From: Jan Kiszka <jan.kiszka@siemens.com>
To: Henning Schild <henning.schild@siemens.com>,
"Hombourger, Cedric" <Cedric_Hombourger@mentor.com>
Cc: "isar-users@googlegroups.com" <isar-users@googlegroups.com>
Subject: Re: [PATCH 1/1] dpkg: acquire lock when calling reprepro
Date: Wed, 22 Aug 2018 15:48:07 +0200 [thread overview]
Message-ID: <b91d84ab-6fee-e378-aa86-ea195febc249@siemens.com> (raw)
In-Reply-To: <20180822154145.07d4a2ca@md1pvb1c.ad001.siemens.net>
On 2018-08-22 15:41, Henning Schild wrote:
> My first reaction was the same, but in the other direction. I would
> prefer a bitbake lock over the flock, if possible.
That would at least involve factoring out the critical section into
dedicated tasks - otherwise we to over-locking and serialize harmless
(and potentially lengthy) steps needlessly.
For now, I would be infavor of fixing in-place and trying to cleanup on top.
Jan
>
> Henning
>
> Am Wed, 22 Aug 2018 12:48:14 +0000
> schrieb "Hombourger, Cedric" <Cedric_Hombourger@mentor.com>:
>
>> Thanks Jan for your quick feedback. I have submitted a v2 with
>> isar.lock removed
>>
>> Cedric
>>
>> -----Original Message-----
>> From: Jan Kiszka [mailto:jan.kiszka@siemens.com]
>> Sent: Wednesday, August 22, 2018 2:20 PM
>> To: Hombourger, Cedric <Cedric_Hombourger@mentor.com>;
>> isar-users@googlegroups.com Subject: Re: [PATCH 1/1] dpkg: acquire
>> lock when calling reprepro
>>
>> On 2018-08-22 13:34, Cedric Hombourger wrote:
>>> As bitbake performs multi-threaded builds, packages may be built
>>> while others get deployed. A lock was already acquired within
>>> dpkg_runbuild but accesses to the isar-apt database were not
>>> protected from do_deploy_deb. Move the lock from the buildchroot to
>>> the isar-apt repository and acquire the (same) lock from
>>> dpkg_runbuild, repo_clean and do_deploy_deb.
>>
>> I think I've seen this issue once but forgot to dig into the details.
>> It looks consistent.
>>
>>>
>>> Signed-off-by: Cedric Hombourger <Cedric_Hombourger@mentor.com>
>>> ---
>>> meta/classes/dpkg-base.bbclass | 8 ++++++--
>>> meta/recipes-devtools/buildchroot/files/build.sh | 2 +-
>>> meta/recipes-kernel/linux/files/build-kernel.sh | 2 +-
>>> 3 files changed, 8 insertions(+), 4 deletions(-)
>>>
>>> diff --git a/meta/classes/dpkg-base.bbclass
>>> b/meta/classes/dpkg-base.bbclass index a41df2c..7137968 100644
>>> --- a/meta/classes/dpkg-base.bbclass
>>> +++ b/meta/classes/dpkg-base.bbclass
>>> @@ -71,6 +71,8 @@ do_build() {
>>> mount -t proc none ${BUILDCHROOT_DIR}/proc
>>> fi'
>>>
>>> + # Create database lock file here so we own it (not root) and
>>> perform actual build
>>> + touch ${DEPLOY_DIR_APT}/${DISTRO}/db.lock
>>> dpkg_runbuild
>>>
>>> sudo umount ${BUILDROOT} 2>/dev/null || true @@ -82,7 +84,8
>>> @@ CLEANFUNCS += "repo_clean"
>>> repo_clean() {
>>> PACKAGES=$(cd ${S}/..; ls *.deb | sed 's/\([^_]*\).*/\1/')
>>> if [ -n "${PACKAGES}" ]; then
>>> - reprepro -b ${DEPLOY_DIR_APT}/${DISTRO} \
>>> + flock ${DEPLOY_DIR_APT}/${DISTRO}/db.lock \
>>> + reprepro -b ${DEPLOY_DIR_APT}/${DISTRO} \
>>> --dbdir ${DEPLOY_DIR_DB}/${DISTRO} \
>>> -C main -A ${DISTRO_ARCH} \
>>> remove ${DEBDISTRONAME} \ @@ -93,7 +96,8 @@
>>> repo_clean() {
>>> # Install package to Isar-apt
>>> do_deploy_deb() {
>>> repo_clean
>>> - reprepro -b ${DEPLOY_DIR_APT}/${DISTRO} \
>>> + flock ${DEPLOY_DIR_APT}/${DISTRO}/db.lock \
>>> + reprepro -b ${DEPLOY_DIR_APT}/${DISTRO} \
>>> --dbdir ${DEPLOY_DIR_DB}/${DISTRO} \
>>> -C main \
>>> includedeb ${DEBDISTRONAME} \
>>
>> There is also
>>
>> do_deploy_deb[lockfiles] = "${DEPLOY_DIR_APT}/isar.lock"
>>
>> in that file. I suppose that can now be dropped, can't it?
