info
Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
2 Gb/s fibre channel question
178 views
Skip to first unread message
Hans Vlems
Jun 22, 2014, 4:03:03 AM6/22/14
to
Never too old to learn, so I bought an EMC^2 X2E Fibre channel disk cabinet.
Assuming it would connect to 1 Gb/s fibre channel controllers.
Owing to non-matching fibre connectors that won't work so I need new FC controllers.
I have two questions at this point?
- what model FC 2 Gb/s controller is supported in a DS20E, or a DS10 under VMS 8.3?
- I need manuals (user, installation) for the EMC^2 X2E system (more specifically on how I can set up a serial connection with the box).
Hans
Assuming it would connect to 1 Gb/s fibre channel controllers.
Owing to non-matching fibre connectors that won't work so I need new FC controllers.
I have two questions at this point?
- what model FC 2 Gb/s controller is supported in a DS20E, or a DS10 under VMS 8.3?
- I need manuals (user, installation) for the EMC^2 X2E system (more specifically on how I can set up a serial connection with the box).
Hans
Paul Sture
Jun 22, 2014, 7:46:23 AM6/22/14
to
Windows supports.
To the best of my knowledge VMS has never supported FC-AL, though there may
be adaptors or other widgets which can enable VMS to talk to FC-AL devices.
--
You can't look at a glass as half full or half empty if it's overflowing.
Stephen Hoffman
Jun 22, 2014, 7:54:11 AM6/22/14
to
On 2014-06-22 08:03:03 +0000, Hans Vlems said:
> - what model FC 2 Gb/s controller is supported in a DS20E, or a DS10
> under VMS 8.3?
Quickspecs
> - what model FC 2 Gb/s controller is supported in a DS20E, or a DS10
> under VMS 8.3?
<http://www8.hp.com/h20195/v2/GetDocument.aspx?docname=c04285101>, page
18. SPOCK if you want to try the latest list. Compatible FC HBAs have
been discussed before here; try a few searches of the archives, using
the controller part numbers from the QuickSpecs as search targets.
> - I need manuals (user, installation) for the EMC^2 X2E system (more
> specifically on how I can set up a serial connection with the box).
there's more to the name than that (eg: CLARiiON something-something?),
or it was a rare controller.
If it's a DB9, try the PC pinout, and try different speeds. If it uses
an RJ45, there's a possible pinout posted
<http://www.sunhelp.org/pipermail/rescue/2007-November/121959.html>
Using TQE-capable SCSI controllers in a multi-host configuration would
be the other approach for these boxes. That might be a shade less
involved than getting the UDID disk info and related baggage configured
and going for OpenVMS on the EMC FC SAN controller you're working with,
if that X2E doesn't have documented VMS support. It'll usually also be
less parts, as you can cable directly to a SCSI shelf.
--
Pure Personal Opinion | HoffmanLabs LLC
Stephen Hoffman
Jun 22, 2014, 8:11:18 AM6/22/14
to
On 2014-06-22 11:46:23 +0000, Paul Sture said:
> There is a lot of FC-AL (AL = Arbitrated Loop) kit out there; it's what
> Windows supports.
>
> To the best of my knowledge VMS has never supported FC-AL, though there
> may be adaptors or other widgets which can enable VMS to talk to FC-AL
> devices.
Ayup. VMS doesn't support FC-AL, and expects there to be a FC SAN
> There is a lot of FC-AL (AL = Arbitrated Loop) kit out there; it's what
> Windows supports.
>
> To the best of my knowledge VMS has never supported FC-AL, though there
> may be adaptors or other widgets which can enable VMS to talk to FC-AL
> devices.
switch somewhere in the configuration. Either a separate switch box,
or sometimes the switch can be embedded in the FC SAN storage
controller.
Hans Vlems
Jun 22, 2014, 9:03:42 AM6/22/14
to
Op zondag 22 juni 2014 14:11:18 UTC+2 schreef Stephen Hoffman:
I tried to register at the EMC site and that was down a couple of hours ago.
Usually I stick to gear that I understand and Fibre Channel equipment wasn't
around in the shop I worked for when I left (both the job and VMS for a living).
What I assumed was that even though the EMC is not supported that it would talk to a VMS supported controller. The next step was to try a RAID set up and should that fail, use a JBOD configuration.
But I'm afraid a bit more off than I can handle. A 1 Gb/s FC controller uses a different fibre connector. The single 2 Gb/s controller that I have fits a PCI slot that none of my aging Alpha's have.
Thanks so far for the news!
Hans
Usually I stick to gear that I understand and Fibre Channel equipment wasn't
around in the shop I worked for when I left (both the job and VMS for a living).
What I assumed was that even though the EMC is not supported that it would talk to a VMS supported controller. The next step was to try a RAID set up and should that fail, use a JBOD configuration.
But I'm afraid a bit more off than I can handle. A 1 Gb/s FC controller uses a different fibre connector. The single 2 Gb/s controller that I have fits a PCI slot that none of my aging Alpha's have.
Thanks so far for the news!
Hans
Stephen Hoffman
Jun 22, 2014, 9:25:45 AM6/22/14
to
On 2014-06-22 13:03:42 +0000, Hans Vlems said:
> What I assumed was that even though the EMC is not supported that it
> would talk to a VMS supported controller. The next step was to try a
> RAID set up and should that fail, use a JBOD configuration.
It isn't so much the fibre channel host bus adapter (FC HBA) that chats
> What I assumed was that even though the EMC is not supported that it
> would talk to a VMS supported controller. The next step was to try a
> RAID set up and should that fail, use a JBOD configuration.
with the fibre channel storage area network (FC SAN) controller (though
both technically and obviously, those two widgets do necessarily chat),
the usual wrinkle here is whether the operating system can successfully
chat with the storage controllers.
OpenVMS has specific expectations for the FC SAN controller. Usual
among those is the ability to set the UDID for the disks. VMS didn't
chat with all of the HP and Compaq FC SAN controllers available, for
instance. VMS also doesn't operate via arbitrated loop — a fibre
channel switch is expected, and with whatever zoning that brings to the
configuration established. Specific storage controller firmware could
be required. If there was no UDID support or such in the controller,
then either VMS or the controller would have to be modified. (I don't
recall off-hand if there was a bypass some of these requirements; that
allowed the UDID to be "spoofed" within VMS.)
The OpenVMS FC SAN DGDRIVER device driver sends basically-SCSI I/O
command packets to the target FC SAN controller. These packets are
effectively using the SCSI command-response protocol, but with the
hardware quite possibly operating rather further alway from the OpenVMS
host than with the typical SCSI configuration. (IIRC, DGDRIVER and the
stack is a build-time variation of the SCSI DKDRIVER device driver,
too.)
RAID and the rest are usually entirely secondary to establishing and
configuring the initial support expected by VMS.
