[Info-vax] Orphaned processes on OpenVMS

[Jan-Erik Soderholm]

Paul Sture wrote 2011-05-27 18:42:
> In article<irogrs$41t$1 at news.albasani.net>,
>   Jan-Erik Soderholm<jan-erik.soderholm at telia.com>  wrote:
>
>> Johnny Billquist wrote 2011-05-27 03:47:
>>> On 2011-05-26 13.36, Rob Brown wrote:
>>>> On Thu, 26 May 2011 at 11:50 -0700, Johnny Billquist wrote:
>>>>
>>>>> On 2011-05-26 11.43, Rob Brown wrote:
>>>>>>
>>>>>> In VMS, as Bob says, you must initialize before you can mount.
>>>>>
>>>>> Interesting. So you (or anyone) is allowed to do I/O to an unmounted
>>>>> disk?
>>>>
>>>> Who prevents it in M-Plus? I can't remember when or if I last tried
>>>> this. Can I not, as a privileged or as a non-privileged user, do
>>>> disk-driver QIOs to an unmounted disk? Somebody must be able to,
>>>> otherwise how would MOUNT work?
>>>
>>> No, no I/O is allowed to an unmounted disk in M+. This is controlled by the
>>> QIO directive processing itself.
>>> Check around line 480 (label 46$) and forward in LB:[11,10]DRQIO.MAC
>>>
>>> Mount is not doing only doing I/O. In order for any I/O operation (with a
>>> few exceptions) to be allowed, the volume valid bit must be set.
>>> Mount starts by setting the volume valid, before it do any I/O, and then
>>> resets it again, in case the mount wasn't good.
>>>
>>>>> That would (to me) seem like a weird thing. So when is there ever a
>>>>> point for foreign mounted disks in VMS?
>>>>
>>>> Tradition? Looks like RMS needs the disk to be mounted. I wonder if RMS
>>>> is used for device-mode I/O as well as for file I/O.
>>>
>>> Well, RMS cannot play with foreign disks, or...?
>>>
>>>> $ ld connect deleteme.dsk
>>>> %LD-I-UNIT, Allocated device is OMEGA$LDA2:
>>>> $ dump/block=count=1 lda2: !LDA2 is not mounted
>>>> %DUMP-E-OPENIN, error opening LDA2:[BROWN.TEMP].; as input
>>>> -RMS-E-DNR, device not ready, not mounted, or unavailable
>>>> $ mount lda2 test
>>>> %MOUNT-I-MOUNTED, TEST mounted on _OMEGA$LDA2:
>>>> $ dump/block=count=1 lda2: !LDA2 is mounted files-11
>>>> %DUMP-E-READERR, error reading LDA2:
>>>> -SYSTEM-F-NOPRIV, insufficient privilege or object protection violation
>>>> $ dismount lda2:
>>>> $ mount lda2 test/for
>>>> %MOUNT-I-MOUNTED, TEST mounted on _OMEGA$LDA2:
>>>> $ dump/block=count=1 lda2: !LDA2 is mounted foreign
>>>>
>>>> Dump of device LDA2: on 26-MAY-2011 14:25:21.84
>>>>
>>>> Logical block number 0 (00000000), 512 (0200) bytes
>>>>
>>>> 03039401 001E65C0 11C00200 15C600A0 ......e...... 000000
>>>> ...
>>>
>>> Hmm. Ok. So it does not actually allow I/O to unmounted volumes. So, is
>>> disk file struction initialization special, in order to allow that I/O?
>>>
>>> Johnny
>>>
>>
>> INIT neads the device to be ALLOCated, not MOUNTed, if I'm not wrong.
>
> You don't need to do an ALLOCATE for an INITIALIZE, but it can be
> prudent to do so (to stop other jobs and/or users getting hold of it).
>

OK. There you see. I've always done an ALLOC before any INIT's
the last 20 years... :-)