Just upgraded a VM from 9.2-1 to 9.2-2 which seemed to have worked well.

Bearing in mind this us just a hobbyist system, when I created the VM I
gave it 8GB memory, and an 8Gb system disk

That was fine, and it created a 2Mb pagefile, not expecting it to be used.

The autogen run at the end of the upgrade obviously though this was
rubbish, so tried to create an 8Gb pagefile on the system disk. Clearly
it couldn't, so it created one of over 4Gb - leaving me with 120Mb free.

It also wanted an 8Gb dumpfile.

Is any of this sensible?

--
Chris

On 2/24/2024 2:18 PM, Chris Townley wrote:
> Just upgraded a VM from 9.2-1 to 9.2-2 which seemed to have worked well.
>
> Bearing in mind this us just a hobbyist system, when I created the VM I
> gave it 8GB memory, and an 8Gb system disk
>
> That was fine, and it created a 2Mb pagefile, not expecting it to be used.
>
> The autogen run at the end of the upgrade obviously though this was
> rubbish, so tried to create an 8Gb pagefile on the system disk. Clearly
> it couldn't, so it created one of over 4Gb - leaving me with 120Mb free.
>
> It also wanted an 8Gb dumpfile.
>
> Is any of this sensible?

I believe that is one of the differences between Windows and VMS. On
Windows plenty of RAM and no/small pagefile works fine. VMS want
pagefile backing of virtual memory even if it is not likely to
need it. And running out of pagefile space is bad - very bad. So I got
1 x 8 GB pagefile on my VMS x86-64 and 3 x 1 GB pagefile on
my VMS Alpha (can't remember what I have on my VMS Itanium).

Arne

On 25/02/2024 00:12, Arne Vajhøj wrote:
> On 2/24/2024 2:18 PM, Chris Townley wrote:
>> Just upgraded a VM from 9.2-1 to 9.2-2 which seemed to have worked well.
>>
>> Bearing in mind this us just a hobbyist system, when I created the VM
>> I gave it 8GB memory, and an 8Gb system disk
>>
>> That was fine, and it created a 2Mb pagefile, not expecting it to be
>> used.
>>
>> The autogen run at the end of the upgrade obviously though this was
>> rubbish, so tried to create an 8Gb pagefile on the system disk.
>> Clearly it couldn't, so it created one of over 4Gb - leaving me with
>> 120Mb free.
>>
>> It also wanted an 8Gb dumpfile.
>>
>> Is any of this sensible?
>
> I believe that is one of the differences between Windows and VMS. On
> Windows plenty of RAM and no/small pagefile works fine. VMS want
> pagefile backing of virtual memory even if it is not likely to
> need it. And running out of pagefile space is bad - very bad. So I got
> 1 x 8 GB pagefile on my VMS x86-64 and 3 x 1 GB pagefile on
> my VMS Alpha (can't remember what I have on my VMS Itanium).
>
> Arne

But VSI recommended 8Gb memory, and an 8Gb system disk. So how do we fit
that in? Especially if they want another 8Gb for the dump file

--
Chris

On 2/24/2024 7:33 PM, Chris Townley wrote:
> On 25/02/2024 00:12, Arne Vajhøj wrote:
>> On 2/24/2024 2:18 PM, Chris Townley wrote:
>>> Just upgraded a VM from 9.2-1 to 9.2-2 which seemed to have worked well.
>>>
>>> Bearing in mind this us just a hobbyist system, when I created the VM
>>> I gave it 8GB memory, and an 8Gb system disk
>>>
>>> That was fine, and it created a 2Mb pagefile, not expecting it to be
>>> used.
>>>
>>> The autogen run at the end of the upgrade obviously though this was
>>> rubbish, so tried to create an 8Gb pagefile on the system disk.
>>> Clearly it couldn't, so it created one of over 4Gb - leaving me with
>>> 120Mb free.
>>>
>>> It also wanted an 8Gb dumpfile.
>>>
>>> Is any of this sensible?
>>
>> I believe that is one of the differences between Windows and VMS. On
>> Windows plenty of RAM and no/small pagefile works fine. VMS want
>> pagefile backing of virtual memory even if it is not likely to
>> need it. And running out of pagefile space is bad - very bad. So I got
>> 1 x 8 GB pagefile on my VMS x86-64 and 3 x 1 GB pagefile on
>> my VMS Alpha (can't remember what I have on my VMS Itanium).
>
> But VSI recommended 8Gb memory, and an 8Gb system disk. So how do we fit
> that in? Especially if they want another 8Gb for the dump file

I don't get it. 8 GB RAM makes sense. I would go for 20/30/40/50 GB
disk.

