Liam Slusser | 9 Jul 11:51 2014
Picon

Running a debug kernel

OI Developers -

Does running a debug kernel cause a performance impact?  Basically I build a nightly image with debugging enabled (defaults) and all seems to work great but I'm wondering if there is a performance penalty and I should rebuild without debugging?

thanks,
liam
_______________________________________________
oi-dev mailing list
oi-dev <at> openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev
Cedric Blancher | 7 Jul 18:24 2014
Picon

Import Illumos vmware image into Amazon EC2 cloud?

Has anyone tried to cheat AWS OS choice limitations and import a
vmware disk image with Illumos installed into the AWS cloud? Did it
work? Which problems did you encounter?

Ced
--

-- 
Cedric Blancher <cedric.blancher <at> gmail.com>
Institute Pasteur
Alexander Pyhalov | 6 Jul 08:26 2014
Picon

oi-jds repository

Hello, Milan.

It seems that http://hg.opensolaris.cz/oi-jds/ is unavailable and the 
whole http://opensolaris.cz/ seems to be down.
What is the reason for this? If there are some problems with this host 
perhaps it worths to make oi-jds available
at some other location?
If necessary I can host it at our DC or I'm sure that if you ask Andrzej 
or Adam they can provide hosting at EveryCity.
--

-- 
System Administrator of Southern Federal University Computer Center
Alexander Pyhalov | 5 Jul 15:07 2014
Picon

Re: [OpenIndiana-discuss] success stories about SRSS and OI

Hello.

Jonathan Adams писал 04.07.2014 19:40:
> ran a "dhcpconfig -U" then destroyed the existing "/var/dhcp"
> files/directory, afterwards I cannot create any new DHCP files:
> 
> root <at> jadlaptop:~# dhcpconfig -D -r files -p /var/dhcp -l 28800
> dhcpconfig: Error - files is not a valid resource.
> 
> root <at> jadlaptop:~# dhcpconfig -D -r SUNWfiles -p /var/dhcp -l 28800
> dhcpconfig: Error - creating location: /var/dhcp.
> 
> root <at> jadlaptop:~# dhcpconfig -D -r SUNWbinfiles -p /var/dhcp -l 28800
> dhcpconfig: Error - creating location: /var/dhcp.
> 
> all fail almost instantly ...
> 
> I'm sure it used to work with java1.7, but just in case I removed 
> /usr/java
> and re-linked to the java1.6 on the system, which failed in the same 
> way.
> 
> I've attached the bzip2'd log files from truss ...
> 

BTW, I see the same behavior on /dev with recent illumos-gate (compiled 
with GCC).
However, dhcpconfig works on stock /dev a8.
So, I see two possibilities:
1) something is wrong with using GCC & dhcp Java configuration tool
2) dhcpconfig is broken in current illumos-gate.

---
System Administrator of Southern Federal University Computer Center

_______________________________________________
oi-dev mailing list
oi-dev <at> openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev
DormitionSkete@hotmail.com | 4 Jul 21:28 2014
Picon

Please forgive me.

Ugh.  I meant to post that boot hang message on the OpenIndiana Discussion list.

Please forgive the double post.
DormitionSkete@hotmail.com | 4 Jul 21:22 2014
Picon

Boot hangs.

Hi, everyone.  I have a little problem.  Our primary server hung on something last night, and I don’t know
what.  I could not get into it, even with ssh.  So, maybe I did something stupid, but I powered the machine off. 
Now it will not boot.

Editing the GRUB menu options, removing the console line and adding “-v -m verbose” to the kernel line
gave me the following output, after a very long time.  (Several minutes, which is long for this machine.)

—

Jul 4 11:51:59 svc.startd[10]: svc:/system/boot-archive:default: Method or service exit timed out. 
Killing contract 27.
Jul 4 11:51:59 svc.startd[10]: svc:/system/boot-archive:default: Method
"/lib/svc/method/boot-archive" failed due to signal KILL.
timed out. Killing contract 27.

—

It kicked me into a maintenance boot prompt.

svcs -xv

[...] <snip>

svc:/network/rpc/smserver:default (removable media management)
State: unitialized since July 4, 2014 11:48:58 AM MDT
Reason: Restarter svc:/network/inetd:default is not running.
See: http://illumos.org/msg/SMF-8000-5H
See: man -M /usr/share/man -s 1M rpc.smserverd
Impact: 2 dependent services are not running:
svc:/milestone/multi-user-server:default
svc:/system/zones:default

—

Since it referred to the removable media management, I unplugged the IDE cables from the CD Rom and the
Floppy drive, disabled the IDE controller in the BIOS, and tried booting again.  Same thing.

I tried booting into a previous boot environment.  Same thing.

Could anyone please tell me what I should do next?  If so, I would GREATLY appreciate it.

