This FixPak is in a somewhat different format than previous ones. It is being delivered in the same format as it is within the WSoD Rel 2 product. It is a zip file that should be unzipped to a directory and can be installed directly from that directory. The directory contents will look just as it would if all the fixpak diskettes were copied from the diskette to the directory. The contents of the Fixtool diskette have also been copied into a sub-directory called TOOLS. Service can be started with a command file that has been included, INSLANFP.CMD. All that is needed to install the FixPak is to unpack the zip file into a directory using the '-d' parameter to create stored directories, execute INSLANFP and follow the normal procedures for installing a FixPak. NOTE: If you need to back out this FixPak, do the following: 1. Run INSLANFP and select "Change product list." 2. Select "Backed up products" and then click on OK. 3. Make sure that the proper product components are selected and click on Backout. 4. When this phase of the backout is complete, you will be asked to reboot the machine if there were any locked files. The following APARs are included in FixPak IPx8410: APAR = IC19816 NET32SHAREENUM TRAPS THE GUI ON MACHINES WITH A LARGE NUMBER OF SHARES. When an OS/2 v4.0 client tries to access information about shares on a server with a large number of shares, the GUI may trap when displaying the information. APAR = IC21398 AS NET.ACC FAILS TO SYNC WITH DC WHEN AS UAS CONTAINS ONE ADDTIONAL USER THAN DCS. NET3226 C708 ERRORS. In situations where the UAS on the Additional/Backup Server contains at least one more user than the DC, NET.ACC files fail to sync. This can happen when the Backup has been changed to STANDALONE mode or made primary and users have been added. When its role is made BACKUP, the new DC and the changed server fail to sync UAS (NET.ACC files). Net3226 errors with C708 are logged on AS's. APAR = IC21399 UPM0010C (SYS0065) DURING LOGON WITH EXPIRED PASSWORD. Steps to recreate this problem are as follows: 1. Use a testid 2. Logon on with this userid 3. With UPM, change password to AB1CD 4. Logoff and logon with new password ( AB1CD ) 5. With UPM, change password to AB2CD 6. Logoff and logon with new password ( AB2CD ) 7. With UPM, change password to AB3CD 8. Logoff 9. In policy, change the password to expire after 1 day 10. Change the date of your workstation to the following day's date 11. Logon to test userid 12. You will be prompted to change your password 13. Try AB1CD ( old password ) 14. Try AB2CD ( old password ) 15. Try AB4CD ( which is a good password ) At this point, you will get the UPM0010C error. You must be at a requester not at D/C, and you must try 2 old passwords. APAR = IC21400 LOGOFF RETURNS MESSAGE UPM0010C OR UPM1009C WHEN SEVERAL PROCESSES ARE STARTED. System is running Warp Server 4.0 with IP08265 level FixPak. The customer attempts to logoff from a server that has had 117 processes started with over 500 threads in the system. They get an error from the logoff that says: UPM1009C: The command failed. The exit code returned from the command is: 17 If you do a logoff /l and select the logoff button, you get the message UPM0010C with exit code 99. If the customer terminates some processes, he can then log off successfully. APAR = IC21401 TRAP 000D IN NETWKSTA.200 ON SERVER ACCESSING REDIRECTED DRIVES TO OTHER SERVERS. Customer running OS/2 Warp Server with IP08265 applied gets a trap in NETWKSTA.200 module. The server does five NET USEs to five different servers for applications on the server itself. The application on the server was doing a DosRead of a file on one of those drives. APAR = IC21402 SYS3175 IN NETAPI.DLL ON DOMAIN CONTROLLER RUNNING WARP SERVER SMP WHEN MULTIPLE NET.ACC READ OPERATIONS ARE OCCURRING Customer running Warp Server Advanced SMP with FixPak 32 and IP08503 on a PC Server 320 with two processors would sometimes get a trap in NETAPI.DLL on the domain controller. The trap occurred while the NETSERVR process was reading a user record from NET.ACC for a full sync request from additional servers. APAR = IC21403 INCORRECT RESULTS WHEN NET ACCESS X:\ /TREE IS EXECUTED. If Net Access x:\ /tree is issued, no data results. If Net Access x: /tree is issued, the data returned is correct. APAR = IC21404 NET ACCESS /APPLY DOES NOT APPLY THE AUDIT SETTINGS FROM THE ACL BEING APPLIED TO RESOURCES WITH ACLS BUT NO AUDIT SETTINGS Customer attempted a NET ACCESS d:\dir /APPLY to a directory tree. The ACL also contained audit settings. While the permissions were applied to all directories with or without ACLs, the audit settings were not applied to directories or files with existing ACLs but no audit settings. If the directory had no ACL, one was created that had both the permissions and audit settings of the applier ACL. For example: C:\TEST (USERS:RX Audited - FAILURE:ALL) C:\TEST\TEST1 (USERS:R) C:\TEST\TEST2 (USERS:X Audited - SUCCESS:All) C:\TEST\TEST3 (no ACL) After the apply it should come out as: C:\TEST (USERS:RX Audited - FAILURE:ALL) C:\TEST\TEST1 (USERS:RX Audited - FAILURE:ALL) C:\TEST\TEST2 (USERS:RX Audited - SUCCESS:ALL) APAR = IC21406 NET AUDIT AND NET ERROR COMMANDS MAY GET NET2379 ON IP08506. Customers who have applied IP08506 have seen the following problems with NET AUDIT and/or NET ERROR: 1. "NET2379: This log file is damaged." 2. "There are no entries in the list." when there are entries in it. 3. The NET.EXE traps. 4. Potential infinite loop. If the NET.EXE module is replaced with a previous level such as IP08503 or IP08267 level, the commands display the logs successfully. Typically, the problems show up with greater than 400-500 entries in the logs. APAR = IC21407 LSAUDIT AND LSERROR UTILITIES LOOP FOREVER WITH MANY LOG ENTRIES ON IP08506 After applying IP08506 the following problem occurs: LSAUDIT is loaded and displays a sub window with the following content: Title: Reading file Text : Please wait Processing record XXXXX XXXXX is a 5 digit number which is increased until 99999, then starts over with 1 and only stops as the PC seems to run out of some resources and gets slower and slower until it has to be rebooted. This effect starts suddenly between 400 to 500 records. The LSERROR utility also displays the same symptoms with greater than 400 log entries. This problem also shows up on LS 5.1 GA and LS 5.1 with IP08503 applied. APAR = IC21408 RESTACC DOES NOT RESTORE ACLS WHEN USING A DIRECTORY NAME LONGER THAN 23 CHARACTERS AS INPUT TO THE COMMAND. When a directory name that is longer than 23 characters is supplied as an input parameter to Restacc, Restacc is not able to locate the entry in the Index Table of the .ACL file--even if the entry is valid and the ACL information exists in the file. For example: D:\12345678\12345\12345678 - will not be found D:\12345678\12345\1234 - will be found. Note: The total includes drive letter, colon, and backslashes. APAR = IC21410 NETBIOSSUBMIT() RETURNS ERROR_INVALID_PARAMETER (RC==87) ON SOME NCBS THAT DO NOT REQUIRE AN NCB_BUFFER. Customer was running Warp Connect Peer requester at GA and applied FixPak IP08196. After this, NCBs submitted through the NetbiosSubmit() APIs that do not require an ncb_buffer to be supplied (such as NCBADDNAME) would get rejected with API return code of 87 (ERROR_INVALID_PARAMETER). Replacing NETAPI.DLL with the GA version enabled the application to work again. This problem has been fixed on Warp Connect Peer. It is only available to customers who have purchased a service extension for that product. This problem also occurs on Warp Server.