I don't think you need that big a dump file. It may want that big
a dump file, but if I remember correctly then VMS writes a subset
of data if the dumpfile is not big enough for all data. Obviously
if you want perfect analysis of a crash, then a full size dumpfile
may be nice, but ...

Arne

On 25/02/2024 00:52, Arne Vajhøj wrote:
> On 2/24/2024 7:33 PM, Chris Townley wrote:
>> On 25/02/2024 00:12, Arne Vajhøj wrote:
>>> On 2/24/2024 2:18 PM, Chris Townley wrote:
>>>> Just upgraded a VM from 9.2-1 to 9.2-2 which seemed to have worked
>>>> well.
>>>>
>>>> Bearing in mind this us just a hobbyist system, when I created the
>>>> VM I gave it 8GB memory, and an 8Gb system disk
>>>>
>>>> That was fine, and it created a 2Mb pagefile, not expecting it to be
>>>> used.
>>>>
>>>> The autogen run at the end of the upgrade obviously though this was
>>>> rubbish, so tried to create an 8Gb pagefile on the system disk.
>>>> Clearly it couldn't, so it created one of over 4Gb - leaving me with
>>>> 120Mb free.
>>>>
>>>> It also wanted an 8Gb dumpfile.
>>>>
>>>> Is any of this sensible?
>>>
>>> I believe that is one of the differences between Windows and VMS. On
>>> Windows plenty of RAM and no/small pagefile works fine. VMS want
>>> pagefile backing of virtual memory even if it is not likely to
>>> need it. And running out of pagefile space is bad - very bad. So I got
>>> 1 x 8 GB pagefile on my VMS x86-64 and 3 x 1 GB pagefile on
>>> my VMS Alpha (can't remember what I have on my VMS Itanium).
>>
>> But VSI recommended 8Gb memory, and an 8Gb system disk. So how do we
>> fit that in? Especially if they want another 8Gb for the dump file
>
> I don't get it. 8 GB RAM makes sense. I would go for 20/30/40/50 GB
> disk.
>
> I don't think you need that big a dump file. It may want that big
> a dump file, but if I remember correctly then VMS writes a subset
> of data if the dumpfile is not big enough for all data. Obviously
> if you want perfect analysis of a crash, then a full size dumpfile
> may be nice, but ...
>
> Arne

But I only have system on the system disc. ISTR when we moved to IA64,
we started with 8Gb, but had to increase it to have the Oracle client.

I wonder what VSI have to say. When the resolve the gateway timeout, I
might post there

--
Chris

On 2/24/2024 7:02 PM, Chris Townley wrote:
> On 25/02/2024 00:52, Arne Vajhøj wrote:
>> On 2/24/2024 7:33 PM, Chris Townley wrote:
>>> On 25/02/2024 00:12, Arne Vajhøj wrote:
>>>> On 2/24/2024 2:18 PM, Chris Townley wrote:
>>>>> Just upgraded a VM from 9.2-1 to 9.2-2 which seemed to have worked well.
>>>>>
>>>>> Bearing in mind this us just a hobbyist system, when I created the VM I gave it 8GB memory, and an 8Gb system disk
>>>>>
>>>>> That was fine, and it created a 2Mb pagefile, not expecting it to be used.
>>>>>
>>>>> The autogen run at the end of the upgrade obviously though this was rubbish, so tried to create an 8Gb pagefile on the system disk. Clearly it couldn't, so it created one of over 4Gb - leaving me with 120Mb free.
>>>>>
>>>>> It also wanted an 8Gb dumpfile.
>>>>>
>>>>> Is any of this sensible?
>>>>
>>>> I believe that is one of the differences between Windows and VMS. On
>>>> Windows plenty of RAM and no/small pagefile works fine. VMS want
>>>> pagefile backing of virtual memory even if it is not likely to
>>>> need it. And running out of pagefile space is bad - very bad. So I got
>>>> 1 x 8 GB pagefile on my VMS x86-64 and 3 x 1 GB pagefile on
>>>> my VMS Alpha (can't remember what I have on my VMS Itanium).
>>>
>>> But VSI recommended 8Gb memory, and an 8Gb system disk. So how do we fit that in? Especially if they want another 8Gb for the dump file
>>

I'm not sure where you saw that, but section 1.2.1 of the V9.2-2 Install says:
https://docs.vmssoftware.com/vsi-openvms-x86-64-v922-installation-guide/#d0e99

Memory: 8 GB.

