[PATCH] Add support for the EntryCacheNoWaitRefreshTimeout
by Stephen Gallagher
This timeout specifies the lifetime of a cache entry before it is
updated out-of-band. When this timeout is hit, the request will
still complete from cache, but the SSSD will also go and update
the cached entry in the background to extend the life of the
cache entry and reduce the wait time of a future request.
Support for the EnumCacheNoWaitRefreshTimeout is still forthcoming, but
I wanted to get a formal review on this portion.
--
Stephen Gallagher
RHCE 804006346421761
Looking to carve out IT costs?
www.redhat.com/carveoutcosts/
14 years, 8 months
[PATCH] Cleanups for library linking
by Sumit Bose
Hi,
I think the LDAP provider should link agains a LDAP library. So far it
only worked, because sssd_be exports all symbols and links against
libldb which links against libldap.
bye,
Sumit
14 years, 8 months
[PATCH] more fixes for older libpcre versions
by Sumit Bose
Hi,
older versions of libpcre only support the Python naming style,
?P<name>, for subpatterns. This patch changes our default pattern and
adds some hints about this.
bye,
Sumit
14 years, 8 months
[PATCH] Correcting issues
by Dmitri Pal
Hi,
This patch corrects (hopefully last) set of issues found in the earlier
patches during review.
The ones that are not addressed yet are tracked via explicit tickets or
via FIXME
comments in the code.
The whole stack of the pending patches is:
* ELAPI sinks and providers
* ELAPI Adding file provider and CSV format
* ELAPI Laying foundation for the async processing
* COLLECTION Copy collection flat with concatenated names
* ELAPI Correcting issues
--
Thank you,
Dmitri Pal
Engineering Manager IPA project,
Red Hat Inc.
-------------------------------
Looking to carve out IT costs?
www.redhat.com/carveoutcosts/
14 years, 8 months
[PATCH] ELAPI Adding file provider and CSV format
by Dmitri Pal
Hi,
This patch is next significant chunk of code.
See patch comments for the main description.
It requires previous patch that has been reviewed but was not pushed yet.
It addresses some of the suggestions risen during the review.
I also added couple tickets: to eventually remove all the "Fixme"
comments and to refactor places where I allocated context structs and
initialize them.
Checked with make check, valgrind and did in tree and parallel builds.
All works, no leaks.
Sorry that it is too big.
I need to move on and there are many things that are still missing.
I will start working on the next set of the functionality tomorrow
mainly adding async processing. That would definitely affect some of the
arguments passed all along so expect changes to many modules.
--
Thank you,
Dmitri Pal
Engineering Manager IPA project,
Red Hat Inc.
-------------------------------
Looking to carve out IT costs?
www.redhat.com/carveoutcosts/
14 years, 8 months
[PATCH] Split database file to one per domain
by Simo Sorce
Hello list.
So far we have been using a single database file for all our caches as
well as for the local domain.
Initially I used a single database because I thought we could optimize
some search queries when it came to enumerations. That didn't prove true
and in effect we are doing always separate searches for different
domains (in fact the patch was rather simple to write from this pov as
no searches needed to be touched at all).
Using a single database file for all domains is otherwise a
disadvantage.
>From a pure "security" point of view each backend has now it's own
database and cannot "screw up" other domains data.
Each domain access its backend separately so there is less contention
when you want to start a transaction.
The various files are clearly separated, so you can easily just delete
all the ones named cache-NAME.ldb and not risk deleting the persistent
data in sssd.ldb
The patch includes also upgrade routines to automatically backup and
convert the old file into split files, so switching to the new code is
painless.
All the code was converted to handle the new list of database
structures. One part that felt odd was the sss_tools. They should really
only operate on the 'local' domain and nothing else. So attaches is also
a patch that basically reverts most of the changes and just opens the
'local' database instead. I have not merged them in a single patch
because I didn't want to mix format changes with functional changes and
because apparently we may want to be a bit more radical and eliminate
also the shadow-utils handling at the same time.
I have tested the change and the upgrade as well as starting from
scratch and all seem to work properly.
Simo.
--
Simo Sorce * Red Hat, Inc * New York
14 years, 8 months
[PATCH] Remove shadow-utils support from tools
by Jakub Hrozek
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
The attached patch applies on top of Simo's two patches.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org/
iEYEARECAAYFAkqmZcgACgkQHsardTLnvCWRbgCePRy9eDg32BZIW0DVu4NYJNGk
5qAAn2rm0OXo45s9Q75nAi4cF6hmdmnw
=kDDk
-----END PGP SIGNATURE-----
14 years, 8 months