* * * READ ME * * *
               * * * Veritas Operations Manager 5.0 * * *
                          * * * Hot Fix * * *
                         Patch Date: 2013-07-30


This document provides the following information:

   * PATCH NAME
   * OPERATING SYSTEMS SUPPORTED BY THE PATCH
   * PACKAGES AFFECTED BY THE PATCH
   * BASE PRODUCT VERSIONS FOR THE PATCH
   * SUMMARY OF INCIDENTS FIXED BY THE PATCH
   * DETAILS OF INCIDENTS FIXED BY THE PATCH
   * INSTALLATION PRE-REQUISITES
   * INSTALLING THE PATCH
   * REMOVING THE PATCH


PATCH NAME
----------
Veritas Operations Manager 5.0 Hot Fix


OPERATING SYSTEMS SUPPORTED BY THE PATCH
----------------------------------------
RHEL4 i686
RHEL4 x86-64
RHEL5 i686
RHEL5 x86-64
RHEL6 x86-64
SLES9 i586
SLES9 x86-64
SLES10 i586
SLES10 x86-64
SLES11 x86-64
HP-UX 11i v2 (11.23)
HP-UX 11i v3 (11.31)
Solaris 8 SPARC
Solaris 9 SPARC
Solaris 10 SPARC
Solaris 10 X86
Solaris 11 SPARC
Solaris 11 X86
AIX 5.2 ppc
AIX 5.3 ppc
AIX 6.1 ppc
AIX 7.1 ppc
Windows 2003 32-bit
Windows 2003 IA64
Windows 2003 X64
Windows 2008 32-bit
Windows 2008 IA64
Windows 2008 X64
Windows Server 2008 R2 X64


BASE PRODUCT VERSIONS FOR THE PATCH
-----------------------------------
   * Veritas Operations Manager 5.0


INSTALLING THE PATCH
--------------------
IMPORTANT NOTE : Please take a backup of the database using the instructions given in the Admin guide before installing this Hotfix. 

This Hotfix is applicable for VOM 5.0 Managed Hosts as well as VOM 5.0 Management Server.

1. Download the file HF050001960-35.sfa
2. Launch a browser and login to the VOM management server.
3. Navigate to Settings ->         Deployment Management tab.
4. Upload the add-on to the VOM CMS using the "upload" button.
   The HF HF050001960-35 should be visible in the table below.
5. HF install instructions
    - In the Deployment Management page,  click on the HF HF050001960-35 and it will take you to the Deployment Summary page. Select the 5.0 Managed hosts on which you want to apply the HF and click on the Install button.


REMOVING THE PATCH
------------------
Un-installation and rollback of this Hotfix is not supported. 


SPECIAL INSTRUCTIONS
--------------------
1. It requires approximately around 1.5 GB of disk space to upload HF050001960-35. Please ensure that Management Server has atleast 1.5 GB of free disk space where CMS is installed. 
2. This Hotfix deprecates the older version of Storage Insight and Virtual Business Services Availability add-ons. If you have these add-ons installed on your Management Server, please install the newer versions of Storage Insight (VRTSsfmsi-5.0.196.01.sfa) and Virtual Business Services Availability (VRTSBEopr-5.0.196.01.sfa) add-ons after installing this Hotfix.
3. If you are using Scripting add-on, please install the newer version of Scripting add-on (VRTSScripting-5.0.196.01.sfa) which is applicable on this Hotfix.
4. This Cumulative Hotfix includes the ASM feature that was part of the HSCL pack add-on.
5. If you see the following Warning in the output console, please ignore it as it is expected and it will not cause any functional impact.
"Warning: Found database metadata differences: Orig_Objs_Count(xx), Changed_Objs_Count(yy)".


