Search Results: "henrich"

15 March 2013

Hideki Yamane: pbuilder/cowbuilder and piuparts failed

Today, I've failed to create pbuilder's base.tgz and cowbuilder's base.cow, and log says as below

Get:1 http://ftp.jp.debian.org sid Release.gpg [836 B]
Hit http://ftp.jp.debian.org sid Release
Ign http://ftp.jp.debian.org sid Release
Hit http://ftp.jp.debian.org sid/main amd64 Packages/DiffIndex
Hit http://ftp.jp.debian.org sid/main Translation-en
Fetched 836 B in 0s (3011 B/s)
Reading package lists...
W: GPG error: http://ftp.jp.debian.org sid Release: The following signatures were invalid: BADSIG AED4B06F473041FA Debian Archive Automatic Signing Key (6.0/squeeze) <ftpmaster debian.org="">
I: Obtaining the cached apt archive contents</ftpmaster>
"The following signatures were invalid: BADSIG" ?

apt (0.9.7.8) unstable; urgency=criticial

* SECURITY UPDATE: InRelease verification bypass
- CVE-2013-1051

[ David Kalnischk ]
* apt-pkg/deb/debmetaindex.cc,
test/integration/test-bug-595691-empty-and-broken-archive-files,
test/integration/test-releasefile-verification:
- disable InRelease downloading until the verification issue is
fixed, thanks to Ansgar Burchardt for finding the flaw
Maybe this cause it (just wild guess).

And piuparts also failed.


8m16.5s DEBUG: Starting command: ['chroot', '/tmp/tmpivurbJ', 'eatmydata', 'umount', '/proc']
8m16.5s DUMP:
umount: /proc: device is busy.
(In some cases useful info about processes that use
the device is found by lsof(8) or fuser(1))
8m16.5s DEBUG: Command failed (status=1), but ignoring error: ['chroot', '/tmp/tmpivurbJ', 'eatmydata', 'umount', '/proc']
8m16.5s DEBUG: Starting command: ['rm', '-rf', '--one-file-system', '/tmp/tmpivurbJ']
8m16.6s DUMP:
rm: skipping '/tmp/tmpivurbJ/proc', since it's on a different device
8m16.6s ERROR: Command failed (status=1): ['rm', '-rf', '--one-file-system', '/tmp/tmpivurbJ']
rm: skipping '/tmp/tmpivurbJ/proc', since it's on a different device

8m16.7s DEBUG: Starting command: ['chroot', '/tmp/tmpivurbJ', 'umount', '/proc']
8m16.8s DUMP:
chroot: failed to run command 'umount': No such file or directory


It mounts binfmt_misc mount under /proc and try to umount /proc without umount it again, again and again... then failed. What causes this annoying behavior?

Please let me know.

12 March 2013

Hideki Yamane: Open Source Conference 2013 Tokushima ( )

Some Debian/Ubuntu folks had participated in Open Source Conference 2013 Tokushima, in (Tokushima), which is in Shikoku ( , smallest of the four main islands of Japan).

(Oh, there's Matz. Can you find it? :)





I have not been there but I hope they had a good time...

10 March 2013

Hideki Yamane: great IOPS device and if we'd have it...?


I know some folks work at Fusion IO, maybe you know, one of the leader of PCIe flash storage system. And they say their products are "great", since their device put amazing IOPS as they press released (9,000,000 IOPS, and Wikipedia says "SAS HDDs put 200 IOPS". If you have a SSD in your laptop/desktop, you'd love it. It puts almost 10,000 IOPS or so). Such PCIe SSD can bring "Moon shot", change the rules for managing CPU, Memory, Disk and Network for servers, IMHO.


Then, I'm dreaming - if we Debian would have such MAGICAL device, what is the most efficient way to improve/accelerate/Kaizen our project? To use mailing list server? BTS? UDD? CI?

Dear lazyweb, please let me know what is the bottleneck of our development, and great IOPS capacity will fix it.

(and please don't say about things based on human resource, since PCIe SSD devices cannot improve such things, you know ;)

8 March 2013

Hideki Yamane: Re: checking package with clang

As I wrote before, I've posted a patch to BTS that add an option use clang compiler to pbuilder. However, pbuilder maintainer Junichi wasn't satisfied with it because he thought it seems to be overhead, introduce complexity and also hook script is enough.

I didn't agree with his opinion since I thought every time we need to add and remove its script, but Thorsten Glaser shows just specifying environment variables can solve it, and moreover, he points other compiler can be supported via hook, too. Sounds good, right? :)

Then I post a new hook script, it works with gcc-4. 4-8 and clang (gcc-4.8 needs experimental apt line for pbuilder). Please test it if you'd be interested.

