VERITAS StorageCentral (tm) 5.0.219 Service Pack 1

Article:TECH91717  |  Created: 2003-01-06  |  Updated: 2004-01-06  |  Article URL http://www.symantec.com/docs/TECH91717
Article Type
Technical Solution

Environment

Issue



VERITAS StorageCentral (tm) 5.0.219 Service Pack 1

Solution



This service pack is an update for StorageCentral build 219. Please note that it will only install on an existing build 219 of StorageCentral. Because this update includes new drivers, a reboot is required once installation is complete so that these new drivers can be loaded.


Steps to install the service pack:

1. Run SC5_Build219_SP1.exe
2. Follow installation wizard
3. Reboot machine

Note: If you have one or more of the following QFEs already installed,  you will need to uninstall them prior to installing this service pack: Q219001, Q219011, or Q219013. While uninstalling these QFEs,  you can hold off on the requested reboot until after the last one has been uninstalled.


      Files Updated
File Name            File Date
========            =======
InstallSC5Cluster.exe      6/13/02
RemoveSC5Cluster.exe      4/29/02
SCAlloc.exe            5/31/02
SCBlock.exe            5/31/02
SCReport.exe            5/31/02
SCSRM.exe            5/31/02
WQFSSvr.exe            4/26/02
WQQASvr.exe            6/14/02
WQRptSvr.exe            5/7/02
FileSystem.dll            6/13/02
QuotaScan.dll            6/14/02
SCControl.dll            6/5/02
SCSnapin.dll            6/13/02
WQRptGen.dll            5/14/02
WQRptRes00.dll            5/13/02
WQSC5Type.dll            6/12/02
FileScreenFilter.sys      4/25/02
NmSvFsf.sys            2/28/02
QaFilter.sys            5/1/02
StorageCentralSRM5.CAB      6/13/02


The update provided in this service pack includes all of the fixes in the previous QFEs Q219001, Q219002, ..., Q219015 for build 219 (Note that Q219010 and Q219012 were never released). A summary of the QFE fixes included are as follows:

Q219001 FIX: Server hang while rebooting or remote connection dropping
Q219002 FIX: Management Console hang on connection to NT 4 machine and reporting updates
Q219003 FIX: When file blocking is set to passive, no alarm messages are triggered
Q219004 FIX: Transparent policies can now be applied using the scalloc command line
Q219005 FIX: The Mailto field doesn't use ";" as a separator. Trigger below alarm was causing problems with other alarm messages
Q219006 FIX: Browser Services no longer used to validate list of computers in Active Directory. Addition of an option for user to adjust refresh rate of Managed Objects
Q219007 FIX: Managed objects still appear on the failed node after cluster fail over
Q219008 FIX: The notify user net send and MailTo features do not work when a user ID has got a space between the names
Q219009 FIX: Errors applying policies to partitions in a cluster environment. File Screening objects are only deleted on one node in a cluster
Q219011 FIX: Server hangs while running DFS or constantly reboots after installing StorageCentral 5.0
Q219013 FIX: Windows 'Access Denied' error or stop error 0x1e & 0x50 while deleting nested folders
Q219014 FIX: Scheduled reports do not run more than once
Q219015 FIX: StorageCentral reports now display proper user profile information from the Active Directory along with OU information and file ownership

Additional updates included in this service pack include:

- Resolution of the "Server Busy" errors received when attempting to expand the list of drives during the creation of a managed object or while selecting an object to run a report against. Also resolved "Server Busy" error received when connecting to an NT machine via Windows 2000 preferred machine
- Resolution of issue with FileBlocking objects not being copied to other node during cluster fail over
- Resolution of the issue where the policy associated with a managed object would not appear in the management console
- Resolution of the disappearing managed objects issue. The possible corruption of the quota database has been resolved
- Resolution of the time out error message when drives were expanded



Details on cause and resolution of issues resolved with the QFEs rolled up into the service pack are as follows:

Q219001: Server hang while rebooting or remote connection dropping
Your application may experience problems after installing StorageCentral SRM 5.0 and rebooting where the server will hang at the "Preparing Network Connections" screen. In addition, you may be experiencing remote connections being dropped or denied. These issues were resolved with the fix provided in Q219001.