Note that if your installation fails due to the lack of memory, you can increase the amount of it for the duration of the installation and then, if needed, reduce it back.

Operating system: Other 64-bit.

A virtual disk for the system files. The disk must be at least 15 GB in size

>> I don't get it. 8 GB RAM makes sense. I would go for 20/30/40/50 GB
>> disk.
>>
>> I don't think you need that big a dump file. It may want that big
>> a dump file, but if I remember correctly then VMS writes a subset
>> of data if the dumpfile is not big enough for all data. Obviously
>> if you want perfect analysis of a crash, then a full size dumpfile
>> may be nice, but ...
>>
>> Arne
>
> But I only have system on the system disc. ISTR when we moved to IA64, we started with 8Gb, but had to increase it to have the Oracle client.
>
> I wonder what VSI have to say. When the resolve the gateway timeout, I might post there
>

On my V9.2-2 install (not upgrade)I created a 24GB system disk. I allocated 12GB of memory and it creates a 2MB page file. Install left 10.2GB free so Install uses about 14GB on my system. This includes an 11.74GB dump file to match the 11.74GB usable memory. Autoconfig at the end did not try to make a bigger page file nor create a swapfile.

Disk NORTON$DKA0:, device type VMware Virtual disk, is online, mounted, file-
oriented device, shareable, available to cluster, error logging is enabled.

Error count 0 Operations completed 11874
Owner process "" Owner UIC [SYSTEM]
Owner process ID 00000000 Dev Prot S:RWPL,O:RWPL,G:R,W
Reference count 30 Default buffer size 512
Current preferred CPU Id 1 Fastpath 1
Total size 24.00GB Sectors per track 0
Total cylinders 0 Tracks per cylinder 0
Logical Volume Size 24.00GB Expansion Size Limit 383.99GB

Volume label "NORTON_A0" Relative volume number 0
Cluster size 3 Transaction count 182
Free space 10.27GB Maximum files allowed 16711679
Extend quantity 5 Mount count 1
Mount status System Cache name "_NORTON$DKA0:XQPCACHE"
Extent cache size 64 Max blocks in extent cache 2153983
File ID cache size 64 Blocks in extent cache 30345
Quota cache size 0 Maximum buffers in FCP cache 4734
Volume owner UIC [1,1] Vol Prot S:RWCD,O:RWCD,G:RWCD,W:RWCD

Volume Status: ODS-5, subject to mount verification, protected subsystems
enabled, file high-water marking, write-through XFC caching enabled,
write-through XQP caching enabled, hard links enabled, special files
enabled.

$ show mem
System Memory Resources on 24-FEB-2024 22:05:53.79

Physical Memory Usage (bytes): Total Free In Use Modified
Main Memory (GB) 11.74 11.15 0.58 0.00

Extended File Cache (Time of last reset: 16-FEB-2024 22:59:52.90)
Allocated (MBytes) 32.50 Maximum size (MBytes) 6014.39
Free (MBytes) 0.64 Minimum size (MBytes) 3.12
In use (MBytes) 31.85 Percentage Read I/Os 84%
Read hit rate 59% Write hit rate 0%
Read I/O count 7545 Write I/O count 1368
Read hit count 4482 Write hit count 0
Reads bypassing cache 26 Writes bypassing cache 1
Files cached open 205 Files cached closed 137
Vols in Full XFC mode 0 Vols in VIOC Compatible mode 2
Vols in No Caching mode 0 Vols in Perm. No Caching mode 0

Granularity Hint Regions (bytes): Total Free In Use Released
S0 Execlet data (MB) 16.00 14.53 1.46 0.00
S0 Executive data (MB) 36.00 0.15 35.84 0.00
S0 Executive RO data (MB) 8.00 6.54 1.45 0.00
S0 Resident image code (MB) 24.00 23.35 0.64 0.00
S0 Resident image data (MB) 4.00 4.00 0.00 0.00
S0 Resident RO image data (MB) 8.00 8.00 0.00 0.00
S2 Execlet code (MB) 32.00 15.16 16.82 0.00
S2 Execlet data (MB) 32.00 32.00 0.00 0.00
S2 Executive data (MB) 8.00 0.00 8.00 0.00
S2 Resident image code (MB) 32.00 11.80 20.19 0.00
S2 Resident image data (MB) 4.00 4.00 0.00 0.00