SUMMARY OF INCIDENTS FIXED BY THE PATCH
---------------------------------------
Patch ID: HF050001960-35
* 2822611 (2829120) The state of a disk is incorrectly reported when the disk is shared between two different clusters/hosts.
* 2887094 (2894130) Unable to login after fresh install of VOM 5.0.
* 2906689 (2975047) Unable to add host to VOM CS from UI.
* 2908591 (2922721) MH shown as At Risk in GUI is without support Fault/At Risk condition.
* 2926661 (2932853) DS8800 Storage Insight deep dive does not show array details.
* 2930831 (2923139) Vomgather collects all data in var dir including DB and stores directory which increases the size of the archive.
* 2939361 (2943751) UUID instead of hostname observed.
* 2941489 (2975368) Cross Mirror check is not returning violation for Mirror which have subdisk in plex on different enclosure.
* 2968906 (2981057) luxadm should not be invoked by VOM to check tape devices.
* 2969824 (2981329) Modify vombackup to archive single copy of database.
* 2976267 (3027169) Vulnerability on port 5634/tcp over SSL.
* 2985578 (2986954) The vbs script does not handle passwords with special characters correctly.
* 3000090 (3004297) sfmsecd core dumps after upgrading from VOM 4.1 to VOM 5.0 due to bad hash for AdminUserPassword.
* 3001140 (3013488) xprtld is reported offline where process is running.
* 3011255 (3014934) VOM does not display information about VxVM and Filesystems if VXDCLID is reported down.
* 3011375 (3011197) If tablespace is created with same name in multiple databases the capacity is double counted for tablespaces.
* 3013115 (3069347) Vulnerability on port 5634/tcp over SSL on AIX MH hosts.
* 3013212 (3020472) vxdclid has been crashing with SIGSEGV fault in process_iostats_for_dmpstats() at 30 minutes interval on several MHs when CS polls the subject MHs.
* 3030371 (3030200) vxdclid spins when Stats set to false in dcli_conf.ini.
* 3046832 (3067192) vomgather is failing with tar error while archiving the required files and directories.
* 3048841 (3048888) Agentless configuration fails to connect to remote host using WMI commands.
* 3073174 (3079981) Unable to collect logs via vomgather due to out of memory errors.
* 3080750 (3099364) Discrepancy between VOM and VEA.
* 3082536 (3083463) hostgid.pl -set is calling netstat -ai, this causes several minutes or so in timeouts due to the call to netstat -ai, use netstat -ain instead to eliminate name resolution timeouts.
* 3083291 (3099464) Volume Migration does not work when multiple volumes are selected for migration.
* 3087219 (3099468) VOM is not reporting all of the volumes that exist in VEA.
* 3089901 (3089904) VOM 5.0 ( vxdisk -o thin, fssize list) triggers panic/memory corruption.
* 3093373 (3093370) Plug a security loop hole in xprtld.
* 3093852 (3099435) HA FireDRILL does not complete (stays in progress).
* 3098688 (3099409) Unable to delete FireDrill Schedule with localized characters.
* 3103093 (3113733) After patching from 5.1SP1RP1 to 5.1SP1RP3, the system is now logging the vxkeyless nagware message to the syslog.
* 3111579 (3149346) gendeploy script fails when configuring some MHs to secondary VOM CMS.
* 3117726 (3101454) Arbitrary File Read as ROOT.
* 3142907 (3032850) [xprtld] consumes enormous cpu when polling non-existent CMS.
* 3154957 (3162553) Replications view does not show secondary for RVG after successful migrate/ takeover VVR operations.
* 3184832 (3213855) Policy Template Signature(On Disk Config Size) reports violations incorrectly.
* 3185995 (3205035) Hotfix HF050001960-12 causes loss of virtualization (zone) details on VOM console.
* 3186189 (3186349) Javascript errors using Performance Statistics Tab on VOM Server.
* 3193430 (3205179) VOM GUI is not showing proper 'Free Space' for volumes.
* 3207832 (3210965) Plain text vCenter password shows up in ps command for vmwareCollector binary.
* 3225782 (3225767) Need to make SSL in vxdclid thread safe.
* 3234866 (3235078) VOM service is sending the excessive invalid DNS queries to DNS and need help in eliminating them.
* 3236091 (3237310) Updating latest JRE.
* 3263284 (3082055) Arbitrary File Download via XML External Entity Injection.


DETAILS OF INCIDENTS FIXED BY THE PATCH
---------------------------------------
This patch fixes the following Symantec incidents:

Patch ID: HF050001960-35

* 2822611 (Tracking ID: 2829120)

SYMPTOM:
VOM is not correctly showing the state of the disk.

DESCRIPTION:
The state of a disk is incorrectly reported when the disk is shared between two different clusters/hosts.

RESOLUTION:
VOM uses a new attribute provided by VxVM (called - Seq No) to correctly report the disks state and attributes that are shared  between two different clusters/hosts.