Peter, hieromonk
Alexander Pyhalov | 4 Jul 12:27 2014
Picon

[HEADSUP] GCC 4.8.3 is a default oi-userland compiler now

GCC 4.8.3 was added to oi-userland.
oi-userland was switched to use it as a default compiler.
So, don't forget to switch your compiler.
--

-- 
Best regards,
Alexander Pyhalov,
system administrator of Computer Center of Southern Federal University
Alexander Pyhalov | 3 Jul 17:13 2014
Picon

Review GCC 48

Hello.

I think that I'm finally finished with gcc 4.8.3.

Changes are here: https://github.com/pyhalov/oi-userland/compare/gcc48 .
Please, review.

The component is almost a copy of GCC 4.7. It's based on OmniOS GCC 4.8 
version.

There are several changes compared to OmniOS version:
1) adopting  patches/unwind.patch (a copy of patch from 
dilos-userland-review) - fixes 
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59788.
2) I don't include old libstdc++.so versions in g++-4-runtime as they 
are useless
3) preserving fixinc.in.patch - don't know if it's necessary.

I was able to rebuild oi-userland with GCC 4.8 and I suppose to switch 
default compiler to 4.8.3 as soon as it's available (as current compiler 
can't build clang).

Whether the whole oi-userland should be rebuilt after updating GCC - is 
an open question. I don't know.
I run desktop VM with GCC 4.8 runtime libs and GCC4.7-compiled soft. No 
problems so far.

Also I have some clang patches (including clang 3.4.2 component), which 
depend on this work. However, as testing shows, clang is still  a crap. 
It can compile some code (for example, clang), but clang-compiled clang 
already dies on compiling boost. GCC-compiled clang can almost can 
compile boost (it seems there are some problems with headers - unlike 
gcc it doesn't like overriding funcion declaration in global namespace 
(when using B::A overrides already defined A).

--

-- 
Best regards,
Alexander Pyhalov,
system administrator of Computer Center of Southern Federal University
Udo Grabowski (IMK | 2 Jul 13:39 2014

Problem with sfe-enc wine vs. sfe libgsm

Hi,

sfe-enc wine cannot be updated because it wants libgsm from
sfe-enc, but installed is libgsm from sfe. The problem is that
sfe-enc indeed has no libgsm binary package, only the source...
So deinstalling libgsm from sfe (and dependances) does not
help until a binary sfe-enc libgsm is available. [oi 151a9]
--

-- 
Dr.Udo Grabowski   Inst.f.Meteorology & Climate Research IMK-ASF-SAT
http://www.imk-asf.kit.edu/english/sat.php
KIT - Karlsruhe Institute of Technology           http://www.kit.edu
Postfach 3640,76021 Karlsruhe,Germany T:(+49)721 608-26026 F:-926026

Attachment (smime.p7s): application/pkcs7-signature, 2773 bytes
_______________________________________________
oi-dev mailing list
oi-dev <at> openindiana.org
http://openindiana.org/mailman/listinfo/oi-dev
Alexander Pyhalov | 30 Jun 10:02 2014
Picon

gcc 4.8 update and binary compatibility

Hello.
I'm looking at adding gcc 4.8 package.

I have a big question about g++-4-runtime package.
For example, OmniOS g++-4-runtime includes
/usr/lib/libstdc++.so.6.0.18
/usr/lib/libstdc++.so.6.0.17
and so on.

But if we look, for example, at percona:
$ ldd /usr/percona-server/5.6/bin/mysql
         libthread.so.1 =>        /lib/libthread.so.1
         libsocket.so.1 =>        /lib/libsocket.so.1
         libssl.so.1.0.0 =>       /lib/libssl.so.1.0.0
         libcrypto.so.1.0.0 =>    /lib/libcrypto.so.1.0.0
         libcurses.so.1 =>        /lib/libcurses.so.1
         libstdc++.so.6 =>        /usr/lib/libstdc++.so.6
         libm.so.2 =>     /lib/libm.so.2
         libgcc_s.so.1 =>         /usr/lib/libgcc_s.so.1
         libc.so.1 =>     /lib/libc.so.1
         libnsl.so.1 =>   /lib/libnsl.so.1
         libdl.so.1 =>    /lib/libdl.so.1
         libmp.so.2 =>    /lib/libmp.so.2
         libmd.so.1 =>    /lib/libmd.so.1

we can see that it is linked to libstdc++.so.6. So, if we update link, 
runtime linker will load binary incompatible library.

Are there any ideas how to preserve binary compatibility on this update?
Or perhaps we should apply gcc-05-LINK_LIBGCC_SPEC-46.diff patch and 
avoid providing runtimes for gcc in /usr/lib ?
--

-- 
Best regards,
Alexander Pyhalov,
system administrator of Computer Center of Southern Federal University
Alexander Pyhalov | 27 Jun 11:12 2014
Picon

TODO list and JDS conversion status

Hello.

I've just created TODO list with some ideas: 
http://wiki.openindiana.org/oi/TODO+list

and status page with current progress of JDS migration: 
http://wiki.openindiana.org/oi/JDS+migration

Please, add your ideas to TODO list and don't hesitate to implement one 
or two points from the list :)

--

-- 
Best regards,
Alexander Pyhalov,
system administrator of Computer Center of Southern Federal University

Gmane