Slot Usage (slots): Total Free Resident Swapped
Process Entry Slots 949 936 13 0

Dynamic Memory Usage: Total Free In Use Largest
Nonpaged Dynamic Memory (MB) 24.00 21.56 2.43 21.50
USB Addressable Memory (KB) 1024.00 1022.87 1.12 1022.87
Paged Dynamic Memory (MB) 11.74 6.80 4.93 6.80
Lock Manager Dyn Memory (KB) 592.00 306.03 285.96
S2 Dynamic Memory Usage (MB) 7.97 7.69 0.28 7.69

Buffer Object Usage (bytes): In Use Peak
32-bit System Space Windows (S0/S1) (bytes) 0.00 0.00
64bit System Space Windows (S2) (bytes) 0.00 0.00
Physical bytes locked by buffer objects (bytes) 0.00 0.00

Memory Reservations (bytes): Group Reserved In Use Type
Total (0 bytes reserved) 0 0

Paging File Usage (bytes): Index Free Size
DISK$NORTON_A0:[SYS0.SYSEXE]PAGEFILE.SYS;1
(MB) 254 2.18 2.18
Total committed paging file usage:(MB) 18.71

Of the physical memory in use, 572.25 MB are permanently allocated to OpenVMS.
$ dir DISK$NORTON_A0:[SYS0.SYSEXE]PAGEFILE.SYS;1

Directory DISK$NORTON_A0:[SYS0.SYSEXE]

PAGEFILE.SYS;1 2.24MB/2.24MB 27-JAN-2024
19:55:46.08 [1,1] (RWED,RWED,,)

Total of 1 file, 2.24MB/2.24MB

$ dir DISK$NORTON_A0:[SYS0.SYSEXE]sysdump.dmp

Directory DISK$NORTON_A0:[SYS0.SYSEXE]

SYSDUMP.DMP;1 11.74GB/11.74GB 27-JAN-2024
21:41:45.65 [1,1] (RWED,RWED,,)

Total of 1 file, 11.74GB/11.74GB

--
John H. Reinhardt

On Sat, 24 Feb 2024 19:12:48 -0500, Arne Vajhøj wrote:

> On Windows plenty of RAM and no/small pagefile works fine.

Really?? Last I heard, Windows would not run without a page/swap file. Did
they fix that?

On 25/2/24 11:33, Chris Townley wrote:
> On 25/02/2024 00:12, Arne Vajhøj wrote:
>> On 2/24/2024 2:18 PM, Chris Townley wrote:
>>> Just upgraded a VM from 9.2-1 to 9.2-2 which seemed to have worked well.
>>>
>>> Bearing in mind this us just a hobbyist system, when I created the VM
>>> I gave it 8GB memory, and an 8Gb system disk
>>>
>>> That was fine, and it created a 2Mb pagefile, not expecting it to be
>>> used.
>>>
>>> The autogen run at the end of the upgrade obviously though this was
>>> rubbish, so tried to create an 8Gb pagefile on the system disk.
>>> Clearly it couldn't, so it created one of over 4Gb - leaving me with
>>> 120Mb free.
>>>
>>> It also wanted an 8Gb dumpfile.
>>>
>>> Is any of this sensible?
>>
>> I believe that is one of the differences between Windows and VMS. On
>> Windows plenty of RAM and no/small pagefile works fine. VMS want
>> pagefile backing of virtual memory even if it is not likely to
>> need it. And running out of pagefile space is bad - very bad. So I got
>> 1 x 8 GB pagefile on my VMS x86-64 and 3 x 1 GB pagefile on
>> my VMS Alpha (can't remember what I have on my VMS Itanium).
>>
>> Arne
>
> But VSI recommended 8Gb memory, and an 8Gb system disk. So how do we fit
> that in? Especially if they want another 8Gb for the dump file
>

In a production environment, by putting the page file (and secondary
page files) and the dump file on disks other than the system disk. The
last thing you want is more paging to and from the system disk which
does enough demand zero paging as it is due to image activation. With
real hardware, locally connected disks are nice for page files as long
as you can provide redundancy for them (hardware mirroring with a raid
controller, for example).