Hideki Yamane: memo: comparing pbuilder with tmpfs and cowbuilder


Test: build GNU hello 20 times with pbuilder/cowbuilder

pbuilder (with tmpfs)
real 9m11.320s
user 5m10.215s
sys 1m21.989s
cowbuilder
real 8m31.191s
user 4m27.629s
sys 1m15.661s

Conclusion: probably it's better to use cowbuiler (at least for me)

23 February 2013

Hideki Yamane: Open Source Conference 2013 Tokyo/Spring

Well, some Debian folks have participated in Open Source Conference 2013 Tokyo/Spring, at Meisei Univ.( ).

Takahide Nojima talked "Debian Update", various thing during about Debian during this 1 or 2 years. Participates are not so much, because it was first session and Meisei Univ is a little bit far from central Tokyo (it takes almost 40 minutes by train...), it was tough (at least for me ;)


Also we had a booth, chatted with users (and potential users) - 1000 participates were there in Meisei Univ.

Thanks folks.

22 February 2013

Hideki Yamane: Postgresql + GPU / SELinux approach

I've participated an event (18th, Feb), Kohei Kaigai from NEC talked about 3 topics.


  • Places to Visit in Europe
  • PG-Strom: GPU Accelerated AsynchronousSuper-Parallel Query
  • Row-Level Security: DB Security, prevent information leak

I've interested in PG-Strom, it can accelerated query 8 times with $100 GPU card, for more detail, see above link.

This event was hosted in Red Hat office at Yebisu( ), Tokyo. Thanks to Ryo Fujita from Red Hat for his coordination.


16 February 2013

Hideki Yamane: "Debian Hot Topics"


As I wrote before, I've started a series on "Software Design" magazine to introduce something inside Debian, news items, essay and so on. It's named "Debian Hot Topics" but it was decided after finished writing my article, so it comes to just an essay about distribution development, instead of "Hot Topics" ;-)

Anyway, it's a good chance for advertising (after all, no cost but get fee :)

Hideki Yamane: Developers Summit 2013 in Tokyo


"Developers Summit 2013" was held in (Meguro Gajoen), Tokyo. I've participated to this event to get information about development method and know-how.

Keyword is: Agile, Scrum, Continuous Delivery, Deploy 1000 times in an hour




Well, some talks were about Agile development (especially Scrum), the experience that how they tried to introduce "Agile" approach to their work to improve thier problems/troubles and what they've learned and got from it. I think it's very exciting if we Debian can deploy (=release) stable anytime we want! To achieve it, we should get Agility by learn their way of values and practice. I hope I'll go to DebConf13 then give a talk about it :)

Some community booth were there, I went and chatted with Vine Linux developer and community member for a while (Vine Linux is a Japanese local distribution since and we Debian get VLGothic font as default Japanese font from them :) They promoted "Handbook for developers" and it looks so good shape. Maybe you know, we also have Debian Develpers Reference(and I've completely translated it to Japanese), but their one is more catchy, of course I think we can borrow something from it to improve our DDR.

And I've got a book about CloudStack named "CloudStack " (Introduction to CloudStack, for beginners, users and professionals) at CloudStack (Japan CloudStack users group) booth, by winning a prize :)

Also bought some books at publishers booth because 10+Tax5% off was so attractive...

13 February 2013

Hideki Yamane: introducing debian-mirror.sakura.ne.jp

Now we Debian JP (local Debian community in Japan) are pleased to announce new Debian full-mirror server named debian-mirror.sakura.ne.jp is running, kindly offered by Sakura Internet, Inc. which is one of most famous VPS/hosting provider in Japan.

So, their VPS/hosting user can use it without any setting now via magical GeoMirror as ftp.jp.debian.org (thanks to Yasuhiro Araki).

It is run at Ishikari DC, northen island Hokkaido.



I had a chance to visit there 3 month ago (you can see my entry in Google+ if you want), it was so interesting and exciting. They are trying to make power-saved data center as Google, Facebook or so, and it had succeed now since its PUE (Power Usage Effectiveness) is same level as Google one, wow :-)

I hope we will provide stable and comfortable Debian environment to their users and they would use Debian more and more, then we'll get their "wants" and "needs" to improve Debian itself.

Thanks to folks in Sakura Internet!

11 February 2013

Hideki Yamane: checking package with clang

Sylvestre Ledru wrote an article "Rebuild of Debian using clang 3.2". Probably package maintainers want to know about their package status with clang in PTS, Paul Wise has already posted to BTS#699936.

And, today I've hit upon an idea, then created a tiny patch for clang support for pbuilder. It may help you to check your package with clang.

