RX2620 No boot after VMS83I_UPDATE-V0400

[Max Pierre]

Hi,

On a RX2620 we installed the patchs VMS83I_LIBRAR-V0200 followed by VMS83I_UPDATE-V0400. The reboot was not successfull (system stopped after the OpenVMS banner). We were fortunately able to boot from the DVD and could de-install the update v4 patch. System was bootable again from the system disk ( direct attached, raid 1)

Any idea of what could be wrong ?

Thanks in advance

[Max Pierre]

I forgot to tell that the system time is one year back in the past. A temporary license expired a few days ago and we have , as workaround, put the system date in the past. The patchs were installed with a system date set somewhere in 2006. We have to keep the date of system in the past until we get the "unlimited" licenses ordered few weeks ago.

[Volker Halle · Accepted solution]

Max,

welcome to the OpenVMS ITRC forum - a 30 year old operating system and still new users showing up ;-)

I installed those patches in the reverse order on a rx2660 on wednesday this week and I had no problem booting: upgraded from VMS83I_UPDATE-V0200 plus patches to VMS83I_UPDATE-V0400, rebooted and then installed VMS83I_LIBRAR-V0200 and others.

Did you capture the console output of the patch installations and boot ? Any errors ?

For troubleshooting: once it's hung, you could crash the machine (use CTRP-P CTRP-P y or better CTRL-B CM TC) and try to find out what was hanging).

Volker.

[Max Pierre]

Volker,

Many thanks so far for your tests.
I am new here but working on VMS since 1985, mainly system management. I don't have much experience on integrity.

I did the patch installations from a PC and maybe was the capture active. I'll check that this evening. I didn't noticed anything special, there were a lot of messages regarding the EFI partition, but no errors (if my memory deserves me well).
We tried Ctrl P when the system was hanging but got the IPC prompt where nothing else than quorum adjustments or crash dump seems available.
We also had to redefine the EFI boot environment after the installation of the patchs. We did it with the dcl utility when the rx2620 was rebooted from the DVD.

I'll have the opportunity to retry the installations next week, I'll post the results.
Maybe should I try to install the update-v0400, but one patch at the time with a reboot between each installation ?

And finally, we decided to install all "mandatory" patchs because we have big LAT problems. On a same DS700 port, one session to VAX is fine, on the integrity it is very slow and sometimes the session is disconnected, but that's another story :-)

[Volker Halle]

Max,

to force a crash, just type CTRL-P while at the IPC prompt. The system will respond with:

Crash y/n:

And will crash once you type Y. Using the MP command TC (Transfer of Control) provides a cleaner forced crash footprint (bugcheck code: CPUINT_INIT) and should be preferred.

If you expect (or experienced) a boot error, you might want to boot with the extended diagnostics flags:

fsn:\EFI\VMS> VMS_LOADER.EFI -flags 0,30000

and capture the full console output.

I did not have to redefine an EFI boot settings after installing VMS83I_UPDATE-V0400.

Volker.

[Max Pierre]

Volker,

Thanks for the tips. I'll try them next week. I'll also write down all firmware revisions of the system and come back with all the results.

Regards and as we are so far from it ... good weekend.

[Max Pierre]

Hi,

I was today on site and able to reproduce the problem (hang at boot) but EFI settings were not affected this time.
I was not able to use the MP serial port, so the system was crashed from IPC.
I did a boot with fl 0,30000 and noticed, but that's maybe normal, several messages like %READ_CONFIG-I-NO_ENTRY, could not find ID (lo) 0x00E01033, (hi) 0x00000000, mask (lo) 0xFFFFFFFF, (hi) 0x00000000, adp 68 (PCI).

I tried to install the patchs contained in the V4 update one by one but was unabled to do it as the update V2 (or later) has to be installed before applying the patch.
I was unable to find the previous versions (V2 or V3). Are these patchs still available somewhere ?

Thanks in advance,

Max

[Volker Halle]

Max,

it's quite obvious from your console log file, that the installation of VMS83I_UPDATE-V0400 had caused the system to hang during boot. Removing that patch again allowed it to boot.

You would need to analyse the crash forced during the hang situation. Consider to create a boot -fl 0,30000 session file with a working boot and try to find out what's happening NEXT after:

%SWAPPER-I-SHUFFLE, executing SWAPPER initialization code
%SWAPPER-I-SYSDISK, checking status of system disk
...

Then look at the dump to try to find out, why THAT is NOT happening.

Consider to log a call with HP and provide the forced crashdump and your session log file as data for the call.

Volker.

[Volker Halle]

Max,

after displaying the

%SWAPPER-I-SYSDISK

message, the swapper loops and waits for the ONLINE bit to be set in the UCB of the system disk.

$ ANAL/CRASH SYS$SYSTEM
SDA> sho dev/addr=@exe$gl_sysucb
...
Device status: 88021810 online,...

Does is show online in your forced crash ?


Superseeded patches are typically deleted from the hp ITRC FTP site, so there is no way to obtain old patches, if you don't keep them yourself.

Volker.

[Max Pierre]

Volker,

Thanks for your help.
I'll do the checks you proposed to do next week and post the results.

Best regards,

Max

[Max Pierre]

Volker,

I was yesterday on site and the problem is solved by setting the system date back to "normal ( from 11/2006 to 11/2007)

Unfortunately, I wasn't able to put back the system in the past to double check this "feature". From the beginning, I had doubts about the time set in the past. It would be interesting to see if this is can be reproduced at will. Set the time one year back in the past, install the patch and reboot.

So what I did yesterday was :

- Install definitive licenses
- Set the system back in 2007
- Reboot the system
- Install the global patch
- Reboot the system (no hang this time)

And for info

...
%SWAPPER-I-SHUFFLE, executing SWAPPER initialization code
%SWAPPER-I-SYSDISK, checking status of system disk
%SWAPPER-I-POOL, initializing paged pool
...

Thanks for the help

Best regards,

Max

[Volker Halle]

Max,

thanks for the update. You didn't check the online bit of the system disk in the forced dump ? Maybe you can still do this next time you'll be onsite.

Volker.

[Max Pierre]

Volker,

I don't know if I'll ever go there soon.
I was called to replace someone that was sick for a few weeks but that person takes his job over next week ...

Max