Running stand alone diva v 0.7.0.2 Post_fixes on windows server. Using Hippo. When i try to hg other places e.g. germangrid.de:8002:Tamaland the address shows on the world map and ava arrives but after about 5 or 10 seconds i am logged out.
I have used the domain name 3dcolab.com as the external hostname in the region file and in all places in the myworld.ini. So as far as i can determine the ip address is not used in any place that I set up. the coords for the 4 regions are "6833,6176" ,etc
The first thing I noticed was that you are attempting to jump from
an OpenSim distribution which is using a hypergrid version different
from the intended destination. 0.7.0.2 uses HG 1.5i6 while
GermanGrid is running on 0.7.1 which uses HG 1.5i7. Actually, I'm
surprised you were even able to make contact with the Tamaland
region as most attempts simply return a "none found" in the map when
the origin and destination are using incompatible HG versions.
Second, in your log file, it seems that when you request the
GermanGrid location your server says "Region already exists in coordinates 28960 0" which leads
me to believe that you had previously linked that location to
your home map from the console (?). I could be wrong but I recall
similar log entries from my own server. If this is the case, it may
be that you are attempting to jump to regions that have since
upgraded their version of OpenSim. This would likely cause a
partial teleport with eventual fail as you would be using a link
which seems perfectly fine to you but no longer valid on
GermanGrid. It's like clicking an outdated bookmark in your web
browser and getting a "404" page in response.
If any of this makes sense or even applies to your situation, I hope
The HG version is controlled by the version of Opensim you're running, so changing Opensim versions is the only way to change your HG version - and yes, which means at the moment there are three incompatible versions (actually for a while I thought OSGrid was maybe a 4th? Or maybe they were just inaccessible to anyone.)
I frequently have the experience of Hypergridding to a new sim and actually seeing it rez before I crash, and sometimes repeating the process works and sometimes it doesn't. I never fully understand what the problem is, usually I just assume they are on a different HG version or something.
I've also never gotten the Hypergates to work on FleepGrid for some reason, not even to what appears to be other HG 1.5 (i6) grids. I know people can get TO FleepGrid from compatible grids, though, so I'm assuming my set up is ok.
I think generally HG is still in such early development that we can expect lots of changes and lack of backwards compatibility between versions is just the price we pay for advancement. Hooray for the bleeding edge. ;)
The security introduced in HG 1.5 makes the viewer crash if the
departing grid doesn't run 1.5. The receiving simulator performs some
checks that only exist in 1.5, and if they fail, the viewer crashes, per
On 4/23/2011 8:27 PM, Fleep Tuque wrote:
> Hi Paul,
> As far as I know, this is correct:
> 0.6.9 = hg 1.0
> 0.7.0.2 = hg 1.5i6
Between these two there was a complete redesign of the HG. They are
intentionally incompatible: 1.0 had very weak security.
> 0.7.0.2 = hg 1.5i6
> 0.7.1 = hg 1.5i7
Between these two, the incompatibility was unintentional, but
unavoidable. The culprit was actually libomv. There is a bug in the
libomv that is packaged in 0.7.0.2 and that was fixed some time later.
The fix makes TPs to older versions (that have the bug) fail. As such,
this incompatibility was unavoidable; we couldn't keep the bug in that
library for the sake of backwards compatibility in HG.
I don't like the naming i6/i7, because the HG interface didn't change.
There were some changes in the way appearance is packaged, but that, by
itself, would not make things incompatible. libomv's bug was the
dominant factor, and I could not get around the consequences of that bug
fix. But if those strings i6/i7 make people happy, I'm not going to
> From: Diva Canto <[hidden email]>
> I don't like the naming i6/i7, because the HG interface didn't change.
> There were some changes in the way appearance is packaged, but that, by
> itself, would not make things incompatible. libomv's bug was the
> dominant factor, and I could not get around the consequences of that bug
I agree.. i7 naming is not really right and may introduce more
confusion by using i7 when the interface is still actually 6 in the
migration/version tables. I have before suggested HG1.5/i6b just to
show its different, but I really think the HG number itself ought to
alter whenever a version of OpenSim is incompatible with an earlier
version even if it's not HG code or protocols themselves that alter.
For 0.7.1 would it make sense to bump the HG version number before
full release to 1.6?