Skim the (two) VMS cluster manuals in the VMS doc set
<http://www.hp.com/go/openvms/doc> for an overview of the FC SAN
support, from the perspective of VMS. There have been various fibre
channel pages and documents posted at the VMS web site over the years,
as well.
Multihost SCSI will be easier and will console less power and space and
parts and cables.
Hans Vlems
Jun 28, 2014, 5:01:22 AM6/28/14
to
Op zondag 22 juni 2014 15:25:45 UTC+2 schreef Stephen Hoffman:
OK, it's entirely different (in my head at least) than I'd assumed FC would work.
I put a KGPSA in a DS10 running VMS 8.4, no patches applied.
The HBA is recognised, no errors were reported. The HBA was connected directly with a SAN controller. A green led lit up when the cable was inserted. VMS
did not notice, at least not on the console.
Not knowing any better I tried:
SYSMAN> io auto /log
%SYSMAN-I-OUTPUT, command execution on node OZON
%IOGEN-I-PREFIX, searching for ICBM with prefix SYS$
%IOGEN-I-PREFIX, searching for ICBM with prefix DECW$
%IOGEN-I-SCSIPOLL, scanning for devices through SCSI port PKA0
%IOGEN-I-FIBREPOLL, scanning for devices through FIBRE port PGA0
%IOGEN-F-FTLIOERR, fatal I/O error while trying to access device
SYSMAN>
Which proves your point in the previous post, right?
Any suggestions that I could try (other than RTFM's, which I'm doing)?
Objective: just a bunch of disks visible to VMS. No need for RAID.
Hans
> On 2014-06-22 13:03:42 +0000, Hans Vlems said:
>
>
[snip]
>
>
>
>
>
>
>
> --
>
> Pure Personal Opinion | HoffmanLabs LLC
Hoff,
>
>
>
>
> --
>
> Pure Personal Opinion | HoffmanLabs LLC
OK, it's entirely different (in my head at least) than I'd assumed FC would work.
I put a KGPSA in a DS10 running VMS 8.4, no patches applied.
The HBA is recognised, no errors were reported. The HBA was connected directly with a SAN controller. A green led lit up when the cable was inserted. VMS
did not notice, at least not on the console.
Not knowing any better I tried:
SYSMAN> io auto /log
%SYSMAN-I-OUTPUT, command execution on node OZON
%IOGEN-I-PREFIX, searching for ICBM with prefix SYS$
%IOGEN-I-PREFIX, searching for ICBM with prefix DECW$
%IOGEN-I-SCSIPOLL, scanning for devices through SCSI port PKA0
%IOGEN-I-FIBREPOLL, scanning for devices through FIBRE port PGA0
%IOGEN-F-FTLIOERR, fatal I/O error while trying to access device
SYSMAN>
Which proves your point in the previous post, right?
Any suggestions that I could try (other than RTFM's, which I'm doing)?
Objective: just a bunch of disks visible to VMS. No need for RAID.
Hans
Colin Butcher
Jun 28, 2014, 6:32:55 AM6/28/14
to
To get FC devices working with a VMS system there are several steps that
need to happen:
- FC adapters in the system (KGPSA for Alpha, 1 or 2 Gbps; Emulex / Qlogic
for Integrity, 2 / 4 / 8 Gbps, adapter model depends on system type)
- FC adapters in "fabric" mode. VMS does not support "arbitrated loop" mode.
- FC adapters are often referred to as HBAs. HBA = host bus adapter.
- FC switches between the systems and the storage devices (disc array or
tape).
- FC switch zoning configured to pass traffic between storage device and FC
adapter. Zoning is usually "single initiator / multiple target", where an
initiator is the FC adapter and a target is a storage array or tape device.
Zoning is usually done with "soft zoning" by specifying the WWPN / WWNN of
the devices as viewed by the switch when the devices are powered up and
connected to the fabric switch. WWPN = world-wide port name, WWNN =
world-wide node name. A WWPN can have multiple child WWNNs - eg: HSG80 has a
single WWNN representing the controller pair with 4x child WWPNs
representing the FC ports. the WWNN and WWPN are generically referred to as
WWIDs. WWID = world-wide ID.
- You might choose to explicitly set the FC switch port speed, or the
adapter port speed, or both, or just let it auto-negotiate. I prefer to
explicitly set the port speed because it can help to identify fibre
connection problems and the port won't come up (or log in to the fabric) if
the requested speed isn't working.
- Tape devices need to be "found" using SYSMAN IO FIND and it's usually best
to specify the MGA device number explicitly. There's stuff you need to do
with WWIDs if you replace a drive (with a different WWID) and want to keep
the same MGA device name. See SYSMAN IO help. The data file is
SYS$SPECIFIC:[SYSEXE]SYS$DEVICES.DAT
- Disc devices (storage arrays) need to be configured to create storage
units out of the set of available physical discs, then the storage units
need to be presented to the systems they will be accessed by. In EVA
terminology (different arrays are just different enough to be confusing),
create "virtual discs" with a given level of redundancy for data integrity
(vraid1 is best for both availability and performance - least overhead in
controller processing, everything on two physical discs). Discs are cheap,
so use vraid1 and double sparing within a single big disc group of idential
discs on an EVA for best performance.
- A "vdisc" needs to be "presented" to the host system FC adapters by
reference to the HBA WWPN for each adapter. VMS clusters are a special case
to most people because VMS allows the same "vdisc" to be presented to
multiple systems (cluster member nodes), thus giving direct path access to
the disc. When you present the vdisc, you must also specify the UDID (or
OS_UNIT_ID) value, which is used by VMS to set the DGAnnnnn device name.
It's best to explicitly set the LUN_ID and UDID, so that all nodes in a
cluster see the same vdisc using the same values.
The LUN_ID is storage array wide. Beware that the UDID (which controls the
VMS device name) is * fabric wide *. DO NOT have more that one instance of a
vdisc presented with a given UDID in the entire set of fabrics that a node
is connected to. If you do make that mistake, you will see the path count
increase, but not a new device. So, when you write to the DGAnnnnn device,
which actual vdisc you write to depends on which path it's using at that
instant in time.
Note: 3PAR storage arrays currently do not allow you to change the UDID
value once you've created the 3PAR equivalent of a vdisc. With an EVA the
LUN_ID and UDID are set at the time of presentation of the vdisc to a
system. With 3PAR, the UDID is set at the time of creation of the 3PAR
equivalent of a vdisc. So, with 3PAR, if you want to change the UDID, you
have to create another piece of storage and use the array to copy the data
from old to new, then delete the old.
Connectivity: each system should have 2x HBA (or more for greater
bandwidth). Alternating adapters should be connected to alternate fabrics:
eg: FGA + FGC to Fabric A, FGB + FGD to Fabric B. Alterating controller
ports should be connected to alternate fabrics: eg: HSG80 controller ports 1
to fabric A and ports 2 to Fabric B, EVA 6500 controller ports 1 + 3 to
Fabric A and ports 2 + 4 to Fabric B.