In a hobbyist situation, do you need a page file that large? Unless you
have an app that uses masses of memory, you can probably get away with
something smaller. I have 6GB of memory with 3GB of page file, and I
rarely use more than 1% of the page file. See warnings about no dump
file and selective dumps in the documentation (qv) if you do this!

In a hobbyist situation, do you need a dump file? Unless you're doing
kernel mode development, perhaps not. Set DUMPBUG=0 in MODPARAMS.DAT,
reboot, and delete the dump file.

If you do want a dump file, you can probably get away with something
much smaller by specifying an appropriate value for DUMPSTYLE in
MODPARAMS.DAT. You should at least compress the dump by setting bit 3,
and choose other bits for that parameter based on your circumstances.

Remember, if a dump file doesn't exist either in SYS$SYSTEM: or in a
DOSD location that you've specified, VMS will use the page file to write
the crash dump.

There's an entire chapter devoted to all this in the "OpenVMS System
Manager's Manual Vol 2". Find it here:

https://vmssoftware.com/docs/VSI_SYS_MGMT_MANUAL_VOL_II.PDF

Jim.
--
eight-cubed.com

On 2/25/2024 12:30 AM, Lawrence D'Oliveiro wrote:
> On Sat, 24 Feb 2024 19:12:48 -0500, Arne Vajhøj wrote:
>> On Windows plenty of RAM and no/small pagefile works fine.
>
> Really?? Last I heard, Windows would not run without a page/swap file. Did
> they fix that?

First time I heard about it was in the early 00's.

Arne

On 2/24/2024 11:11 PM, John H. Reinhardt wrote:
> On 2/24/2024 7:02 PM, Chris Townley wrote:
>> On 25/02/2024 00:52, Arne Vajhøj wrote:
>>> I don't get it. 8 GB RAM makes sense. I would go for 20/30/40/50 GB
>>> disk.
>>>
>>> I don't think you need that big a dump file. It may want that big
>>> a dump file, but if I remember correctly then VMS writes a subset
>>> of data if the dumpfile is not big enough for all data. Obviously
>>> if you want perfect analysis of a crash, then a full size dumpfile
>>> may be nice, but ...
>>>
>>> Arne
>>
>> But I only have system on the system disc. ISTR when we moved to IA64,
>> we started with 8Gb, but had to increase it to have the Oracle client.
>>
>> I wonder what VSI have to say. When the resolve the gateway timeout, I
>> might post there
>>
>
> On my V9.2-2 install (not upgrade)I created a 24GB system disk.  I
> allocated 12GB of memory and it creates a 2MB page file.  Install left
> 10.2GB free so Install uses about 14GB on my system.  This includes an
> 11.74GB dump file to match the 11.74GB usable memory.  Autoconfig at the
> end did not try to make a bigger page file nor create a swapfile.

I got a very small pagefile as well by default. But I increased the
size.

I believe it is still:

memory usage exceed WSEXTENT => application can not allocate memory and
crash

running out of PAGEFILE => system crashes

so WSEXTENT < sizeof(PAGEFILE) makes sense to me.

Arne

On 2/25/2024 3:37 AM, Jim Duff wrote:
> On 25/2/24 11:33, Chris Townley wrote:
>> On 25/02/2024 00:12, Arne Vajhøj wrote:
>> VMS want
>>> pagefile backing of virtual memory even if it is not likely to
>>> need it. And running out of pagefile space is bad - very bad. So I got
>>> 1 x 8 GB pagefile on my VMS x86-64 and 3 x 1 GB pagefile on
>>> my VMS Alpha (can't remember what I have on my VMS Itanium).
>>
>> But VSI recommended 8Gb memory, and an 8Gb system disk. So how do we
>> fit that in? Especially if they want another 8Gb for the dump file

> In a hobbyist situation, do you need a page file that large?  Unless you
> have an app that uses masses of memory, you can probably get away with
> something smaller.  I have 6GB of memory with 3GB of page file, and I
> rarely use more than 1% of the page file.

On my system @sys$startup:tomcat$startup and
@sys$startup:activemq$startup without even using them
bump usage to 600 MB.

