Hi Tobias, Am 01.09.2012 um 10:58 schrieb Tobias Oetiker <[email protected]>:
> Today Dr. Jens Langner wrote:
>
>> As we had some private communication already I knew about your
>> move to ThinLinc. And to be honest, our organization (ca. 800
>> employees with about 200 sunrays) has already decided to move
>> away from the Sunray technology due to the bad licensing and
>> similar technical issues you report at your blog. In fact, I am
>> currently evaluating the Thinlinc technology and its potential.
>> Our first tests look quite promising, but we have to do some more
>> testing and financial calculations to judge if moving to ThinLinc
>> and a different Thinclient would really do good for us. In the
>> end we don't want to end at the same one way road which we are
>> currently in due to the disappointing political decisions Oracle
>> made with the otherwise stunning sunray technology.
>
> I agree, in that respect it would be reassuring if cendio were not
> obfuscating their python scripts ... at least the core technologies
> are opensource, and they do take our support calls very seriously.
Oh, they are obfuscating their python scripts on purpose? This seems to be a
rather bad approach as apart from the java and openldap stuff the sunray server
software can be nicely tuned by changing the tons of shell scripts they use.
However, if they just obfuscate their python script there should be a
deobfuscating tool around somewhere to get the scripts formatted nicely again
;) But I agree, Cendio should not go the same one way route like Oracle by
obfuscating/hiding things. Flexibility and freedom is very important for us. So
thanks for pointing that out to me so that I can have a look at that during my
evaluation of the ThinLinc software.
>> In our opinion, Oracle have to support more than their Solaris or
>> Oracle Linux, otherwise the Sunray platform has definitly no
>> future.
>
> indeed ... I still think the sunray 3plus is the best thinclinet HW
> available …
I fully agree, just the server software (SRS) seems to be badly maintained
IMHO. The hardware group seems to be way better than the software crew.
Unfortunately the best hardware is just as good as the software it uses. So
IMHO Oracle should spent a hell more resources into getting the sunray server
software uptodate. Besides badly missing OpenGL (VirtualGL) support, the
SunRay3 hardware can in principle utilize full USB2 speed (they even advertise
USB2 speed), but the sunray protocol seems to be not capable of transferring
faster than USB1.1 speed which nowadays is way too slow even for simply memory
stick operations. So if Oracle would really consider supporting all kind of
Linux derivates (which should be quite easy and straight forward) and if they
would adapt their licensing behavior to be more user friendly (separation of
hardware, software and even firmware licenses are non-sense) the SunRay
platform would probably be used by way more institutions like today. However, I
fear that changes like that are likely to never happen since it cannot be
expected that a former database/software only company understands the
requirements of application/hardware driven customers like us. They just want
to sell stuff in short term and I think as soon as the number of sunray
customers will drop to a certain level they will abandon the whole sunray
product line instead of analyzing why customers are moving to other solutions.
So if I would be a shareholder of Cendio&Co I would be quite happy to see the
value of these companies raising during the next years because of the
paralyzing status of Oracle in respect to the SunRay platform. And keep in
mind, the SunRay3 client line was still developed by Sun and I think it would
have never be released by Oracle if it weren't almost finished by the time
Oracle acquired Sun.
>> Nevertheless, we can't simply throw away these 200 sunrays today
>> but have to keep them running for a few more years. Therefore I
>> can only hope that I will be able to find a workaround for the
>> current datastore problem I have been running into due to my
>> tries to get SRS 5.3.1 running on Ubuntu 12.04.
>
> fortunately our deployment is not as large so that our decision to
> move was not financially restricted, only emotionally.
Indeed. However, as I said. We almost decided to completely move away from
Oracle products (not only Sunrays, but even servers, etc.) because of the bad
general support and licensing politics Oracle provides to customers. So during
the next few month we will probably decide to move to ThinLinc as soon as our
evaluation is finished. And as I read the documentation of ThinLinc it should
be supported to have the SunRay Server software installed alongside the
ThinLinc software so that we can move over slowly as the SunRay clients fade
out over the years.
>> Have you done some more invstigation on that issue? At least I
>> tried to copy the datastore binary from a 5.2.1 SRS disribution
>> (which is running fine on Ubuntu 11.04) to our new 12.04
>> testinstallation. However, the same problem shows up. The
>> datastore daemon simply refuses to load its own valid config
>> file. Thus, the problem must be in some shared library the
>> datastore daemon uses to parse the config file.
>
> I tried the same things ... including adding old libraries and
> stuff ... but to no avail ... what I did not try, but might be
> interesting, is substituting the sunray ds with openldap
Well, at least I have some good news. By simply replacing the double quotes (")
in the conf file with single quotes (') I was able to get the utdsd daemon
running and properly parsing the config file. However, I am now stuck in
getting the ldap permissions correctly so that "utconfig" will be able to
add/modify entries in the ldap via "ldapadd". Hopefully that should be quite
straight forward as soon as I find more time to dig into that. But don't ask me
why the new SRS 5.3.1 utdsd gets confused as soon as there are double quotes
used in the acl config file.
best regards,
jens
--
Dr. Jens Langner
Institute of Radiopharmacy
Department of Positron Emission Tomography
Helmholtz-Zentrum Dresden-Rossendorf
Bautzner Landstraße 400 | 01328 Dresden
http://www.hzdr.de | +49 351 260 2757
smime.p7s
Description: S/MIME cryptographic signature
_______________________________________________ SunRay-Users mailing list [email protected] http://www.filibeto.org/mailman/listinfo/sunray-users