Example: 4 HBAs with an EVA6500 (4 ports per controller): Zone FGA to EVA
6500 ports 1 + 3 ( 4 paths), FGB to ports 2 + 4 (4 paths), FGC to ports 1 +
3 (4 paths), FGD to ports 2 + 4 (4 paths). That gives a total of 16 FC
paths, 8 paths per fabric. Now set up the vidiscs and present them to all
four adapters. Finally use SYSMAN IO AUTO /LOG and you should see the
devices come online. There are some FC utilities in SYS$EXAMPLES: as well,
such as FIBRE_SCAN.EXE
As you can expect, firmware versions in HBA, FC switches, storage arrays /
tape devices all need to be compatible too, plus VMS needs appropriate
device support, hence patch kits for FIBRE_SCSI. HP's SPOCK database
attempts to keep track of all that compatibility and interoperability stuff
for HP kit. EMC etc. have their own specific versions for VMS compatibility
as well.
It's all good fun, mostly.
--
Cheers, Colin.
---------------------
Legacy = Stuff that works properly!
need to happen:
- FC adapters in the system (KGPSA for Alpha, 1 or 2 Gbps; Emulex / Qlogic
for Integrity, 2 / 4 / 8 Gbps, adapter model depends on system type)
- FC adapters in "fabric" mode. VMS does not support "arbitrated loop" mode.
- FC adapters are often referred to as HBAs. HBA = host bus adapter.
- FC switches between the systems and the storage devices (disc array or
tape).
- FC switch zoning configured to pass traffic between storage device and FC
adapter. Zoning is usually "single initiator / multiple target", where an
initiator is the FC adapter and a target is a storage array or tape device.
Zoning is usually done with "soft zoning" by specifying the WWPN / WWNN of
the devices as viewed by the switch when the devices are powered up and
connected to the fabric switch. WWPN = world-wide port name, WWNN =
world-wide node name. A WWPN can have multiple child WWNNs - eg: HSG80 has a
single WWNN representing the controller pair with 4x child WWPNs
representing the FC ports. the WWNN and WWPN are generically referred to as
WWIDs. WWID = world-wide ID.
- You might choose to explicitly set the FC switch port speed, or the
adapter port speed, or both, or just let it auto-negotiate. I prefer to
explicitly set the port speed because it can help to identify fibre
connection problems and the port won't come up (or log in to the fabric) if
the requested speed isn't working.
- Tape devices need to be "found" using SYSMAN IO FIND and it's usually best
to specify the MGA device number explicitly. There's stuff you need to do
with WWIDs if you replace a drive (with a different WWID) and want to keep
the same MGA device name. See SYSMAN IO help. The data file is
SYS$SPECIFIC:[SYSEXE]SYS$DEVICES.DAT
- Disc devices (storage arrays) need to be configured to create storage
units out of the set of available physical discs, then the storage units
need to be presented to the systems they will be accessed by. In EVA
terminology (different arrays are just different enough to be confusing),
create "virtual discs" with a given level of redundancy for data integrity
(vraid1 is best for both availability and performance - least overhead in
controller processing, everything on two physical discs). Discs are cheap,
so use vraid1 and double sparing within a single big disc group of idential
discs on an EVA for best performance.
- A "vdisc" needs to be "presented" to the host system FC adapters by
reference to the HBA WWPN for each adapter. VMS clusters are a special case
to most people because VMS allows the same "vdisc" to be presented to
multiple systems (cluster member nodes), thus giving direct path access to
the disc. When you present the vdisc, you must also specify the UDID (or
OS_UNIT_ID) value, which is used by VMS to set the DGAnnnnn device name.
It's best to explicitly set the LUN_ID and UDID, so that all nodes in a
cluster see the same vdisc using the same values.
The LUN_ID is storage array wide. Beware that the UDID (which controls the
VMS device name) is * fabric wide *. DO NOT have more that one instance of a
vdisc presented with a given UDID in the entire set of fabrics that a node
is connected to. If you do make that mistake, you will see the path count
increase, but not a new device. So, when you write to the DGAnnnnn device,
which actual vdisc you write to depends on which path it's using at that
instant in time.
Note: 3PAR storage arrays currently do not allow you to change the UDID
value once you've created the 3PAR equivalent of a vdisc. With an EVA the
LUN_ID and UDID are set at the time of presentation of the vdisc to a
system. With 3PAR, the UDID is set at the time of creation of the 3PAR
equivalent of a vdisc. So, with 3PAR, if you want to change the UDID, you
have to create another piece of storage and use the array to copy the data
from old to new, then delete the old.
Connectivity: each system should have 2x HBA (or more for greater
bandwidth). Alternating adapters should be connected to alternate fabrics:
eg: FGA + FGC to Fabric A, FGB + FGD to Fabric B. Alterating controller
ports should be connected to alternate fabrics: eg: HSG80 controller ports 1
to fabric A and ports 2 to Fabric B, EVA 6500 controller ports 1 + 3 to
Fabric A and ports 2 + 4 to Fabric B.
Example: 4 HBAs with an EVA6500 (4 ports per controller): Zone FGA to EVA
6500 ports 1 + 3 ( 4 paths), FGB to ports 2 + 4 (4 paths), FGC to ports 1 +
3 (4 paths), FGD to ports 2 + 4 (4 paths). That gives a total of 16 FC
paths, 8 paths per fabric. Now set up the vidiscs and present them to all
four adapters. Finally use SYSMAN IO AUTO /LOG and you should see the
devices come online. There are some FC utilities in SYS$EXAMPLES: as well,
such as FIBRE_SCAN.EXE
As you can expect, firmware versions in HBA, FC switches, storage arrays /
tape devices all need to be compatible too, plus VMS needs appropriate
device support, hence patch kits for FIBRE_SCSI. HP's SPOCK database
attempts to keep track of all that compatibility and interoperability stuff
for HP kit. EMC etc. have their own specific versions for VMS compatibility
as well.
It's all good fun, mostly.
--
Cheers, Colin.
---------------------
Legacy = Stuff that works properly!
Hans Vlems
Jun 28, 2014, 8:30:21 AM6/28/14
to
Colin,
thanks, that is a very helpful summary.
In the mean time I've read the wwidmgr manual. I put the KGPSA in loop mode.
One noticable result: I dont get timeout messages the various wwidmgr -show commands.
But the main issue seems to be to connect to the EMC box. That is essential to configure the actual disk devices. The way I try to get it is that, like a Mylex 960 controller, you must set up devices: RAID or JBOD. No devices on the controller none can be presented to VMS.
Back to the EMC2 documentation.