> In a hobbyist situation, do you need a dump file?  Unless you're doing
> kernel mode development, perhaps not.  Set DUMPBUG=0 in MODPARAMS.DAT,
> reboot, and delete the dump file.
>
> If you do want a dump file, you can probably get away with something
> much smaller by specifying an appropriate value for DUMPSTYLE in
> MODPARAMS.DAT.  You should at least compress the dump by setting bit 3,
> and choose other bits for that parameter based on your circumstances.
>
> Remember, if a dump file doesn't exist either in SYS$SYSTEM: or in a
> DOSD location that you've specified, VMS will use the page file to write
> the crash dump.

Yes. Unless one really are into reading crash dumps, then
dumps are not so much fun.

Arne

On 25/02/2024 04:11, John H. Reinhardt wrote:
> On 2/24/2024 7:02 PM, Chris Townley wrote:
>> On 25/02/2024 00:52, Arne Vajhøj wrote:
>>> On 2/24/2024 7:33 PM, Chris Townley wrote:
>>>> On 25/02/2024 00:12, Arne Vajhøj wrote:
>>>>> On 2/24/2024 2:18 PM, Chris Townley wrote:
>>>>>> Just upgraded a VM from 9.2-1 to 9.2-2 which seemed to have worked
>>>>>> well.
>>>>>>
>>>>>> Bearing in mind this us just a hobbyist system, when I created the
>>>>>> VM I gave it 8GB memory, and an 8Gb system disk
>>>>>>
>>>>>> That was fine, and it created a 2Mb pagefile, not expecting it to
>>>>>> be used.
>>>>>>
>>>>>> The autogen run at the end of the upgrade obviously though this
>>>>>> was rubbish, so tried to create an 8Gb pagefile on the system
>>>>>> disk. Clearly it couldn't, so it created one of over 4Gb - leaving
>>>>>> me with 120Mb free.
>>>>>>
>>>>>> It also wanted an 8Gb dumpfile.
>>>>>>
>>>>>> Is any of this sensible?
>>>>>
>>>>> I believe that is one of the differences between Windows and VMS. On
>>>>> Windows plenty of RAM and no/small pagefile works fine. VMS want
>>>>> pagefile backing of virtual memory even if it is not likely to
>>>>> need it. And running out of pagefile space is bad - very bad. So I got
>>>>> 1 x 8 GB pagefile on my VMS x86-64 and 3 x 1 GB pagefile on
>>>>> my VMS Alpha (can't remember what I have on my VMS Itanium).
>>>>
>>>> But VSI recommended 8Gb memory, and an 8Gb system disk. So how do we
>>>> fit that in? Especially if they want another 8Gb for the dump file
>>>
>
> I'm not sure where you saw that, but section 1.2.1 of the V9.2-2 Install
> says:
> https://docs.vmssoftware.com/vsi-openvms-x86-64-v922-installation-guide/#d0e99
>
> Memory: 8 GB.
>
> Note that if your installation fails due to the lack of memory, you can
> increase the amount of it for the duration of the installation and then,
> if needed, reduce it back.
>
> Operating system: Other 64-bit.
>
> A virtual disk for the system files. The disk must be at least 15 GB in
> size
>
>
>>> I don't get it. 8 GB RAM makes sense. I would go for 20/30/40/50 GB
>>> disk.
>>>
>>> I don't think you need that big a dump file. It may want that big
>>> a dump file, but if I remember correctly then VMS writes a subset
>>> of data if the dumpfile is not big enough for all data. Obviously
>>> if you want perfect analysis of a crash, then a full size dumpfile
>>> may be nice, but ...
>>>
>>> Arne
>>
>> But I only have system on the system disc. ISTR when we moved to IA64,
>> we started with 8Gb, but had to increase it to have the Oracle client.
>>
>> I wonder what VSI have to say. When the resolve the gateway timeout, I
>> might post there
>>
>
> On my V9.2-2 install (not upgrade)I created a 24GB system disk.  I
> allocated 12GB of memory and it creates a 2MB page file.  Install left
> 10.2GB free so Install uses about 14GB on my system.  This includes an
> 11.74GB dump file to match the 11.74GB usable memory.  Autoconfig at the
> end did not try to make a bigger page file nor create a swapfile.
>
>
>
>
> Disk NORTON$DKA0:, device type VMware Virtual disk, is online, mounted,
> file-
>     oriented device, shareable, available to cluster, error logging is
> enabled.
>
>     Error count                    0    Operations
> completed              11874
>     Owner process                 ""    Owner UIC
> [SYSTEM]
>     Owner process ID        00000000    Dev Prot
> S:RWPL,O:RWPL,G:R,W
>     Reference count               30    Default buffer
> size                 512
>     Current preferred CPU Id       1
> Fastpath                              1
>     Total size               24.00GB    Sectors per
> track                     0
>     Total cylinders                0    Tracks per
> cylinder                   0
>     Logical Volume Size      24.00GB    Expansion Size Limit
> 383.99GB
>
>     Volume label         "NORTON_A0"    Relative volume
> number                0
>     Cluster size                   3    Transaction
> count                   182