For the server hang while rebooting, there was a "race condition" in the NmSvFsf.sys where it was possible for a paged lookaside list to be initialized twice. This caused the lookaside list structure to be inserted into the system-wide list of lookaside lists twice, generating a cycle in the list. Some time later, the balance set manager decided to walk that list while holding a spinlock, hit the cycle, and therefore kept all I/O completion DPCs from executing. This causes what appears to be a hang in the SCSI driver. It usually appears at boot, but may also be seen during server operation.

For the dropping or denial of remote connections, there was a problem with the volume dismount code in QaFilter driver builds 96 through 99, which can cause the process doing the dismount to hang. Basically, the thread is waiting for the dismount IRP to complete, when that IRP has actually already completed. Depending on what process is doing the dismount (e.g. the Cluster Service), this could cause remote connections to be dropped or denied.


Q219002 FIX: Management Console hang on connection to NT 4 machine and reporting updates
Your application may experience a Management Console hang upon connection with an NT 4 machine or may receive a "There are no computers with StorageCentral SRM installed within this Active Directory level" error when trying to create a managed object on an NT 4 server. This issue was resolved with the fix provided in Q219002.

Additional issues resolved include updating the "Space by Space Allocation" and "Out of Allocated Space" reports so that the "remote procedure call failed" error no longer occurs. Also additional data sources to the "Chargeback" report have been added.


Q219003 FIX: When file blocking is set to passive, no alarm messages are triggered
Your application may experience problems receiving alarm messages when file blocking is set to passive. This issue was resolved with the fix provided in Q219003. The cause of the problem was that the source code condition to send the notification message on passive file blocking was incorrect.


Q219004 FIX: Transparent policies can now be applied using the scalloc command line
Your application may experience problems applying transparent policies using the command scalloc from the command line. This enhancement was added with Q219004. This enhancement had not been available before because the logic differentiating between transparent and regular policies had not been implemented within the scalloc command line.


Q219005 FIX: The Mailto field doesn't use ";" as a separator. Trigger below alarm was causing problems with other alarm messages
This QFE added the ability to now use ";" as a separator between multiple email addresses in the "mailto" field. In addition, this QFE resolved the alarm message problem seen when the "Trigger Below" value (eg: set at Alarm 3, at 50%), when checked in the policy. After checking the trigger below, all other Alarm messages within that policy would stop functioning. The problem here was caused by the fact that the source code condition for "trigger below" alarm was incorrect.


Q219006 FIX: Browser Services no longer used to validate list of computers in Active Directory. Addition of an option for user to adjust refresh rate of Managed Objects
Your application may experience a problem displaying a list of available computers when Browser Services have been disabled within the network. This QFE updated StorageCentral so that Browser Services are no longer required to display the list of available computers. An additional issue resolved is the addition of an option allowing the user to adjust the default refresh rate of the display of managed objects. With Managed Objects selected in the left pane of the MMC window, right-click in the right pane to bring up the shortcut menu. Select the Auto Refresh Frequency� menu option. Set the desired refresh interval in the Auto Refresh Frequency dialog and click OK. StorageCentral was using the Browser Services to verify the list of computers received from the Active Directory. When Browser Services are disabled none of the computers in the Active Directory are validated and displayed.


Q219007 FIX: Managed objects still appear on the failed node after cluster fail over
Your application may experience problems after you fail over in a cluster environment. You may find managed objects still appearing on the failed node. The problem was that the server was showing all objects, even those that are not accessible. This QFE addressed this issue so that non-accessible objects are now not shown.


Q219008 FIX: The notify user net send and MailTo features do not work when a user ID has got a space between the names
With build 219, net send/pop up messages or emails do not work when there is a space between the user ID, i.e. John Doe versus John.Doe. This QFE updated source code that did not allow for spaces with User IDs. As a result, StorageCentral now allows for spaces within user IDs.


