Installing Maglev 1.1 using rbenv

[Patru]

Hi all

It just took longer than it should to install Maglev 1.1 using rbenv so I thought I would share this.

I installed rbenv and ruby-build which gives me a nice list with

•   maglev-1.0.0
  
•   maglev-1.1.0-dev
  
•   maglev-2.0.0-dev
  

which is only slightly missleading, as maglev-2.0.0-dev will acutally install maglev-1.2.0-dev (which I take as a Good Thing (tm) :-). 

However as the current build of 1.2.0 did not seem very reliable last week (it even failed *sometimes* on a few trivial benchmarks I am running) I actually wanted to install 1.1.0 today for another account as that one seems to run fairly reliably (though still with 1.8 syntax).

As it should be `rbenv install` knows even how to autocomplete maglev-1.1.0-dev for me and it hops along. There do not appear to be any troubles, but your `maglev start` will just not ... start. It seems to be missing its Gemstone which is linked to

~/.rbenv/versions/GemStone-29699.Darwin-i386

which seems fine at first sight. However, that was just a *file* and not a *directory* as it should have been. So I looked for the file myself and found it in 

http://seaside.gemtalksystems.com/maglev/GemStone-29699.Darwin-i386.tar.gz

which did not quite match the 

http://glass-downloads.gemstone.com/maglev/$gss_file

from update.sh. I guess the name change of the company is now reflected in the DNS, but unfortunately no new Maglev-packages have been built of the "old" versions. and they still reference the old site.

All is not lost however, you can download the GemStone manually from the above link and put it in ~/.rbenv/versions/, note however that your friendly Macintosh might have inflated the .gz already (mine had in any case). After re-gzipping the tar you can run update.sh and it will generate a perfect stone as it always used to (as it is nice enough not to download something that is already there).

I guess this will be fixed soon, but in the meantime you can circumvent the problems like this if you need to reinstall Maglev anywhere.

Greetings

:: Patru ""

[Matthias Springer]

Hi Patru,

can you describe shortly what kind of benchmarks you ran and what failed, so that we can narrow down the issue?

Matthias

--
You received this message because you are subscribed to the Google Groups "MagLev Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to maglev-discuss...@googlegroups.com.
To post to this group, send email to maglev-d...@googlegroups.com.
Visit this group at http://groups.google.com/group/maglev-discussion.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

-- 
Best regards,
Matthias Springer

Student B.Sc. IT-Systems Engineering (Hasso Plattner Institute)
+49-1577-3814255 * Feldweg 4, 85391 Unterkienberg, Germany

[Patru]

Oops, just noticed that I did not reply to this for an awfully long time.

I have tried to run the simple benchmarks in

https://github.com/Patru/persbench

Maglev 1.0.0 was pretty fast when saving a lot of objects in a single transaction, but it was very slow if I threw a lot of single object transactions at it without (obviously) maxing out any CPU resource. After a few crashes with maglev-1.2.0 I gave up on it. As described in my previous post I am currently using maglev-1.1.0

Greetings

:: Patru ""