One side question: I have no FC switch so the HBA is connected directly to the SAN controller. Is this by definition an arbitrated loop configuration?
(which is not supported by VMS, but it might work?)
Hans
thanks, that is a very helpful summary.
In the mean time I've read the wwidmgr manual. I put the KGPSA in loop mode.
One noticable result: I dont get timeout messages the various wwidmgr -show commands.
But the main issue seems to be to connect to the EMC box. That is essential to configure the actual disk devices. The way I try to get it is that, like a Mylex 960 controller, you must set up devices: RAID or JBOD. No devices on the controller none can be presented to VMS.
Back to the EMC2 documentation.
One side question: I have no FC switch so the HBA is connected directly to the SAN controller. Is this by definition an arbitrated loop configuration?
(which is not supported by VMS, but it might work?)
Hans
Stephen Hoffman
Jun 28, 2014, 3:20:21 PM6/28/14
to
On 2014-06-28 12:30:21 +0000, Hans Vlems said:
> One side question: I have no FC switch so the HBA is connected directly
> to the SAN controller. Is this by definition an arbitrated loop
> configuration?
> (which is not supported by VMS, but it might work?)
You need a switch.
> One side question: I have no FC switch so the HBA is connected directly
> to the SAN controller. Is this by definition an arbitrated loop
> configuration?
> (which is not supported by VMS, but it might work?)
Fibre Channel Arbitrated Loop (FC-AL) configurations are not supported
by OpenVMS, and are not expected to work with OpenVMS.
If you want to experiment with OpenVMS and Fibre Channel storage, then
visit eBay or otherwise and get a switch.
Or get a FC SAN storage server that has an embedded switch, as some do offer.
Hans Vlems
Jun 28, 2014, 3:28:36 PM6/28/14
to
Op zaterdag 28 juni 2014 21:20:21 UTC+2 schreef Stephen Hoffman:
OK. Any suggestions, or will any 2 Gb/s FC switch do?
Stephen Hoffman
Jun 28, 2014, 4:04:02 PM6/28/14
to
On 2014-06-28 19:28:36 +0000, Hans Vlems said:
> OK. Any suggestions, or will any 2 Gb/s FC switch do?
Preferably one that's cheap and supported by OpenVMS, but one that's
> OK. Any suggestions, or will any 2 Gb/s FC switch do?
cheap and feature-simplistic will probably work fine, too.
Some of the older supported options (which'll also give you some search
keywords) are DSGGA-AA / -AB and DSGGB-AA / -AB
You're probably not looking for and probably don't want fancy features
and lots of ports here, after all.
VMS support hasn't changed all that much in the last few years, and the
order gear will probably be cheaper in any case.
For an overview of FC SANs and OpenVMS and clustering:
<http://h71000.www7.hp.com/doc/82final/6318/6318pro_009.html>
Also see the "HP StorageWorks SAN Design Reference Guide" for a rather
more gonzo view of the FC SAN hardware. I don't see a copy of that
document available at HP (probably got renamed or retired?), but
there's a copy here:
<http://community.brocade.com/dtscp75322/attachments/dtscp75322/fibre/13967/1/SDG_FullBook_AA-RWPYE-TE_2009_11_13.pdf>
There was once a fibre channel page
<http://h71000.www7.hp.com/openvms/fibre/index.html> on the VMS web
site, but that page is gone.
Here's the old you-need-a-switch answer:
<http://h71000.www7.hp.com/wizard/wiz_4552.html>
There's also Colin's quite-useful wall-of-text reply, too.
You're also probably going to have some "fun" figuring out how to
configure that storage controller, particularly if it's not one that
doesn't already include details of configuring it for use with OpenVMS.
For just a couple of boxes, I'd personally scrounge a TQE-capable
SCSI controller and a shared bus, and skip the FC SAN.
Hans Vlems
Jun 29, 2014, 11:29:49 AM6/29/14
to
I'm a hobbyist Hoff and the breed is known and feared for entertaining strange ideas and conducting dubious experiments. Remember the shared SCSI VAXcluster experiment? That was me :)
That said, I bought the EMC2 FC SAN without knowing anything at all about FC technology. All I needed was a controller, right?
"FC ain't SCSI" I should write that down sometime....
It'd be a waste of time (and some money) to stop here. So a switch will be bought.
But the main problem right now is getting EMC2's Navisphere software, it's a prerequisite tp configure the FC disks.
Thanks so far for all the help!
Hans
That said, I bought the EMC2 FC SAN without knowing anything at all about FC technology. All I needed was a controller, right?
"FC ain't SCSI" I should write that down sometime....
It'd be a waste of time (and some money) to stop here. So a switch will be bought.
But the main problem right now is getting EMC2's Navisphere software, it's a prerequisite tp configure the FC disks.
Thanks so far for all the help!
Hans
Stephen Hoffman
Jun 29, 2014, 12:06:20 PM6/29/14
to
On 2014-06-29 15:29:49 +0000, Hans Vlems said:
> I'm a hobbyist Hoff and the breed is known and feared for entertaining
> strange ideas and conducting dubious experiments.
Then consider ending the quest for certainly and support here. You
> I'm a hobbyist Hoff and the breed is known and feared for entertaining
> strange ideas and conducting dubious experiments.
have reached the final Fibre Channel SAN frontier, and can boldly go
where no hobbyist has gone before. You might find that FC-AL does even
work for this particular case, too. Though do plan to lose a few
"red shirts" in the process of your various quests — you might possibly
and occasionally fry some equipment. Probably not here, but it's well
worth budgeting for the occasional Enterprise Hardware Incident.
Colin Butcher
Jun 29, 2014, 2:14:58 PM6/29/14
to
No switch = loop mode, with which VMS does not work. You must have a switch
and be in fabric mode.
Booting is a special case. Some storage arrays work with VMS as data
devices, but not as boot devices. It's all in the firmware. Your EMC array
may be usable as a data device, but not as a boot device.
On Alpha you have WWIDMGR to set boot paths, adapter modes etc.
On IA64 you have to get into each adapter in turn from EFI Shell, then set
up each adapter. You can also scan the fibre to find discs with bootable
partitions from EFI Shell, but it's vastly easier to boot a minimal system
from USB or local disc, then use @sys$manager:boot_options.com to set the
boot paths.
The firmware has to know what to do on boot, so FC path information for the
boot device paths gets stashed in the adapter NVRAM by WWIDMGR (on Alpha) or
by BOOT_OPTIONS (or EFI Shell and adapter menu) on IA64.
You might want to get an old HSG80 and a couple of the 8 port 2Gbps FC
switches. It'll probably work with a whole lot less pain than trying to get
the EMC array to work with VMS - and you'll know that the FC stuff is good
before you try to figure out the EMC specific bits.
Have fun!
and be in fabric mode.
