12-03-2003, 04:23 PM
We upgraded our 4-CPU 820 last weekend, using the Virtual Optical Drive. I was able to do nearly everything from home, using PC Anywhere to access our Operations Console. It had some glitches. Prior to the upgrade, I needed to RCLSTG *DBXREF. Then the first time through about 7am, it only installed the LIC. That may have been because it somehow thought the previous powerdown was abnormal. Or it may have been because I got CPF37A1: Cannot install from optical device OPTVRT01. One or more files are damaged and cannot be processed. CPI3D8C Automatic installation not complete. Then when the DB server jobs started, I got a lot of CPF32A6 Cross-reference notification 1061 for request 2, problem type 1 I would guess that was due to the new QADBXREF not being installed. DSPSFWRSC showed that only the LIC (5722999) had been installed; all else was backlevel. I started the installation process again. It installed everything but 5722SS1 option 01, "OS/400 - Extended Base Support" and 02, "OS/400 - Online Information". A joblog showed something about trying to create some database thing that already existed. I deleted one of the two failed programs, then did another RCLSTG *DBXREF. That appeared to help things. There were several messages in QHST about "CPF7001 Damage was detected but has been ignored" during journal recovery. Then I was finally able to completely install it. I installed the Cume/Hiper/DB2 PTFs from the virtual drive. I was finally able to have our testing team come in then, about 3pm. Our development partition was already V5R2, so I did a compare-and-copy for PTFs not on the Primary partition. I had to download a few that had no SAVF to copy. Overall there were about 400 PTFs. Another reboot. INZSYS completed about 8pm. AFTERWARDS: The upgrade deleted the 'as-sts' Service Tools Server from the server table entries; it is needed to do partition & disk stuff from OpsNav. An IBM APAR says that this is a bug in V5R2 installation that will be fixed in V5R3. I manually added it back. One of our HTTP servers would not start, due to SSL initialization failure. I was able to have the programmer who best understood it log on from home, and he was eventually able to fix it. Not quite sure what the problem was; it had something to do with the digital certificate we'd installed the week before. QSYSOPR showed a lot of TCP2617 "TCP/IP connection to remote system 192.168.1.101 closed, reason code 2". This was due to using OpsNav from my PC at home; the management central server keeps trying to contact it. It'll be fixed in some later release, apparently. I turned off the messages with CHGTCPA TCPCNNMSG(*NONE). PCs running early service packs of Client Access Express V5R1 would get a harmless Windows message. They will get a dialog box titled CWBUNPLA with this message: Error occurred while retrieving usage information for [USER PROFILE] on [SYSTEM NAME]. The solution is to upgrade to a newer service pack or upgrade them to iSeries Access V5R2. A couple of programs used interactively are having sporadic CPU-eating problems; it may be due to V5R2 changes. Otherwise, all our queries & programs work, nothing has blown up, and the pure-SQL database guys report that it works faster.