Q219009 FIX: Errors applying policies to partitions in a cluster environment. File Screening objects are only deleted on one node in a cluster
The following problems are addressed by this QFE:
1) Adding a managed object: When you add a managed object with a policy, you can select a disk partition from the browse option. The browse pop-up identifies ALL disk partitions available to cluster in shared storage. If you then select a disk partition owned by Node A (Node that you are administering on), then the policy is applied successfully. If you select a disk partition that is not owned by Node A, you may receive the following message pop-up:
      A Runtime Error has occurred
      Do you wish to debug?
      Line 12
      Error: 'document.frm.explorer.notify fn is not an object.
This error was caused by the server showing all partitions from the StorageCentral browse window, even those that are not accessible. This has been addressed such that non-accessible partitions are now not shown.

2) Deleting an object: If you delete a File Screening object from only one node, and there is a fail-over, the other node will still have the object. This has been addressed by the QFE so that if the object is deleted only on Node A, it will then be removed when a fail-over occurs to Node B.


Q2190011 FIX: Server hangs while running DFS or constantly reboots after installing StorageCentral 5.0
After installing StorageCentral 5.0 build 219 the server may keep rebooting and causing a memory dump. You may also experience hangs if you are running DFS. These crashes were caused by a FileScreen data structure becoming corrupt during scanning. The boot problems were due to drivers not being loaded and changes in the XP/.NET IoRegisterFsRegistrationChange functionality.  The hang with DFS was due to StorageCentral enabling APC calls. The scanning procedure has been changed in the FileScreen driver so that it handles the data structure correctly. The driver also loads properly now and the Move CreateThread before IoRegisterFsRegistrationChange in DriverEntry procedure has changed, so if it fails, cleanup is done properly.  APCs are now disabled as DFS uses these.


Q219013 (Reboot Required) FIX: Windows 'Access Denied' error or stop error 0x1e & 0x50 while deleting nested folders
Your application may experience problems deleting nested folders when a space allocation limit has been reached. Windows might return an 'access denied file may be in use' error while deleting these nested folders. It is also possible to eventually get a stop 0x0000001e, kmode_exception not handled in NTOSKRNL.exe error. Another possible stop error that could occur is 0x00000050. These issues were resolved with this QFE. The cause of these problems were that the QuotaAdvisor driver was unable to handle the situation where a quota object is renamed on one handle and subsequently deleted on a different simultaneously open handle.


Q219014 FIX: Scheduled reports do not run more than once
Your application may experience scheduled reports not running correctly. You may find that the report will only run once and that further scheduled reports will fail to run. This problem was caused by the script/text file being deleted after the report ran once. This issue has been resolved with this QFE and the script/text file located in the Wquinn/StorageCentral SRM/5.0/ Reports is no longer deleted after running once. Thus all occurrences of scheduled reports will run.


Q219015 FIX: StorageCentral reports now display proper user profile information from the Active Directory along with OU information and file ownership
Your application may experience issues reporting on the user profile fields in the Active Directory along with properly reporting on file ownership of files from mapped drives. In addition, information on which Organizational Unit (OU) a user belongs was not available. This QFE addresses all of these issues. The problem was determined to be caused by the StorageCentral report generation code not properly pulling information from the user profile in the Active Directory. In addition, there was no code in place to return information on the OU to which a user belonged. Finally, the source code had incorrect logic in place to report on the file ownership of files contained on mapped drives. After installing software update Q219015, StorageCentral report definitions and report sets can be created that include the OU along with user profile information from the Active Directory such as: user account information like user name, telephone number, email, account expiration, title, department, and company. In addition, reports will now display correct file ownership on mapped as well as local drives.

Attachments

SC5_Build219_SP1_258976.exe (27.2 MBytes)


Legacy ID



258976


Article URL http://www.symantec.com/docs/TECH91717


Terms of use for this information are found in Legal Notices