Booting is a special case. Some storage arrays work with VMS as data
devices, but not as boot devices. It's all in the firmware. Your EMC array
may be usable as a data device, but not as a boot device.
On Alpha you have WWIDMGR to set boot paths, adapter modes etc.
On IA64 you have to get into each adapter in turn from EFI Shell, then set
up each adapter. You can also scan the fibre to find discs with bootable
partitions from EFI Shell, but it's vastly easier to boot a minimal system
from USB or local disc, then use @sys$manager:boot_options.com to set the
boot paths.
The firmware has to know what to do on boot, so FC path information for the
boot device paths gets stashed in the adapter NVRAM by WWIDMGR (on Alpha) or
by BOOT_OPTIONS (or EFI Shell and adapter menu) on IA64.
You might want to get an old HSG80 and a couple of the 8 port 2Gbps FC
switches. It'll probably work with a whole lot less pain than trying to get
the EMC array to work with VMS - and you'll know that the FC stuff is good
before you try to figure out the EMC specific bits.
Have fun!
Jan-Erik Soderholm
Jun 29, 2014, 4:57:51 PM6/29/14
to
Colin Butcher wrote 2014-06-29 20:14:
> No switch = loop mode, with which VMS does not work. You must have a switch
> and be in fabric mode.
>
> Booting is a special case. Some storage arrays work with VMS as data
> devices, but not as boot devices. It's all in the firmware. Your EMC array
> may be usable as a data device, but not as a boot device.
>
> On Alpha you have WWIDMGR to set boot paths, adapter modes etc.
>
> On IA64 you have to get into each adapter in turn from EFI Shell, then set
> up each adapter. You can also scan the fibre to find discs with bootable
> partitions from EFI Shell, but it's vastly easier to boot a minimal system
> from USB or local disc, then use @sys$manager:boot_options.com to set the
> boot paths.
>
> The firmware has to know what to do on boot, so FC path information for the
> boot device paths gets stashed in the adapter NVRAM by WWIDMGR (on Alpha) or
> by BOOT_OPTIONS (or EFI Shell and adapter menu) on IA64.
>
> You might want to get an old HSG80 and a couple of the 8 port 2Gbps FC
> switches...
> No switch = loop mode, with which VMS does not work. You must have a switch
> and be in fabric mode.
>
> Booting is a special case. Some storage arrays work with VMS as data
> devices, but not as boot devices. It's all in the firmware. Your EMC array
> may be usable as a data device, but not as a boot device.
>
> On Alpha you have WWIDMGR to set boot paths, adapter modes etc.
>
> On IA64 you have to get into each adapter in turn from EFI Shell, then set
> up each adapter. You can also scan the fibre to find discs with bootable
> partitions from EFI Shell, but it's vastly easier to boot a minimal system
> from USB or local disc, then use @sys$manager:boot_options.com to set the
> boot paths.
>
> The firmware has to know what to do on boot, so FC path information for the
> boot device paths gets stashed in the adapter NVRAM by WWIDMGR (on Alpha) or
> by BOOT_OPTIONS (or EFI Shell and adapter menu) on IA64.
>
> You might want to get an old HSG80 and a couple of the 8 port 2Gbps FC
Probably also available in the Netherlands, but my customer decommisioned
a HSG80 system complete with 4 shelfs with 9 and 18 GB disks. There are
also complete FC net with DEC FC (1Gbps) switches and AlphaServer HBA's.
Our local brooker didn't even want the stuff in return for free, so I
think the plan is/was to simply trash it all.
There was also a HSZ80 shelf with the same disks. The two HSx80
systems was shadowed using VMS HBVS.
Anyway, if there is any interest in these thinhs, I could ask on Monday
and see what happend with the stuff.
Jan-Erik.
John E. Malmberg
Jun 29, 2014, 7:25:00 PM6/29/14
to
On 6/29/2014 1:14 PM, Colin Butcher wrote:
> No switch = loop mode, with which VMS does not work. You must have a switch
> and be in fabric mode.
Loop mode is unsupported on VMS. No storage arrays configurations with
> No switch = loop mode, with which VMS does not work. You must have a switch
> and be in fabric mode.
loop mode were ever officially qualified.
Which means if you want to use it at your own risk, it is totally on
you. But make sure you have good backups and up to date firmware on
your storage devices.
Regards,
-John
wb8...@qsl.network
Personal Opinion Only
Paul Sture
Jun 30, 2014, 2:34:02 AM6/30/14
to
On 2014-06-29, Hans Vlems <hvl...@freenet.de> wrote:
> I'm a hobbyist Hoff and the breed is known and feared for entertaining
> strange ideas and conducting dubious experiments.
LOL!
> I'm a hobbyist Hoff and the breed is known and feared for entertaining
> strange ideas and conducting dubious experiments.
And very well put.
John E. Malmberg
Jun 30, 2014, 8:39:27 AM6/30/14
to
On 6/29/2014 10:29 AM, Hans Vlems wrote:
> I'm a hobbyist Hoff and the breed is known and feared for
> entertaining strange ideas and conducting dubious experiments.
> Remember the shared SCSI VAXcluster experiment? That was me :)
> That said, I bought the EMC2 FC SAN without knowing anything at all
> about FC technology. All I needed was a controller, right?
> "FC ain't SCSI" I should write that down sometime.... It'd be a waste
> of time (and some money) to stop here. So a switch will be bought.
FC-AL is implemented on VMS, just not supported because no storage array
> I'm a hobbyist Hoff and the breed is known and feared for
> entertaining strange ideas and conducting dubious experiments.
> Remember the shared SCSI VAXcluster experiment? That was me :)
> That said, I bought the EMC2 FC SAN without knowing anything at all
> about FC technology. All I needed was a controller, right?
> "FC ain't SCSI" I should write that down sometime.... It'd be a waste
> of time (and some money) to stop here. So a switch will be bought.
was qualified.
As no devices were qualified, there may be unknown dragons lurking.
This is the same case for using any third-party storage on VMS that has
not been qualified by HP for this use.
The difference between FC-AL and a FC Switch is that it is easier for a
small problem to introduce packet corruption on a FC-AL configuration,
and if there is a bug in handling corrupt packets anywhere, then you
will get data corruption on the device.
Think of it like the difference between a coaxial Ethernet and a modern
switch.
> But the main problem right now is getting EMC2's Navisphere
> software, it's a prerequisite tp configure the FC disks.
requirements that VMS has for Fibre Channel storage and possible
solutions using Ubuntu.
Subject: "Re: Any Fibre Channel hackers out there?"
Hans Vlems
Jun 30, 2014, 10:19:26 AM6/30/14
to
Op zondag 29 juni 2014 18:06:20 UTC+2 schreef Stephen Hoffman:
What's "losing a red shirt" mean?
