We are just running a regular Windows share.
I installed Spark on a workstation, then I copied over the \Program Files\Spark directory (minus the uninstall file in the root) over to our servers shared folder.
We have a registration program that automatically creates users in Wildfire, puts an entry into the Run key in HKCU (in the form of
server\share\Spark.exe …) and then finally launches Spark from the server share.
All users run Spark from the Windows share, and it works great. Plus upgrades are a snap (only one place to upgrade for 200+ users).
Hope this helps.
All are XP SP2.
That would explain why it works. It’‘s not working in Windows 2000 for any user that doesn’'t actually install Spark (even admins unless they install it under their credentials.) I feel like it has to be something done in the installation that is happening on a per-user instead of a machine wide basis (HKCU instead of HKLM.)
Jimmy Martin
Okay, it looks like I’'ve guessed correctly. If you create a registry file with the following registry entries and merge it into the registry, Spark works for any user you merge these settings for. So the real questions are: Which registry entry is it that keeps it from working and can this be changed from HKCU to HKLM?
“InstallStarted”=dword:00000000
[HKEY_CURRENT_USER\Software\ej-technologies\exe4j\jvms\c:/program
files/java/jre1.5.0_06/bin/java.exe]
“LastWriteTime”=hex:00,ff,a9,ff,e0,ff,96,13,ff,e6,ff,c5,01
“Version”=“1.5.0_06”
“@DeskTop
spark_~1.exe”=dword:000002e8
“Spark”="@ProgFiles
Spark
Spark.exe"
are the JavaSoft keys needed?
And there’'s also VMware keys?
We are just running a regular Windows share.
I installed Spark on a workstation, then I copied over the \Program Files\Spark directory (minus the uninstall file in the root) over to our servers shared folder.
Do you grant write access to this share or will spark run ok if it’'s read-only.
Plus upgrades are a snap (only one place to upgrade for 200+ users).
So do you repeat the process for an upgrade? (Install on a workstation - copy files out to share)
I’‘m updating this post…I found the ONE registry key that is required to make the client work. This one has me scratching my head as to the cause, but if you add this registry key to the user, it works fine. Rights aren’‘t an issue, just this key (and it’‘s the one I figured didn’'t matter at all.)
“Spark”=“C:
Program Files
Spark
Spark.exe”
Go figure.
The funny thing is, adding the key resolves the problem but removing the key doesn’‘t recreate the problem. It’‘s like the key has to be there for one run of Spark and then it’'s not needed any more in W2K.
Jimmy Martin
Message was edited by: jimmym
Do you grant write access to this share or will spark
run ok if it’'s read-only.
We didn’'t set the read only attribute, we rely on NTFS permissions. But it should work read only, except for the log files.
So do you repeat the process for an upgrade? (Install
on a workstation - copy files out to share)
Yep.
We didn’'t set the read only attribute, we rely on
NTFS permissions. But it should work read only,
except for the log files.
What NTFS permissions did you give to this share?
Message was edited by: dsteven
What NTFS permissions did you give to this share?
The AD group that all users are a member of was assigned Read & Execute.
.I found the ONE registry key that is required to make the client work.
Can confirm that this works for me also…
That entry makes Spark start on login, right?
.I found the ONE registry key that is required to
make the client work.
Can confirm that this works for me also…
That entry makes Spark start on login, right?
That’‘s the only use that registry key has in Windows. I’'m not sure why Spark looks for it the first time, but if you run Spark once with that entry and then delete the registry entry, Spark seems to work fine every time thereafter.
At least I’'ve got a fix.
Jimmy