10 February 2013

Hideki Yamane: license for MIBs

Unfortunately, most of MIBs doesn't meet DFSG since RFC is non-free. However, RFC5378 was released and it says "code components in RFCs are licensed under Simplified BSD license (BSD-3-clause)", See Trust Legal Provisions (TLP) Documents page.

Code components are

ABNF definitions, XML Schemas, XML DTDs, XML RelaxNG definitions, tables of values, MIBs, PIBs, ASN.1, and classical programming source code
So, recent MIBs can be included to Debian package :)

But it isn't enough because those RFCs are almost 30/260, 90% are not DFSG-free. And 190/260 RFCs are owned by ISOC and IETF trustee, I hope thay will release old MIBs code in RFCs under DFSG-free license as same as after RFC5378.

29 January 2013

Hideki Yamane: upstream and distro has a "same" goal


Tom Marble has a nice blog entry "Crowdsourcing Upstream Refactoring". It's interesting and I agree with most, but I want to say against one thing in "conclusion" in his presentation.

He said "upstreams and distros have different goals" but I don't think so. We distro and upstream has a same goal, "Deliver the value to users", but we distro have much criteria than upstream, like license, non-duplicate library, etc.

If we (distro and upstream) cannot share this view, then we would be failed, IMHO.

28 January 2013

Hideki Yamane: convention over configuration

" convention over configuration" is an approch for software design. Now I recognize why dh7 style is good.

24 January 2013

Hideki Yamane: net-snmp package

Current net-snmp package maintainers are busy, so now I'm trying to update it. One of maintainers Jochen Friedrich kindly checked my experimental package (thank!), and point MIB files should be removed to upload to Debian out (at least).

Anyway, current changelog is here (a little bit long). If you want to try it, get dsc file and build it.


net-snmp (5.7.2~dfsg-1~0.1) UNRELEASED; urgency=low
* Non-maintainer upload.
* New upstream version 5.7.2 (Closes: #557348, #631063, #684388, #599929,
#673197, #581185, #558356, #568550, #514842, #445608, #557186, #411858,
#428824, #611837)
* debian/control
- set "Standards-Version: 3.9.4"
- set "Build-Depends: debhelper (>= 9)" to enable hardening
- set "Build-Depends: automake", instead of automake1.9
- drop "Build-Depends: python-central", use dh_python2 by default, instead
- remove duplicational "Priority:" and "Section:" fields
- make it "Multi-Arch" enable
- add description for non-free snmp-mibs-downloader for users' convinience
(Closes: #561124)
- add "Build-Depends: libmysqld-dev" to support MySQL
- add "Build-Depends: dh-autoreconf"
- use python-all (2.6.6-3~) instead of python-all-dev (>= 2.5.4-1~), python
(>=2.3.5-7) for Build-Depends
- add "Build-Depends: python2.7-dev"
- s/libsnmp-python/python-netsnmp/ as Python Policy compliant (Closes: #661899)
* debian/compat
- set 9
* debian/*.install
- split into *.manpages
* debian/fixman: drop it.
* debian/*.manpages (Closes: #505149)
- remove unnecessary"snmp" from tail of all man pages
* debian/libsnmp-base.install
- move /usr/share/mibs to /usr/share/snmp/mibs (probably, previous setting
was just wrong...)
- install all mibs/*.txt
* debian/libsnmp-dev.install
- change from "usr/lib/*" to "usr/lib/*/*" to deal with Multi-Arch
* debian/libsnmp30.install
- change from "usr/lib/*.so" to "usr/lib/*/*.so" to deal with Multi-Arch
* debian/snmp.dirs: remove it since unnecessary
* debian/snmpd.dirs: remove lintian overrides direcotry, add /etc/snmp
* debian/libsnmp-dev.dirs: add it
* debian/libsnmp-perl.examples: add it since dh_installexamples target in
previous debian/rules doesn't work correctly.
* debian/patches
- remove all *.README files
- add fix_typo_in_snmpd.conf.patch (Closes: #603593)
Thanks to Slavko <linux@slavino.sk>
- add fix_logging_option.patch (Closes: #616437)
- add fix_snmpcheck_perl_path.patch to provide snmpcheck (Closes: #44373)
- add net-snmp-config_multi-arch.patch to enable Multi-Arch
- add fix_regular_expression.patch enable to build under Multi-Arch
- add to snmptranslate.1.patch fix lintian "hyphen-used-as-minus-sign"
warnings
- add fix_spelling_error.patch to fix typo
- adjust and refresh 03_makefiles.patch
- drop 25_duplicate_iftable.patch and 44_nlist_kvm.patch since it cannot be
applied to current code.
- drop 32_mnttab_path.patch since it was merged to upstream as
configure.d/config_os_misc4
- drop 56_manpage.patch since most of patches are merged, and others are
unnecessary because upstream files are disappeared.
- refresh 61_vacm_missing_dependency_check.patch
- drop 63_fix_shell.patch since it was merged to upstream
- reapplied 64_missing_lib.patch since upstream source has been changed.
- drop 65_CVE-2012-2141.patch since it was merged to upstream
- drop 66_formatstrings.patch since almost merged to upstream, some of
code are changed in upstream and become unnecessary.
- move 08_defaultconfig.patch to debian/snmptrapd.conf
* debian/rules
- enable hardening
(TODO: building perl module would be failed without -pie)
- specify LDFLAGS to enable hardening
- add "--with autotools-dev,autoreconf"
- export DEB_BUILD_MAINT_OPTIONS to enable hardening
- drop "dh --with python-central" (Closes: #616913)
- drop "include /usr/share/python/python.mk"
- remove "$(PYVERS:%=debian/python-install-stamp-%)"
- move exist python targets under override_dh_install
- remove unnecessary .PHONY lines
- enable tests by removing no instructions with dh_auto_test line
- set DEB_HOST_MULTIARCH to enable Multi-Arch
- enable install snmpcheck
- don't specify to copy files but install them by using .install file
(mostly)
- enable mysql support
- enable AES support (Closes: #447705)
- don't make symlink for /usr/share/doc/libsnmp-perl since other are okay
but it has examples files.
- snmpd doc files link to libsnmp (Closes: #453124)
- adjust dh_strip for dbg package.
- adjust dh_clean target
- remove override_dh_auto_build since LD_RUN_PATH is clearly specified
during building perl modules
- add mibII/mta_sendmail to build modules (Closes: #641608)
* debian/clean
- most listed files are cared by autotools-dev, so removed.
* debian/snmp.conf: fix typo (Closes: #623499, #647468)
* debian/snmpd.init
- force remove pid files (Closes: #528104)