Hans
> On 2014-06-29 15:29:49 +0000, Hans Vlems said:
>
>
>
> > I'm a hobbyist Hoff and the breed is known and feared for entertaining
>
> > strange ideas and conducting dubious experiments.
>
>
>
> Then consider ending the quest for certainly and support here. You
>
> have reached the final Fibre Channel SAN frontier, and can boldly go
>
> where no hobbyist has gone before. You might find that FC-AL does even
>
> work for this particular case, too. Though do plan to lose a few
>
> "red shirts" in the process of your various quests -- you might possibly
>
>
>
> > I'm a hobbyist Hoff and the breed is known and feared for entertaining
>
> > strange ideas and conducting dubious experiments.
>
>
>
> Then consider ending the quest for certainly and support here. You
>
> have reached the final Fibre Channel SAN frontier, and can boldly go
>
> where no hobbyist has gone before. You might find that FC-AL does even
>
> work for this particular case, too. Though do plan to lose a few
>
>
> and occasionally fry some equipment. Probably not here, but it's well
>
> worth budgeting for the occasional Enterprise Hardware Incident.
>
>
>
> --
>
> Pure Personal Opinion | HoffmanLabs LLC
Crossing the support boundary was intended :)
> and occasionally fry some equipment. Probably not here, but it's well
>
> worth budgeting for the occasional Enterprise Hardware Incident.
>
>
>
> --
>
> Pure Personal Opinion | HoffmanLabs LLC
What's "losing a red shirt" mean?
Hans
Hans Vlems
Jun 30, 2014, 10:41:53 AM6/30/14
to
Op zondag 29 juni 2014 20:14:58 UTC+2 schreef Colin Butcher:
Fun it sure is Colin, and good fun at that, but I'm unwilling to try the really impossible.
I use a DS10 with a KGPSA-CA. VMS recognizes the HBA and both the PGA0 and FGA0 device drivers get loaded and report no errors.
wwidmgr on the DS10 talks to the EMC2 FC SAN controller, at least I think so.
wwidmgr -show adapter and wwidmgr -set adapter work and changed settings remain changed.
In FABRIC mode the -show wwid command didn't return anything except a timeout error.
In LOOP mode the SAN returns one entry though with a udid value of -1.
At that point >>> SHO DEV also shows a DGA disk device. However, an init is
required after wwidmgr was run and that destroys the DGA disk....
What I need is Navisphere from EMC2 !!!!!!!!!
Hans
I use a DS10 with a KGPSA-CA. VMS recognizes the HBA and both the PGA0 and FGA0 device drivers get loaded and report no errors.
wwidmgr on the DS10 talks to the EMC2 FC SAN controller, at least I think so.
wwidmgr -show adapter and wwidmgr -set adapter work and changed settings remain changed.
In FABRIC mode the -show wwid command didn't return anything except a timeout error.
In LOOP mode the SAN returns one entry though with a udid value of -1.
At that point >>> SHO DEV also shows a DGA disk device. However, an init is
required after wwidmgr was run and that destroys the DGA disk....
What I need is Navisphere from EMC2 !!!!!!!!!
Hans
Hans Vlems
Jun 30, 2014, 10:45:03 AM6/30/14
to
Op zondag 29 juni 2014 22:57:51 UTC+2 schreef Jan-Erik Soderholm:
I'm interesetd but shipping is going to spoil this. I'm willing to travel by car.
On one day it is possible to drive about 1200 kms. So Sweden is too far away :-(
On one day it is possible to drive about 1200 kms. So Sweden is too far away :-(
Hans Vlems
Jun 30, 2014, 10:53:30 AM6/30/14
to
Op maandag 30 juni 2014 01:25:00 UTC+2 schreef John E. Malmberg:
Unsupported does not scare me away John. I run a two node NI VAXcluster each with its own locally attached quorum disk. That's not just unsupported but also deemed very dangerous.
So all I'm trying to do is to try and make it work. It's not supposed to be used in a production environment.
If it won't work, I'll get myself an HSG80 and put the disks in that.
Hans
So all I'm trying to do is to try and make it work. It's not supposed to be used in a production environment.
If it won't work, I'll get myself an HSG80 and put the disks in that.
Hans
Stephen Hoffman
Jun 30, 2014, 11:29:16 AM6/30/14
to
On 2014-06-30 14:19:26 +0000, Hans Vlems said:
> What's "losing a red shirt" mean?
http://en.wikipedia.org/wiki/Redshirt_(character)
> What's "losing a red shirt" mean?
Jan-Erik Soderholm
Jun 30, 2014, 12:28:58 PM6/30/14
to
The customer site is close to J�nk�ping in the south of Sweden.
J�nk�ping/Amsterdam is 1114 Kms according to Google Earth.
Including the ferry R�dby/Puttgarden...
Let me know if I should ask them... :-)
Jan-Erik.
Simon Clubley
Jun 30, 2014, 1:09:46 PM6/30/14
to
On 2014-06-30, Hans Vlems <hvl...@freenet.de> wrote:
> Op zondag 29 juni 2014 22:57:51 UTC+2 schreef Jan-Erik Soderholm:
>>
> Op zondag 29 juni 2014 22:57:51 UTC+2 schreef Jan-Erik Soderholm:
>>
>> Probably also available in the Netherlands, but my customer decommisioned
>> a HSG80 system complete with 4 shelfs with 9 and 18 GB disks. There are
>> also complete FC net with DEC FC (1Gbps) switches and AlphaServer HBA's.
>>
>> Our local brooker didn't even want the stuff in return for free, so I
>> think the plan is/was to simply trash it all.
>>
There was a time not too many years ago when people would be scrambling
>> a HSG80 system complete with 4 shelfs with 9 and 18 GB disks. There are
>> also complete FC net with DEC FC (1Gbps) switches and AlphaServer HBA's.
>>
>> Our local brooker didn't even want the stuff in return for free, so I
>> think the plan is/was to simply trash it all.
>>
for that equipment. The times are indeed changing.
I'm surprised brokers don't want the equipment in order to sell on to
other DEC customers. Has the DEC market really collapsed to that extent ?
>> There was also a HSZ80 shelf with the same disks. The two HSx80
>> systems was shadowed using VMS HBVS.
>>
>> Anyway, if there is any interest in these thinhs, I could ask on Monday
>> and see what happend with the stuff.
>>
>
> I'm interesetd but shipping is going to spoil this. I'm willing to travel by car.
> On one day it is possible to drive about 1200 kms. So Sweden is too far away :-(
As an aside, it's interesting how technology has redefined what "far away"
> On one day it is possible to drive about 1200 kms. So Sweden is too far away :-(
means. My personal maximum walking distance is about 25 miles (40 km) in
a day so that one day car journey limit would take me about a month.
OTOH, a civilian aircraft can cover, _well_ within a day, the distance
it would take a car a week to travel.