>     Free space               10.27GB    Maximum files allowed
> 16711679
>     Extend quantity                5    Mount
> count                           1
>     Mount status              System    Cache name
> "_NORTON$DKA0:XQPCACHE"
>     Extent cache size             64    Max blocks in extent cache
> 2153983
>     File ID cache size            64    Blocks in extent
> cache            30345
>     Quota cache size               0    Maximum buffers in FCP
> cache       4734
>     Volume owner UIC           [1,1]    Vol Prot
> S:RWCD,O:RWCD,G:RWCD,W:RWCD
>
>   Volume Status:  ODS-5, subject to mount verification, protected
> subsystems
>       enabled, file high-water marking, write-through XFC caching enabled,
>       write-through XQP caching enabled, hard links enabled, special files
>       enabled.
>
> $ show mem
>               System Memory Resources on 24-FEB-2024 22:05:53.79
>
> Physical Memory Usage (bytes):     Total        Free      In Use
> Modified
>   Main Memory             (GB)     11.74       11.15        0.58
> 0.00
>
> Extended File Cache  (Time of last reset: 16-FEB-2024 22:59:52.90)
>  Allocated (MBytes)             32.50    Maximum size (MBytes)
> 6014.39
>  Free (MBytes)                   0.64    Minimum size
> (MBytes)             3.12
>  In use (MBytes)                31.85    Percentage Read
> I/Os                84%
>  Read hit rate                     59%   Write hit
> rate                       0%
>  Read I/O count                  7545    Write I/O
> count                   1368
>  Read hit count                  4482    Write hit
> count                      0
>  Reads bypassing cache             26    Writes bypassing
> cache               1
>  Files cached open                205    Files cached
> closed                137
>  Vols in Full XFC mode              0    Vols in VIOC Compatible
> mode         2
>  Vols in No Caching mode            0    Vols in Perm. No Caching
> mode        0
>
> Granularity Hint Regions (bytes):   Total        Free      In Use
> Released
>   S0 Execlet data           (MB)    16.00       14.53
> 1.46        0.00
>   S0 Executive data         (MB)    36.00        0.15
> 35.84        0.00
>   S0 Executive RO data      (MB)     8.00        6.54
> 1.45        0.00
>   S0 Resident image code    (MB)    24.00       23.35
> 0.64        0.00
>   S0 Resident image data    (MB)     4.00        4.00
> 0.00        0.00
>   S0 Resident RO image data (MB)     8.00        8.00
> 0.00        0.00
>   S2 Execlet code           (MB)    32.00       15.16
> 16.82        0.00
>   S2 Execlet data           (MB)    32.00       32.00
> 0.00        0.00
>   S2 Executive data         (MB)     8.00        0.00
> 8.00        0.00
>   S2 Resident image code    (MB)    32.00       11.80
> 20.19        0.00
>   S2 Resident image data    (MB)     4.00        4.00
> 0.00        0.00
>
> Slot Usage (slots):                Total        Free    Resident
> Swapped
>   Process Entry Slots                949         936
> 13           0
>
> Dynamic Memory Usage:              Total        Free      In Use
> Largest
>   Nonpaged Dynamic Memory (MB)     24.00       21.56        2.43
> 21.50
>   USB Addressable Memory  (KB)   1024.00     1022.87        1.12
> 1022.87
>   Paged Dynamic Memory    (MB)     11.74        6.80        4.93
> 6.80
>   Lock Manager Dyn Memory (KB)    592.00      306.03      285.96
>   S2 Dynamic Memory Usage (MB)      7.97        7.69        0.28
> 7.69
>
> Buffer Object Usage (bytes):                              In Use
> Peak
>   32-bit System Space Windows (S0/S1)     (bytes)
> 0.00        0.00
>   64bit System Space Windows (S2)         (bytes)
> 0.00        0.00
>   Physical bytes locked by buffer objects (bytes)
> 0.00        0.00
>
> Memory Reservations (bytes):       Group    Reserved      In Use
> Type
>   Total (0 bytes reserved)                         0           0
>
> Paging File Usage (bytes):                     Index        Free
> Size
>   DISK$NORTON_A0:[SYS0.SYSEXE]PAGEFILE.SYS;1
>                                     (MB)         254        2.18
> 2.18
>   Total committed paging file usage:(MB)
> 18.71
>
> Of the physical memory in use, 572.25 MB are permanently allocated to
> OpenVMS.
> $ dir DISK$NORTON_A0:[SYS0.SYSEXE]PAGEFILE.SYS;1
>
> Directory DISK$NORTON_A0:[SYS0.SYSEXE]
>
> PAGEFILE.SYS;1                                     2.24MB/2.24MB
> 27-JAN-2024
>  19:55:46.08  [1,1]            (RWED,RWED,,)
>
> Total of 1 file, 2.24MB/2.24MB
>
> $ dir DISK$NORTON_A0:[SYS0.SYSEXE]sysdump.dmp
>
> Directory DISK$NORTON_A0:[SYS0.SYSEXE]
>
> SYSDUMP.DMP;1                                     11.74GB/11.74GB
> 27-JAN-2024
>  21:41:45.65  [1,1]            (RWED,RWED,,)
>
> Total of 1 file, 11.74GB/11.74GB
>