* 2887094 (Tracking ID: 2894130)

SYMPTOM:
Login to VOM 5.0 Management Server UI does not work.

DESCRIPTION:
Login to VOM 5.0 Management Server UI does not work when md5 encryption is used for password.

RESOLUTION:
Apply this HF to apply the updated libraries used within VOM to allow the VOM UI login.

* 2906689 (Tracking ID: 2975047)

SYMPTOM:
If a managed host is added from VOM UI, it's unable to add the host.

DESCRIPTION:
The Add host from UI does not work and the operation times out. 
However using gendeploy script, the managed host gets added to CMS.

RESOLUTION:
The issue was with parsing of the JSON output if it contained a "JOB" substring in the hostname. The parsing logic has now been fixed.

* 2908591 (Tracking ID: 2922721)

SYMPTOM:
After suppressing fault, managed host is still shown as faulty.

DESCRIPTION:
When the faults are suppressed for a managed host, the state of the
managed host does not change.

RESOLUTION:
This is because one of the database views was not taking into account
the suppressed faults table. The issue has been resolved by correcting the
database view.

* 2926661 (Tracking ID: 2932853)

SYMPTOM:
Discovery of DS8800 array is failing due to out of memory issue in Perl process.

DESCRIPTION:
Perl process is consuming around 3.5 to 3.7 GB of memory during DS8800 array
discovery. Since VRTSsfmh perl is 32 bit binary, it hits memory limit which
results in perl process to be out of memory error. This issue occurs only if the
array configuration is huge.

RESOLUTION:
Optimized the memory consumption taken by array_driver.pl Perl process.
Following changes are done only for array discovery.
1. Removed pretty formatting of JSON files.
2. Disabled sfm.dump and of SQL statements generation as both of them are not
used for array discovery.

* 2930831 (Tracking ID: 2923139)

SYMPTOM:
Size of the archive and time to gather logs are too large.

DESCRIPTION:
Vomgather collects all data in var dir including DB and stores directory. It 
increases the size of the archive and it takes too long to archive.

RESOLUTION:
Modify vomgather.pl and vom_bkup.pl with default option to not collect DB 
and stores and provide an option "--full" to collect everything including DB and 
stores.

* 2939361 (Tracking ID: 2943751)

SYMPTOM:
Incorrectly showing UUID instead of hostname.

DESCRIPTION:
Faults raised using fault.pl were incorrectly showing UUID instead of host name for the source attribute.

RESOLUTION:
Use API "get_mh_hostname" to read sfm_resolv.conf file and get the value for mh_hostname. If we are not able to find the mh_hostname from any of the domain entries, we will be falling back to get_fqdn().

* 2941489 (Tracking ID: 2975368)

SYMPTOM:
Cross Mirror check is not returning correct violation.

DESCRIPTION:
Cross Enclosure Mirroring Policy check was not showing correct violations for Mirror which have subdisk in plex from same enclosure.

RESOLUTION:
Modified the Cross Enclosure Policy to show violations for Mirror which have subdisk in plex from same enclosure by using the plex <-> subdisk co-relation.

* 2968906 (Tracking ID: 2981057)

SYMPTOM:
luxadm is invoked by VOM to check tape devices.

DESCRIPTION:
luxadm display invoked by VOM discovery on tape device causes it to dump core.

RESOLUTION:
Modified the code to exclude /dev/rmt and /dev/sg from getting discovered during SF discovery on Solaris platform so luxadm will not get invoked for these devices.

* 2969824 (Tracking ID: 2981329)

SYMPTOM:
DB is getting collected multiple times with vombackup utility in HA environment.

DESCRIPTION:
During HA configuration old var dirs are archived into SFMha_dir/SFMha.tar.
During backup, this file is also getting backed up, this causes backing up of DB multiple times.

RESOLUTION:
Modify vom_bkup.pl to exclude the SFMha.tar file from getting backed up.

* 2976267 (Tracking ID: 3027169)

SYMPTOM:
Vulnerability on port port 5634/tcp over SSL.

DESCRIPTION:
There is no SSLCipherSuite entry in VRTSatlocal.conf file on Solaris X86 MH hosts and there is no support for reading SSLCipherSuite entry in AT binaries in VOM 5.0 Sol X86.