Simon.
--
Simon Clubley, clubley@remove_me.eisner.decus.org-Earth.UFP
Microsoft: Bringing you 1980s technology to a 21st century world
David Froble
Jun 30, 2014, 1:22:58 PM6/30/14
to
Hans Vlems wrote:
> I'm interesetd but shipping is going to spoil this. I'm willing to travel by car.
> On one day it is possible to drive about 1200 kms. So Sweden is too far away :-(
For what you guys are paying for gas, I've got to question the wisdom of
> I'm interesetd but shipping is going to spoil this. I'm willing to travel by car.
> On one day it is possible to drive about 1200 kms. So Sweden is too far away :-(
being anti-shipping. That is, unless shipping there has gotten even
worse than here, which is totally crazy.
Jan-Erik Soderholm
Jun 30, 2014, 1:35:35 PM6/30/14
to
Simon Clubley wrote 2014-06-30 19:09:
> On 2014-06-30, Hans Vlems <hvl...@freenet.de> wrote:
>> Op zondag 29 juni 2014 22:57:51 UTC+2 schreef Jan-Erik Soderholm:
>>>
>>> Probably also available in the Netherlands, but my customer decommisioned
>>> a HSG80 system complete with 4 shelfs with 9 and 18 GB disks. There are
>>> also complete FC net with DEC FC (1Gbps) switches and AlphaServer HBA's.
>>>
>>> Our local brooker didn't even want the stuff in return for free, so I
>>> think the plan is/was to simply trash it all.
>>>
>
> There was a time not too many years ago when people would be scrambling
> for that equipment. The times are indeed changing.
>
> I'm surprised brokers don't want the equipment in order to sell on to
> other DEC customers. Has the DEC market really collapsed to that extent ?
>
Well, he has several stores with a lot of used DEC/CPQ/HP
> On 2014-06-30, Hans Vlems <hvl...@freenet.de> wrote:
>> Op zondag 29 juni 2014 22:57:51 UTC+2 schreef Jan-Erik Soderholm:
>>>
>>> Probably also available in the Netherlands, but my customer decommisioned
>>> a HSG80 system complete with 4 shelfs with 9 and 18 GB disks. There are
>>> also complete FC net with DEC FC (1Gbps) switches and AlphaServer HBA's.
>>>
>>> Our local brooker didn't even want the stuff in return for free, so I
>>> think the plan is/was to simply trash it all.
>>>
>
> There was a time not too many years ago when people would be scrambling
> for that equipment. The times are indeed changing.
>
> I'm surprised brokers don't want the equipment in order to sell on to
> other DEC customers. Has the DEC market really collapsed to that extent ?
>
stuff incl VT-screens and parts for PDP/VAX/Alpha systems. So
he simply alrady had enough of StorageWorks HSZ/HSG stuff.
I left a MSA1000 shelf with fiber controller to the recycling
a few weeks ago. I also asked the broker at that time, and he
had plenty of those things also... :-)
According to him, the market (at least for VMS compatible stuff)
has kind of collapsed. A month ago he was to a large producer of
stock exchange software and fetched 30+ rx2660 boxes, the major
part of the VMS development environment. Only maintenance and
support now. If he could resell 2 or 3 he would break even... :-)
Well, times are changing.
Jan-Erik.
Alessandro Mazzini
Jun 30, 2014, 3:02:04 PM6/30/14
to
>What I need is Navisphere from EMC2 !!!!!!!!!
>Hans
John H. Reinhardt
Jun 30, 2014, 5:58:14 PM6/30/14
to
On 2014-06-30 13:35:35 -0400, Jan-Erik Soderholm said:
> Simon Clubley wrote 2014-06-30 19:09:
>> I'm surprised brokers don't want the equipment in order to sell on to
>> other DEC customers. Has the DEC market really collapsed to that extent ?
>>
>
> Well, he has several stores with a lot of used DEC/CPQ/HP
> stuff incl VT-screens and parts for PDP/VAX/Alpha systems. So
> he simply alrady had enough of StorageWorks HSZ/HSG stuff.
>
> I left a MSA1000 shelf with fiber controller to the recycling
> a few weeks ago. I also asked the broker at that time, and he
> had plenty of those things also... :-)
>
I routinely see MSA1500cs controllers on the US Ebay for less than $200
> Simon Clubley wrote 2014-06-30 19:09:
>> I'm surprised brokers don't want the equipment in order to sell on to
>> other DEC customers. Has the DEC market really collapsed to that extent ?
>>
>
> Well, he has several stores with a lot of used DEC/CPQ/HP
> stuff incl VT-screens and parts for PDP/VAX/Alpha systems. So
> he simply alrady had enough of StorageWorks HSZ/HSG stuff.
>
> I left a MSA1000 shelf with fiber controller to the recycling
> a few weeks ago. I also asked the broker at that time, and he
> had plenty of those things also... :-)
>
in a Buy It Now listing. They aren't selling at the price. 3 years
ago I bought one with a dual controller, 256MB cache and 3 SCSI/SATA
I/O ports for around $500 and another that came as a package with a
MSA20 with 12 750GB SATA drives for $700. The MSA20 and 750GB drives
are still going for around $300 but the MSA's are tough to sell.
I also have a Compaq 2200 storage shelf with dual HSG80 controllers and
3 4354/4254 shelves and 36Gb, 18Gb and 9Gb drives. Plus a DEC RA8000
with dual HSG80's and 24 storage bricks. They are nice units. The
problem is capacity and usage. What I want is the capability to
create a large RAID storage pool and carve smaller logical disks from
them like an EVA or an EMC Clarion type system. And, of course, I need
it in a small space. I'm trying to fit it into about 4u-6u in a
storage rack and the HSG's just don't do that. Hence my other thread
about the Linux FC target software (which I haven't made any progress
on yet). I know i'm being picky, but that's what I have room for now.
I'm having to consolidate my server footprint. The MSA1500 series does
allow me to use smaller drives and get close to the size I need, but as
I've commented other places, the CLI and the capabilities for creating
LUNs is somewhat barbaric.
John H. Reinhardt
John H. Reinhardt
Jun 30, 2014, 6:02:04 PM6/30/14
to
that John probably knows his stuff about the OpenVMS FC capabilites.
His initials are strewn around the comments enough in
"IOGEN-FIBRE-CONFIG.LIS" and others enough.
BTW, thanks John for your hints. I'm still working on decyphering
varios things in the OpenVMS and the SCST code to see if I can get them
to talk to each other.
--
John H. Reinhardt
Hans Vlems
Jul 1, 2014, 4:47:00 AM7/1/14
to
Op maandag 30 juni 2014 19:22:58 UTC+2 schreef David Froble:
Well, I happen to like driving a car. Diesel is EURO 1.40 per liter and the car drives 17 kms per liter. Not sure whether that is cheaper than DHL but for short distances (anything under 500 kms) I prefer to go there myself. I've collected all kinds of nice surprises just by actually showing up.
