs file, and when you synchronize the configuration in the bigdb. For example, type the following command to set this value:

b db set Common.Bigip.CS.save.120.ignore = "regkey.license"

Changing the system IP address and updating the IP address for the CORBA portal in bigdb (CR27037)
If you change the IP address of the system using the Configuration utility, the system does not update the IP address for IIOP and FSSL for the CORBA portal in the bigdb. To change the CORBA address for IIOP and FSSL, run the Setup utility (setup) from the command line, and choose the option (I) Initialize iControl portal.

Adding a switch interface to the admin vlan causes large volumes of traffic (CR27103)
Adding a switch interface to the admin VLAN causes large volumes of traffic. We recommend that you do not add a switch interface to the admin VLAN.

Certain load balancing modes do not honor node connection limits  (CR27124)
When using the observed_member, predictive_member, predictive, or observed load balancing modes, the member and node addresses do not honor node connection limits.

CompactFlash® media drives and logging for named  (CR27132)
When named is running, it generates status and usage messages as part of its normal behavior. If you are running named on a system with a CompactFlash media drive, these messages may fill up the /var/log/messages directory. To avoid this, periodically delete the status and usage messages for named.

RADIUS server configuration and Netscape  (CR27212)
If you configure remote login for RADIUS, and you set an invalid IP address for the primary RADIUS server, and a valid IP address for the secondary RADIUS server, you may not be able to log in using a Netscape browser. This can also happen if your primary RADIUS server is down. We recommend that you use an alternative browser with this type of configuration.

User administration for remote authentication using the Configuration utility  (CR27223)
With remote authentication configured, if you use the Configuration utility to add a new user, you may receive an internal server error message when you press Enter, and then click the Done button. The user is added when you press Enter. When using local authorization, the Enter key is ignored, and you must click the Done button in order to add a new user.

Incorrect UDP checksum generated if incoming request has 0 UDP checksum   (CR27240)
If an incoming UDP request has an initial checksum of 0, when the request is routed back through the BIG-IP system, the UDP checksum may be calculated incorrectly.

Deleting the default gateway pool using the Setup utility (CR27260)
The command line Setup utility, (setup), does not delete the default gateway pool when you remove all of the pool's members. To work around this issue, delete the default gateway pool using the browser-based Configuration utility.

Unsupported system_check tool  (CR27354)
The system_check script is running on all BIG-IP platforms. The system_check script is supported only on IP Application Switch platforms. This script does not have any adverse effect on unsupported platforms.

Changes in US and Canada Daylight Saving Time (CR58321)
The Energy Policy Act of 2005, which was passed by the US Congress in August 2005, changed both the start and end dates for Daylight Saving Time in the United States, effective March 2007. Canada is also adopting this change. The resulting changes are not reflected in this version of the product software. To find out more about this issue, refer to SOL6551: F5 Networks software compliance with the Energy Policy Act of 2005.

---

Work-arounds for known issues

The following sections describe work-arounds for some of the known issues listed in the previous section.

---

Correcting the JSP errors

If you see a JSP 500 error message when you are changing settings on the Link Configuration screens, one or more utilities may have stopped running. To restart the utilities, use the following process.

To restart utilities that may have stopped running due to a JSP 500 error

From the command line, type the following commands, in the order they are presented:

bigstart restart portal
bigstart restart corbalocallb
bigstart restart corbasystem
bigstart restart corbanetwork
bigstart restart corbamgmt
3ndc restart
bigstart restart tomcat

Using the Global Availability or Ratio load balancing mode within a wide IP pool

The following instructions describe how to configure the Global Availability or Ratio load balancing mode within a wide IP pool. You need to use these instructions only if you meet the criteria listed in the Using the Global Availability or Ratio load balancing mode within a pool item in the Known Issues section.

To configure Global Availability or Ratio load balancing within a pool in a new wide IP

1. In the navigation pane, click Wide IPs.
   
   The Wide IP List screen opens.
   
   
2. On the toolbar, click Add Wide IP.
   
   The Add a New Wide IP screen opens.
   
   
3. Type the settings for the new wide IP, and click Next.
   
   The Configure Load Balancing for New Pool screen opens.
   
   
4. Select a load balancing mode other than Global Availability in all of the following lists:
   
   
   • Load Balancing Modes, Preferred
   • Load Balancing Modes, Alternate
   • Load Balancing Modes, Fallback
   Note that you can accept the default settings, rather than changing the settings.
   
   
5. Click Next.
   
   The Select Virtual Servers screen opens.
   
   
6. Once you have finished configuring the virtual servers for the pool, click Finish to save your changes.
   
   
7. On the Wide IP List screen, select the wide IP that you just created.
   
   
8. On the toolbar, click Modify Pool.
   
   The Modify Wide IP Pools screen opens.
   
   
9. Click the pool that you just created.
   The Modify Load Balancing for [pool name] screen opens.
   
   
10. Select Global Availability, as appropriate, in the Load Balancing Modes, Preferred, or the Load Balancing Modes, Alternate, or the Load Balancing Modes, Fallback list, and click Update.
    
    The Modify Virtual Servers screen opens, where you can determine the order in which the 3-DNS Controller load balances to the virtual servers in the pool.

To configure Global Availability or Ratio load balancing within a pool in an existing wide IP

1. In the navigation pane, click Wide IPs.
   
   The Wide IP List screen opens.
   
   
2. On the toolbar, click Add Pool.
   
   The Configure Load Balancing for New Pool screen opens.
   
   
3. Select a load balancing mode other than Global Availability in the following lists:
   
   
   • Load Balancing Modes, Preferred
   • Load Balancing Modes, Alternate
   • Load Balancing Modes, Fallback
   Note that you can accept the default settings; you do not have to change the settings.
   
   
4. Once you have finished configuring the pool, click Finish to save your changes.
   
   The Wide IP List screen opens.
   
   
5. In the Pools column, select the pools for the wide IP that you just modified.
   
   The Modify Wide IP Pools screen opens.
   
   
6. In the Pool Name column, click the name of the pool that you just created.
   
   The Modify Load Balancing for [pool name] screen opens.
   
   
7. Select Global Availability, as appropriate, in the Load Balancing Modes, Preferred list, or the Load Balancing Modes, Alternate list, or the Load Balancing Modes, Fallback list, and click Update.
   
   The Modify Virtual Servers screen opens, where you can determine the order in which the 3-DNS Controller load balances to the virtual servers in the pool.

Copyright © 2003, F5 Networks, Inc., Seattle, Washington. All rights reserved.
F5, F5 Networks, the F5 logo, BIG-IP, 3-DNS, iControl, GLOBAL-SITE, SEE-IT, EDGE-FX, FireGuard, Internet Control Architecture, IP Application Switch, iRules, OneConnect, Packet Velocity, SYN Check, Control Your World, ZoneRunner, uRoam, FirePass, and TrafficShield are registered trademarks or trademarks of F5 Networks, Inc., in the U.S. and certain other countries. All other trademarks mentioned in this document are the property of their respective owners. F5 Networks' trademarks may not be used in connection with any product or service except as permitted in writing by F5.

This product protected by U.S. Patents 6,374,300; 6,473,802. Other patents pending.