RESOLUTION:
A new entry for "SSLCipherSuite" needs to be added into VRTSatlocal.conf. AT binaries shipped with the HF will read the key from 
configuration file for cipher suite.

* 2985578 (Tracking ID: 2986954)

SYMPTOM:
Agentless host is not getting configured in VOM.

DESCRIPTION:
The vbs script is not handling passwords with special characters correctly.

RESOLUTION:
Modified the vbs script to handle passwords with special characters.

* 3000090 (Tracking ID: 3004297)

SYMPTOM:
We do see sfmsecd core dump after upgrading from VOM 4.1 to VOM 5.0 due to bad hash for AdminUserPassword.

DESCRIPTION:
The code for converting plain text password for ldap to encrypted is not working in 64 bit machines like solaris 64 bit.

RESOLUTION:
The code has been fixed by removing cast to int pointer.

* 3001140 (Tracking ID: 3013488)

SYMPTOM:
The managed host appears faulted on the VOM CMS with xprtld not running fault, even though xprtld is running on the mh.

DESCRIPTION:
The mh is not able to send ping requests to the CMS, hence the CMS raises the xprtld not running fault and marks the host as faulted.

RESOLUTION:
On certain Linux machines, the bind on the local port 5634 succeeds even though the port is already taken, and this was causing a failure on a loopback test that was issued.

This loopback check is no longer needed and has been removed, as the xprtld communication mechanism has been made asynchronous.

* 3011255 (Tracking ID: 3014934)

SYMPTOM:
VOM does not display information about VxVM and Filesystems if VXDCLID is reported down.

DESCRIPTION:
Discovery is sending deletes for VxVM objects when vxdclid is down.

RESOLUTION:
The code has been fixed to prevent DB deletes when vxdclid is down.

* 3011375 (Tracking ID: 3011197)

SYMPTOM:
Incorrect capacities are reported in VOM console for some tablespaces of Oracle database.

DESCRIPTION:
If there are Oracle tablespaces with same matching names across multiple Oracle database on the same host, then the vom  oracle-agentlet added up the capacities for these same name Oracle tablespaces.

RESOLUTION:
The code has been changed to not aggregate the capacities of these same name Oracle tablespace.

* 3013115 (Tracking ID: 3069347)

SYMPTOM:
Vulnerability on port 5634/tcp over SSL on AIX MH hosts.

DESCRIPTION:
There is no SSLCipherSuite entry in VRTSatlocal.conf file on AIX MH hosts and there is no support for reading SSLCipherSuite entry in
AT binaries in VOM 5.0 AIX MH.

RESOLUTION:
A new entry for "SSLCipherSuite" needs to be added into VRTSatlocal.conf. AT binaries shipped with the HF will read the key from
configuration file for cipher suite.

* 3013212 (Tracking ID: 3020472)

SYMPTOM:
vxdclid was dumping core while collecting stats.

DESCRIPTION:
vxdclid has been crashing with SIGSEGV fault in process_iostats_for_dmpstats at 30 minutes interval on several MHs when CS polls the subject MHs.

RESOLUTION:
The code has been fixed to prevent vxdclid from dumping core when stats collection is enabled.

* 3030371 (Tracking ID: 3030200)

SYMPTOM:
vxdclid spins when Stats set to false in dcli_conf.ini.

DESCRIPTION:
vxdclid was going into an infinite loop when vxdclid was restarted after disabling the stats collection in the dcli_conf.ini.

RESOLUTION:
The code has been fixed by changing the interval to 60 sec instead of 0.

* 3046832 (Tracking ID: 3067192)

SYMPTOM:
vomgather is failing with tar error while archiving the required files and directories.

DESCRIPTION:
When a file or directory is changing and tar is trying to archive the modified files or directory, it throws an error.

RESOLUTION:
During tar, it should ensure that no files or directories are
changing. Stop the xprtld service during vomgather so that no discovery should
run and temporary files for scheduled discoveries will not be created during tar.

* 3048841 (Tracking ID: 3048888)

SYMPTOM:
Agentless configuration fails to connect to remote host using WMI commands.

DESCRIPTION:
Agentless configuration fails to connect to remote host using WMI commands due to an issue in parsing of EMCPP marker.

RESOLUTION:
Fixed the parsing code to eliminate the requirement of extra space for EMCPP marker.

