Re: [Xen-users] is safe to compile and install the gplpv (changeset939) for production?

[Flako <subforos@xxxxxxxxx>]

2011/7/6 James Harper <james.harper@xxxxxxxxxxxxxxxx>:
>>
>> Hello
>>      Compile the latest version of gplpv (changeset 939) and these
>> weeks I think to try.
>>      As I will not be able to test all complete (not know where to get
>> started) and seeing that there is poor documentation on gplpv (the doc
>> is the source code), I'm not sure if it's a good idea to use this
>> compilation in production.
>>      I really do not know if the version "changeset 939" is stable,
>> testing or a simple snapshot that only a madman would put in
>> production running a DBMS.
>>      Nor is a stable version gplpv_2003x32_0.11.0.238.msi .. :)
>>
>>           You think? is safe to install gplpv (changeset 939) for
> production?
>>
>
> For 2000, XP, and 2003 you should be fine using the latest committed
> changeset as I've run the network and vbd drivers through the MS test
> suite.
>
> I'm almost finished running the Vista+ version drivers through the MS
> testing suite, so should have something I can call stable again for the
> newer OS's soon. All the problems I'm finding in the network drivers so
> far are not crash related though so they should still work, there are
> just a few oddities with statistics reporting and some OID's I should
> support but don't. I haven't run the vbd drivers through the test suite
> yet though.
>
> James
>
>

Hello James
I'm glad you're you, who replied .. gives me confidence in my new
files. msi gplpv_2003x32_0.11.0.295.msi :)

 The compile on Windows XP Version 5.1.2600] 32Bit with WinDDK
7600.16385.1 and I have some doubts errors, if you got want to answer
(in a line) is appreciated.
 Also used to indicate some bugs associated with files .bat used to compile.


1.  Which achieves better performance static or static lib build or
dynamic lib build?
   (After much thought .. I chose the static, but the desicion is
based on 70% of ignorance)

2. Possible bug?
    In the file sign.bat, line 19, generates the line:
"C: \ WinDDK \ 7600.16385.1 \ bin \ x86 \ signtool.exe sign / v / s
PrivateCertStore / n / t
http://timestamp.verisign.com/scripts/timestamp.dll xenpci \
objchk_wnet_x86 \ i386 \ xenpci. sys "

   that when executed returns:  SignTool Error: File not found:

  according 
https://knowledge.verisign.com/support/code-signing-support/index?page=content&id=SO326&actp=search&viewlocale=en_US&searchid=1310076513769
the correct url is  http://timestamp.verisign.com/scripts/timstamp.dll
(without the e), but neither goes ..
   This is really necessary? influencing not have it?


2. Possible bug when compiling with 32Bit Windows XP?
    In archive.bat file in line 4 is the sentence "IF NOT EXIST
%SEVENZIP% GOTO :end" for me is worth the ':'  :)
    For my part not know the syntax, Win XP does not work.


3. Possible bug when building on Windows XP?
   I had to insert the statement 'exit / b 0' at the end of the files:
sign.bat, wix.bat archive.bat and for these to run properly on the
line "cmd /C "%DDK_PATH%\bin\setenv.bat %DDK_PATH%\ chk WLH && CD /D
"%PV_DIR%" && build -cZg && call sign.bat && call archive.bat && call
wix.bat"  File makedist.bat.
   Without that line these files are not retrieved variables
setenv.bat defined variables, nor were run and call archive.bat
wix.bat.

4.  It's normal to win a 2003 32bit domU with 9GB ram (on a 64bit
Linux), the driver xenvbd be 50% less than in disk access performance,
in relation to windows as host?
   It is acceptable performance? I can see graphs of benchmark WindowsGplPv?

 Thanks and sorry for the English

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users