>>
>>> diff --git a/meta/recipes-devtools/buildchroot/files/build.sh
>>> b/meta/recipes-devtools/buildchroot/files/build.sh
>>> index e2dabab..e8a9c17 100644
>>> --- a/meta/recipes-devtools/buildchroot/files/build.sh
>>> +++ b/meta/recipes-devtools/buildchroot/files/build.sh
>>> @@ -50,7 +50,7 @@ install_cmd="apt-get -o
>>> Debug::pkgProblemResolver=yes --no-install-recommends -y
>>>
>>> # Install all build deps
>>> mk-build-deps $set_arch -t "${install_cmd}" -i -r
>>> debian/control -) 42>/dpkg.lock
>>> +) 42>/isar-apt/db.lock
>>>
>>> # If autotools files have been created, update their timestamp to
>>> # prevent them from being regenerated diff --git
>>> a/meta/recipes-kernel/linux/files/build-kernel.sh
>>> b/meta/recipes-kernel/linux/files/build-kernel.sh
>>> index 8b58e93..2ff5a1e 100644
>>> --- a/meta/recipes-kernel/linux/files/build-kernel.sh
>>> +++ b/meta/recipes-kernel/linux/files/build-kernel.sh
>>> @@ -34,7 +34,7 @@ REPACK_LINUX_IMAGE_DIR="${REPACK_DIR}/linux-image"
>>> REPACK_LINUX_HEADERS_DIR="${REPACK_DIR}/linux-headers"
>>>
>>> # Lock-protected because apt and dpkg do not wait in case of
>>> contention -flock /dpkg.lock \
>>> +flock /isar-apt/db.lock \
>>> apt-get install -y -o Debug::pkgProblemResolver=yes \
>>> --no-install-recommends ${KBUILD_DEPENDS}
>>>
>>>
>>
>> Jan
>>
>> --
>> Siemens AG, Corporate Technology, CT RDA IOT SES-DE Corporate
>> Competence Center Embedded Linux
>>
>
--
Siemens AG, Corporate Technology, CT RDA IOT SES-DE
Corporate Competence Center Embedded Linux
next prev parent reply other threads:[~2018-08-22 13:48 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-08-22 11:34 [PATCH 0/1] acquire database " Cedric Hombourger
2018-08-22 11:34 ` [PATCH 1/1] dpkg: acquire " Cedric Hombourger
2018-08-22 12:20 ` Jan Kiszka
2018-08-22 12:48 ` Hombourger, Cedric
2018-08-22 13:41 ` Henning Schild
2018-08-22 13:47 ` Hombourger, Cedric
2018-08-22 13:48 ` Jan Kiszka [this message]
2018-08-26 6:21 ` [PATCH v3 1/3] dpkg: move installation of dependencies into dpkg_prepare Cedric Hombourger
2018-08-26 6:21 ` [PATCH v3 2/3] linux: leverage dpkg_prepare to install build dependencies Cedric Hombourger
2018-08-26 6:21 ` [PATCH v3 3/3] dpkg+linux: use Isar's lock via do_prepare[lockfiles] instead of flock Cedric Hombourger
2018-08-28 12:12 ` [PATCH v3 1/3] dpkg: move installation of dependencies into dpkg_prepare Maxim Yu. Osipov
2018-08-31 8:10 ` Jan Kiszka
2018-08-31 8:15 ` chombourger
2018-08-31 8:21 ` Jan Kiszka
2018-08-31 8:28 ` Jan Kiszka
2018-08-31 8:33 ` chombourger
2018-08-31 8:41 ` Jan Kiszka
2018-08-31 9:05 ` Jan Kiszka
2018-09-01 6:58 ` Jan Kiszka
2018-09-01 8:44 ` Cedric Hombourger
2018-09-01 9:31 ` Jan Kiszka
2018-08-22 11:58 ` [PATCH 0/1] acquire database lock when calling reprepro Alexander Smirnov
2018-08-22 12:07 ` Hombourger, Cedric
2018-08-22 12:47 ` [PATCH v2] dpkg: acquire " Cedric Hombourger
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=b91d84ab-6fee-e378-aa86-ea195febc249@siemens.com \
--to=jan.kiszka@siemens.com \
--cc=Cedric_Hombourger@mentor.com \
--cc=henning.schild@siemens.com \
--cc=isar-users@googlegroups.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox