ehlo,
The simple reproducer is to use getent passwd,group with non-existing entry. Without the patch you will see that "/var/lib/sss/mc/passwd" is opened twice. Onece with mode "rw" opened by sssd_nss and the 2nd time with "r-" opened by sssd-client (getpwnam, getpwuid, getgrnam, getgrgid)
LS
On 08/08/2016 02:40 PM, Lukas Slebodnik wrote:
ehlo,
The simple reproducer is to use getent passwd,group with non-existing entry. Without the patch you will see that "/var/lib/sss/mc/passwd" is opened twice. Onece with mode "rw" opened by sssd_nss and the 2nd time with "r-" opened by sssd-client (getpwnam, getpwuid, getgrnam, getgrgid)
LS
Hello Lukas,
it looks good to me. I am waiting for CI tests.
Regards
On 08/08/2016 03:14 PM, Petr Cech wrote:
On 08/08/2016 02:40 PM, Lukas Slebodnik wrote:
ehlo,
The simple reproducer is to use getent passwd,group with non-existing entry. Without the patch you will see that "/var/lib/sss/mc/passwd" is opened twice. Onece with mode "rw" opened by sssd_nss and the 2nd time with "r-" opened by sssd-client (getpwnam, getpwuid, getgrnam, getgrgid)
LS
Hello Lukas,
it looks good to me. I am waiting for CI tests.
Hi Lukas,
CI failed: http://sssd-ci.duckdns.org/logs/job/51/05/summary.html
On F24 during valgrind negcache-tests
http://sssd-ci.duckdns:.org/logs/job/51/05/fedora24/ci-build-debug/ci-make-c...
I didn't find more descriptive log. Is there any?
Regards
On (08/08/16 15:36), Petr Cech wrote:
On 08/08/2016 03:14 PM, Petr Cech wrote:
On 08/08/2016 02:40 PM, Lukas Slebodnik wrote:
ehlo,
The simple reproducer is to use getent passwd,group with non-existing entry. Without the patch you will see that "/var/lib/sss/mc/passwd" is opened twice. Onece with mode "rw" opened by sssd_nss and the 2nd time with "r-" opened by sssd-client (getpwnam, getpwuid, getgrnam, getgrgid)
LS
Hello Lukas,
it looks good to me. I am waiting for CI tests.
Hi Lukas,
CI failed: http://sssd-ci.duckdns.org/logs/job/51/05/summary.html
On F24 during valgrind negcache-tests
http://sssd-ci.duckdns:.org/logs/job/51/05/fedora24/ci-build-debug/ci-make-c...
I didn't find more descriptive log. Is there any?
Yes, http://sssd-ci.duckdns.org/logs/job/51/05/fedora24/ci-build-debug/src/tests/... http://sssd-ci.duckdns.org/logs/job/51/05/fedora24/ci-build-debug/src/tests/...
Anyway, attached is a patch which does not require changes in unit test.
LS
On 08/08/2016 04:09 PM, Lukas Slebodnik wrote:
On (08/08/16 15:36), Petr Cech wrote:
On 08/08/2016 03:14 PM, Petr Cech wrote:
On 08/08/2016 02:40 PM, Lukas Slebodnik wrote:
ehlo,
The simple reproducer is to use getent passwd,group with non-existing entry. Without the patch you will see that "/var/lib/sss/mc/passwd" is opened twice. Onece with mode "rw" opened by sssd_nss and the 2nd time with "r-" opened by sssd-client (getpwnam, getpwuid, getgrnam, getgrgid)
LS
Hello Lukas,
it looks good to me. I am waiting for CI tests.
Hi Lukas,
CI failed: http://sssd-ci.duckdns.org/logs/job/51/05/summary.html
On F24 during valgrind negcache-tests
http://sssd-ci.duckdns:.org/logs/job/51/05/fedora24/ci-build-debug/ci-make-c...
I didn't find more descriptive log. Is there any?
Yes, http://sssd-ci.duckdns.org/logs/job/51/05/fedora24/ci-build-debug/src/tests/... http://sssd-ci.duckdns.org/logs/job/51/05/fedora24/ci-build-debug/src/tests/...
Anyway, attached is a patch which does not require changes in unit test.
LS
Great,
this looks better to me. (LBTM?)
I am waiting for CI tests.
Regards
On 08/08/2016 04:20 PM, Petr Cech wrote:
On 08/08/2016 04:09 PM, Lukas Slebodnik wrote:
On (08/08/16 15:36), Petr Cech wrote:
On 08/08/2016 03:14 PM, Petr Cech wrote:
On 08/08/2016 02:40 PM, Lukas Slebodnik wrote:
ehlo,
The simple reproducer is to use getent passwd,group with non-existing entry. Without the patch you will see that "/var/lib/sss/mc/passwd" is opened twice. Onece with mode "rw" opened by sssd_nss and the 2nd time with "r-" opened by sssd-client (getpwnam, getpwuid, getgrnam, getgrgid)
LS
Hello Lukas,
it looks good to me. I am waiting for CI tests.
Hi Lukas,
CI failed: http://sssd-ci.duckdns.org/logs/job/51/05/summary.html
On F24 during valgrind negcache-tests
http://sssd-ci.duckdns:.org/logs/job/51/05/fedora24/ci-build-debug/ci-make-c...
I didn't find more descriptive log. Is there any?
Yes, http://sssd-ci.duckdns.org/logs/job/51/05/fedora24/ci-build-debug/src/tests/...
http://sssd-ci.duckdns.org/logs/job/51/05/fedora24/ci-build-debug/src/tests/...
Anyway, attached is a patch which does not require changes in unit test.
LS
Great,
this looks better to me. (LBTM?)
I am waiting for CI tests.
Regards
CI nearly passed: http://sssd-ci.duckdns.org/logs/job/51/07/summary.html The failure is not connected.
=> ACK
Regards
On Tue, Aug 09, 2016 at 07:54:15AM +0200, Petr Cech wrote:
On 08/08/2016 04:20 PM, Petr Cech wrote:
On 08/08/2016 04:09 PM, Lukas Slebodnik wrote:
On (08/08/16 15:36), Petr Cech wrote:
On 08/08/2016 03:14 PM, Petr Cech wrote:
On 08/08/2016 02:40 PM, Lukas Slebodnik wrote:
ehlo,
The simple reproducer is to use getent passwd,group with non-existing entry. Without the patch you will see that "/var/lib/sss/mc/passwd" is opened twice. Onece with mode "rw" opened by sssd_nss and the 2nd time with "r-" opened by sssd-client (getpwnam, getpwuid, getgrnam, getgrgid)
LS
Hello Lukas,
it looks good to me. I am waiting for CI tests.
Hi Lukas,
CI failed: http://sssd-ci.duckdns.org/logs/job/51/05/summary.html
On F24 during valgrind negcache-tests
http://sssd-ci.duckdns:.org/logs/job/51/05/fedora24/ci-build-debug/ci-make-c...
I didn't find more descriptive log. Is there any?
Yes, http://sssd-ci.duckdns.org/logs/job/51/05/fedora24/ci-build-debug/src/tests/...
http://sssd-ci.duckdns.org/logs/job/51/05/fedora24/ci-build-debug/src/tests/...
Anyway, attached is a patch which does not require changes in unit test.
LS
Great,
this looks better to me. (LBTM?)
I am waiting for CI tests.
Regards
CI nearly passed: http://sssd-ci.duckdns.org/logs/job/51/07/summary.html The failure is not connected.
=> ACK
* master: 950716d2087446205c84f00b371f468d6ead1ec2
sssd-devel@lists.fedorahosted.org