-RMS-F-RER, file read error

[Volker Halle]

Dmitry,

please see the OpenVMS I64 V8.3-1H1 Server Upgrade and Installation manual chapter A.2 for how to set up your consoles.

ftp://ftp.hp.com/pub/openvms/doc/BA322-90077.PDF

Volker.

[Volker Halle]

Dmitry,

you need to take the caches into account !

Once you've copied the files and received the parity errors, the disk block data may still be cached somewehere - most likely in the OpenVMS XFC cache. When you then COPY the files again, all the blocks will be taken out of the cache and no real disk access may happen !

This may explain what you're seeing.

Volker.

[Hein van den Heuvel]

Volker wrote>> Once you've copied the files and received the parity errors, the disk block data may still be cached somewehere - most likely in the OpenVMS XFC cache. When you then COPY the files again, all the blocks will be taken out of the cache and no real disk access may happen !


That would seem wrong and NOT the OpenVMS way of doing things. That would imply that a second application could silently get bad data and build on top of that, write it back, and in the process of the write 'fix' the disk through hardware bad block re-vectoring, and all evidence would be gone.

Yikes!

If no one else replies here with more insights, then that is something I'll have to try some say. We could use the LDdriver to inject a parity error, and then test with that.

Hein.

[Volker Halle]

Hein,

you may be right. The cache (XFC) theory was just a good explanation for what Dmitry is seeing. Maybe there are other caches (on disk?) involved as well ? Or the PARITY error is somewhere else and only shows up under certain IO load patterns ?!

Volker.

[Dmitry V. Sinelnikov]

Thanks again, will try to set up console.

As for the cache being involved - I'm not completely sure but it seems that cache is not the case:
1) I 'm copying 100Mb file, after parity error I see only 80Mb size chunk of the initial EXE image in a target directory - COPY operation terminated somewhere on 80th megabyte.
2) I invoke copy operation again and it copies all 100 Mb correctly. If file was cached during first copy operation - there must have been only 80Mb which were cached before the 'parity error' occurred, am I right?

[Volker Halle]

Dmitry,

with your more detailled description given now, I agreee, that the 'cache theory' is wrong.

You need to decode the errlog entries to better understand, what's going wrong.

Volker.

[Dmitry V. Sinelnikov]

Finally had some time to play with 'boot form DVD' configuration... But not sure if I can write it here. Should I start another branch for this issue?

[Volker Halle]

Dmitry,

please start another topic for the 'DVD booting' problem. If necessary, you can still refer to this topic by including a link.

Volker.

[Dmitry V. Sinelnikov]

new branch for 'DVD boot' was created:
http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1370165

[Hoff]

This...

LOADER-W-Conout device path cannot be set to multiple devices
LOADER-I-Select Console Device Paths from the Boot Manager Menu.
...

can be resolved after reading:

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=1338643

and then the documents referenced there...

if you have hardware service for this box, call your provider now.