* 3073174 (Tracking ID: 3079981)

SYMPTOM:
vomgather is failing due to out of memory errors.

DESCRIPTION:
vomgather archive logs with store which makes archive very large and results in out of memory issue.

RESOLUTION:
Excluded store directory from the archive.

* 3080750 (Tracking ID: 3099364)

SYMPTOM:
Windows MH with SFW installed reports incorrect disk serial number.

DESCRIPTION:
On Windows MH, When SFW is installed, the disk serial number does not match serial number reported by VEA.

RESOLUTION:
On Windows, When SFW is installed, pick up the serial number from the UDID.

* 3082536 (Tracking ID: 3083463)

SYMPTOM:
xprtld hangs on AIX systems after reboot.

DESCRIPTION:
netstat -ai hangs when DNS is not configured correctly.

RESOLUTION:
Changed switches for netstat from netstat -ai to netstat -ain to show the IP 
addresses as numbers.

* 3083291 (Tracking ID: 3099464)

SYMPTOM:
Volume Migration fails when multiple volumes are migrated together.

DESCRIPTION:
Volume Migration fails when multiple volumes are migrated together due to a race condition.

RESOLUTION:
Fixed the code for the task to merge two tasks events instead of overwriting.

* 3087219 (Tracking ID: 3099468)

SYMPTOM:
Some volumes are not shown in VOM UI for a Windows managed host.

DESCRIPTION:
On Windows, in a clustered Disk Group setup, some volumes which are 
part of the DG are not visible, even though the DG is in imported state.

RESOLUTION:
The Windows SF discovery removes the volume entries from VOM CMS when 
the DG is deported. These delete queries arrive after the insert queries from the 
other host in the cluster, causing the issue. This has been fixed and the Windows 
SF discovery no longer removes the volume entries on DG deport.

* 3089901 (Tracking ID: 3089904)

SYMPTOM:
Server panics or hangs.

DESCRIPTION:
VOM triggers a panic in Storage Foundation while discovering thin disk capacity.

RESOLUTION:
Remove the VOM code which discovers thin disk capacity.

* 3093373 (Tracking ID: 3093370)

SYMPTOM:
A user without proper privilege was able to send a file to any VOM MH and execute it.

DESCRIPTION:
xprtld allows a user to send a file to any VOM MH and execute it without proper access permission.

RESOLUTION:
Check correctly for user roles for user commands in xprtld.

* 3093852 (Tracking ID: 3099435)

SYMPTOM:
HA FireDrill does not complete (stays in progress).

DESCRIPTION:
HA FireDRILL task was not getting reported to CMS properly because of race 
condition between multiple discoveries trying to report the task.

RESOLUTION:
Implemented the locking mechanism so that only one discovery can process the 
events at a time and the status is reported correctly.

* 3098688 (Tracking ID: 3099409)

SYMPTOM:
Unable to delete FireDrill schedule with localized characters.

DESCRIPTION:
If the name of the scheduled HA FireDrill contains localized characters (eg. 
Spanish), the delete operation fails with the error 'Failed to delete the 
schedule(s)'.

RESOLUTION:
The code has been fixed by decoding the Name for handling the special characters.

* 3103093 (Tracking ID: 3113733)

SYMPTOM:
After patching from 5.1SP1RP1 to 5.1SP1RP3, the system is now logging the 
vxkeyless nagware message to the syslog.

DESCRIPTION:
The system is logging the vxkeyless nagware message because the entry for 
cs_config_name is missing in the file sfm_resolv.conf which is checked by 
licensing remindware library.

RESOLUTION:
The code has been fixed by adding cs_config_name entry in the sfm_resolv.conf file.

* 3111579 (Tracking ID: 3149346)

SYMPTOM:
Cannot attach Managed Hosts to Central Management Server.

DESCRIPTION:
gendeploy script fails when configuring some MHs to secondary VOM CMS.

RESOLUTION:
Added default user as nobody for getvitals url to succeed.

* 3117726 (Tracking ID: 3101454)

SYMPTOM:
A logged in user can change the URL input parameters for archived reports.

DESCRIPTION:
This Hotfix provides enhanced validation of user inputs, and discards requests 
with invalid parameter values.

RESOLUTION:
Enhanced validation of request parameters in  URL.