I'd love to do a trip to Sweden but it's just (a ferry) too far.
Hans
I'd love to do a trip to Sweden but it's just (a ferry) too far.
Hans
Hans Vlems
Jul 1, 2014, 5:06:40 AM7/1/14
to
Yes, John Malmberg's post was very informative. Which is why I went the LOOP path, knowing it is technically unsafe. But I'm not running a production environment, just trying to figure out how things work. I once connected two VAX systems on one shared SCSI-2 bus with a shared SCSI disk. Performance was awful, access was not shared but nothing broke and data was not corrupted.
I bought the EMC2 cabinet without FC knowledge. So I'm trying to figure things out. I can't access the EMC2 FC SAN yet but Alessandro was a great help in getting the tools.
My plan is this: configure the EMC2 as simple as possible. What I'm trying to achieve is this: Each physical disk gets its own udid and in turn each udid wil map to a device identifier and VMS will see a DGA disk with that identifier.
If VMS recognizes these device then I'll change to FABRIC mode and repeat the process.
Why do I start with LOOP? Well I have no switch (yet) and wwidmgr returns a timeout error message on the adapter (PGA0 or FGA0, forgot which) in FABRIC mode. In LOOP mode at least I get responses to wwidmgr commands.
My "if all fails" plan is to get me an HSG80 cabinet and put the FC disks in there.
Hans
I bought the EMC2 cabinet without FC knowledge. So I'm trying to figure things out. I can't access the EMC2 FC SAN yet but Alessandro was a great help in getting the tools.
My plan is this: configure the EMC2 as simple as possible. What I'm trying to achieve is this: Each physical disk gets its own udid and in turn each udid wil map to a device identifier and VMS will see a DGA disk with that identifier.
If VMS recognizes these device then I'll change to FABRIC mode and repeat the process.
Why do I start with LOOP? Well I have no switch (yet) and wwidmgr returns a timeout error message on the adapter (PGA0 or FGA0, forgot which) in FABRIC mode. In LOOP mode at least I get responses to wwidmgr commands.
My "if all fails" plan is to get me an HSG80 cabinet and put the FC disks in there.
Hans
Phillip Helbig---undress to reply
Jul 1, 2014, 5:13:35 AM7/1/14
to
In article <7bb06e6e-bd2c-412c...@googlegroups.com>, Hans
number of liters needed for a fiducial distance. In Europe, fuel
efficiency is usually expressed in the number of liters needed for 100
km. (Divide 240 by this number to get, approximately, miles per gallon;
divide 240 by mpg to get liters for 100 km.) Yes, fuel is more
expensive here, but generally cars are more economical.
> Well, I happen to like driving a car. Diesel is EURO 1.40 per liter and the
The price varies with time and with country and from place to place
within a country, but EUR 1.40 per liter is a good estimate (though I
paid EUR 1.30 recently, about $3.90 per gallon). If I drive 100 km/h
(not much more is allowed anywhere but on parts of the Autobahn in
Germany), I use a bit less than 4 liters per 100 km (about 60 mpg).
This is a 4-door estate (station wagon). (Although it takes a bit to
get there, it will go 190 km/h (about 120 mph), which is all I really
need.)
> I've collected all kinds of nice surprises just by actually showing up.
And I still appreciate the hardware you brought me a while back!
Vlems <hvl...@freenet.de> writes:
> > For what you guys are paying for gas, I've got to question the wisdom of
> > being anti-shipping. That is, unless shipping there has gotten even
> > worse than here, which is totally crazy.
Well, the cost is the distance times the price per liter times the
> > For what you guys are paying for gas, I've got to question the wisdom of
> > being anti-shipping. That is, unless shipping there has gotten even
> > worse than here, which is totally crazy.
number of liters needed for a fiducial distance. In Europe, fuel
efficiency is usually expressed in the number of liters needed for 100
km. (Divide 240 by this number to get, approximately, miles per gallon;
divide 240 by mpg to get liters for 100 km.) Yes, fuel is more
expensive here, but generally cars are more economical.
> Well, I happen to like driving a car. Diesel is EURO 1.40 per liter and the
within a country, but EUR 1.40 per liter is a good estimate (though I
paid EUR 1.30 recently, about $3.90 per gallon). If I drive 100 km/h
(not much more is allowed anywhere but on parts of the Autobahn in
Germany), I use a bit less than 4 liters per 100 km (about 60 mpg).
This is a 4-door estate (station wagon). (Although it takes a bit to
get there, it will go 190 km/h (about 120 mph), which is all I really
need.)
> I've collected all kinds of nice surprises just by actually showing up.
Hans Vlems
Jul 1, 2014, 5:14:38 AM7/1/14
to
Op maandag 30 juni 2014 18:28:58 UTC+2 schreef Jan-Erik Soderholm:
that won't work. I know the ferry because DECUS Europe was in Kopenhagen once and I got there by car. Nice trip and easily done in one day.
I've been planning a trip to the Danmark but since it needs to be a return trip, even Danmark is too faar away. So unless an empty cabinet and/or FC switch is not too expensive and shipping is affordable (< €50) it goes in the bin..
Hans
> Hans Vlems wrote 2014-06-30 16:45:
>
> > Op zondag 29 juni 2014 22:57:51 UTC+2 schreef Jan-Erik Soderholm:
>
> > Op zondag 29 juni 2014 22:57:51 UTC+2 schreef Jan-Erik Soderholm:
> > > On one day it is possible to drive about 1200 kms. So Sweden is too far away :-(
>
> >
>
>
>
> Well, Sweden is 2500 Kms long in itself... :-)
>
> The customer site is close to J�nk�ping in the south of Sweden.
>
> J�nk�ping/Amsterdam is 1114 Kms according to Google Earth.
>
> Including the ferry R�dby/Puttgarden...
>
>
>
> Let me know if I should ask them... :-)
>
>
>
> Jan-Erik.
Jan-Erik,
>
> >
>
>
>
> Well, Sweden is 2500 Kms long in itself... :-)
>
> The customer site is close to J�nk�ping in the south of Sweden.
>
> J�nk�ping/Amsterdam is 1114 Kms according to Google Earth.
>
> Including the ferry R�dby/Puttgarden...
>
>
>
> Let me know if I should ask them... :-)
>
>
>
> Jan-Erik.
that won't work. I know the ferry because DECUS Europe was in Kopenhagen once and I got there by car. Nice trip and easily done in one day.
I've been planning a trip to the Danmark but since it needs to be a return trip, even Danmark is too faar away. So unless an empty cabinet and/or FC switch is not too expensive and shipping is affordable (< €50) it goes in the bin..
Hans
0 new messages