On Sun, 25 Feb 2024 07:54:36 -0500, Arne Vajhøj wrote:

> On 2/25/2024 12:30 AM, Lawrence D'Oliveiro wrote:
>
>> On Sat, 24 Feb 2024 19:12:48 -0500, Arne Vajhøj wrote:
>>
>>> On Windows plenty of RAM and no/small pagefile works fine.
>>
>> Really?? Last I heard, Windows would not run without a page/swap file.
>> Did they fix that?
>
> First time I heard about it was in the early 00's.

Remember when “netbooks” first came along? Initial Linux-based models like
the Asus Eee 701 (I have one) were the first machines with SSDs and no
hard drive. Microsoft had just introduced Vista, which was hopelessly
inept on running on such small machines. So they postponed the death of
XP, and together with Intel, they managed to put limits on the specs of
these machines, so that they could continue running XP.

XP was hard on SSDs, because it required a swap file, and the continual
writing to the SSD shortened its life.

On 2024-02-24, Chris Townley <news@cct-net.co.uk> wrote:
> Just upgraded a VM from 9.2-1 to 9.2-2 which seemed to have worked well.
>
> Bearing in mind this us just a hobbyist system, when I created the VM I
> gave it 8GB memory, and an 8Gb system disk
>
> That was fine, and it created a 2Mb pagefile, not expecting it to be used.
>
> The autogen run at the end of the upgrade obviously though this was
> rubbish, so tried to create an 8Gb pagefile on the system disk. Clearly
> it couldn't, so it created one of over 4Gb - leaving me with 120Mb free.
>

That was a stupid thing for autogen to do.

There should be some inbuilt free space limit so that autogen never leaves
the free space on the disk below something like 10%. That way, you can
continue with the upgrade/installation without trashing the system disk
by filling it up, and hence you get the opportunity to fix that on a reboot.

The other thing I don't like about this is that autogen should run initially
in test mode during the installation/upgrade, and then only commit the
changes after you are happy with them. You know, just like we are all taught
to do when adjusting autogen parameters manually. :-)

> It also wanted an 8Gb dumpfile.
>

That should have been optional. Were you asked if you wanted a dumpfile ?
If not, that's probably something else that should be added.

So, does anyone agree or disagree with the above suggestions ?

Simon.

--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Walking destinations on a map are further away than they appear.

First of all there is no need for a specific dump file at all. Setting SAVEDUMP allows the use of the pagefile for a dump that can then be copied elsewhere to free up the pagefile upon reboot. Also, setting dumpstyle such that the selective dump is selected, will reduce the size requirements and most often save sufficient information to analyze the dump and locate the problem that caused the crash.

Most systems have a small pagefile on the system disk and other larger ones on alternative devices to address any pagefile size requirements as well as performance.

This has been the case for years and should not be news to anyone here. I recall using the above pagefile techniques for VAX systems 30+ years ago. Even the dump file (if you want a specific one) can be on an alternative drive (see DOSD).

Dan