* 3142907 (Tracking ID: 3032850)

SYMPTOM:
High cpu usage by xprtld process on MH when VOM Server is not available.

DESCRIPTION:
When a VOM MH is cross connected to 2 different VOM Servers, and one of
Servers is unavailable, the xprtld on the MH shows high CPU usage.

RESOLUTION:
Return error if socket not available, retry again later.

* 3154957 (Tracking ID: 3162553)

SYMPTOM:
Replications view does not show secondary for RVG after successful migrate.

DESCRIPTION:
On SF Windows VVR setup, when a migrate/takeover VVR operation is performed 
from VEA, VOM does not update the new state of VVR hosts right away. The change 
can take 30 mins to be reflected in VOM console.

RESOLUTION:
VOM was earlier not listening to change events of VEA. This has been fixed and 
the SFW real time discovery is now triggered for change events as well. The VOM 
console now shows the updated state of VVR hosts in a short time of about 2 
minutes after performing the migrate/takeover operation from VEA.

* 3184832 (Tracking ID: 3213855)

SYMPTOM:
Policy Template Signature(On Disk Config Size) reports violations incorrectly.

DESCRIPTION:
On Disk Config Size policy check was showing incorrect violations for Disk 
Groups having correct On Disk Configuration Size.

RESOLUTION:
Modified the policy check to show violations only for those Disk groups which 
contains On Disk Configuration Size as 0.

* 3185995 (Tracking ID: 3205035)

SYMPTOM:
Hotfix HF050001960-12 causes loss of virtualization (zone) details on VOM console.

DESCRIPTION:
HF050001960-12 introduced a new dependency in the tdf file which caused the issue.

RESOLUTION:
Added the tdf files to zone virtualization family.

* 3186189 (Tracking ID: 3186349)

SYMPTOM:
Javascript errors using Performance Statistics Tab on VOM Server.

DESCRIPTION:
The Javascript errors were seen due to missing encoding of few special characters in the URL.

RESOLUTION:
Fixed the code to encode the special characters.

* 3193430 (Tracking ID: 3205179)

SYMPTOM:
VOM GUI is not showing proper 'Free Space' for volumes.

DESCRIPTION:
On SF Windows setup, VOM does not frequently update the Free Space/Used Space 
attribute even when the capacity monitoring for the volume is enabled from VEA. 
The change can take 30 mins to be reflected in VOM console.

RESOLUTION:
VOM was earlier not listening to change events of VEA. This has been fixed and 
the SFW real time discovery is now triggered for change events as well. The VOM 
console now shows the changed attributes in a short time of about 2 minutes.

* 3207832 (Tracking ID: 3210965)

SYMPTOM:
Plain text vCenter password visible when ps -ef command is run on control host.

DESCRIPTION:
When discovering objects from vCenter, the vCenter password passed to 
vmwarecollector cli was in plain text and hence visible when ps -ef command was run on control host.

RESOLUTION:
The password is now encrypted and then sent to vmwarecollector cli. 
Thus when ps -ef command is run, the original password is not visible.

* 3225782 (Tracking ID: 3225767)

SYMPTOM:
vxdclid is spinning on vxlist requests.

DESCRIPTION:
vxdclid can spin or lock in OpenSSL code on vxlist requests.

RESOLUTION:
The issue was fixed by making OpenSSL calls thread-safe.

* 3234866 (Tracking ID: 3235078)

SYMPTOM:
VOM service is sending the excessive invalid DNS queries to DNS.

DESCRIPTION:
xprtld was doing lookups which was sending excessive invalid DNS queries to DNS.

RESOLUTION:
The code has been fixed such that xprtld won't do the lookup which was causing the 
invalid DNS requests.

* 3236091 (Tracking ID: 3237310)

SYMPTOM:
JRE needs to be updated to the latest version which contains security fixes.

DESCRIPTION:
VOM was shipping JRE which needed to be updated with the latest available version.

RESOLUTION:
Updated JRE with the latest available version 1.7.0_25.

* 3263284 (Tracking ID: 3082055)

SYMPTOM:
A user can view the uploaded scripts without logging into VOM.

DESCRIPTION:
A user can view the uploaded scripts without logging into VOM. A user can also 
view scripts in other directories by modifying url.

RESOLUTION:
vom_requirement.xml validates the user before showing the scripts. Allow users to 
see the scripts in the permitted directory only.