-- Hideki Yamane <henrich@debian.org> Wed, 18 Jul 2012 13:39:41 +0900

And if I can put it to Debian, then I hope someone will maintain it...

20 January 2013

Hideki Yamane: digging CVEs

Sometimes I would check security-tracker to find problems that is already fixed but not tagged as so.

Today, I found problems that not in unstable package but not tagged as fixed for Eucalyptus, because it was removed from unstable once but re-introduced to archive again, then data in security-tracker was not updated as well. So update it (CVE-2010-3905 and CVE-2011-0730).

And also found vulnerablity in firebird2.5 was fixed in upstream svn, so pick it and made it as a patch, just report to BTS (for squeeze, too).

For isc-dhcp, it has a vulnerability from BIND9... I've been surprised but dhcp includes bind9 source since 4.2, so BIND9 vulnerability affects isc-dhcp. Pick a diff from bind9 for squeeze security update and made a patch (Bug#698597) .

10 January 2013

Hideki Yamane: 256 -> Debian

I had a cold, so couln't write anything (but back to be normal, almost).

I'm also Ubuntu LoCo team member as Ubuntu Japanese team. Some of team members write articles on gihyo.jp website during the years, and now our team leader Jun Kobayashi(*) celebrate 256 times(!) as Android Ubuntu Debian (Run Ubuntu Debian on Android)... Yes, Debian, you know :)

*) He also wrote very funny article ago, see this...

4 January 2013

Hideki Yamane: As tech writer

Sometimes I write articles for Japanese Tech Magazine, the recent one is for Software Design 2013/01. It's "Overview for 2013", of course I wrote about Debian :-)



And luckyly, editor requested the series (yeah!), so I'll do it for a while. I'm happy to get a chance to provide Debian information to readers, thanks!

... and I hope they'll be a media sponsor for DebConf in Japan (in some day :)

3 January 2013

Hideki Yamane: Hello New year, Hello Planet

This is first post to Planet Debian...

Hi, let me introduce myself a bit... I'm a Debian Developer living in Japan. I'm not a geek, programmer... but love Debian and enjoy to participate this project :-) I want to share something interesting related to Debian.


Recently, also I'm curious about Debian cloud computing service - AmazonEC2, Rackspace, WindowsAzure, Google Compute Engine and so on.
Recent two big topic for me is

Especially Aeolus is huge software stack and need to package a lot of Ruby gems, so I've joined Pkg-ruby-extras-maintainers and started packaging some gems. I'm newbie for Ruby, so please help if there's something wrong with my packaging.

30 December 2012

Hideki Yamane: put, put, put

put, put, put.

Just ITPed(dependency package is not satisfied yet).

Next.

Previous.