|
Date |
16.03.2010 |
Priority |
Normal |
Description
glite-AMGA_oracle, glite-AMGA_postgres, glite-BDII, glite-CONDOR_utils, glite-CREAM, glite-FTA_oracle, glite-FTM, glite-FTS_oracle, glite-HYDRA_mysql, glite-LB, glite-LFC_mysql, glite-LFC_oracle, glite-LSF_utils, glite-MON, glite-MPI_utils, glite-PX, glite-SCAS, glite-SE_dcache_admin_gdbm, glite-SE_dcache_admin_postgres, glite-SE_dcache_info, glite-SE_dcache_pool, glite-SE_dpm_disk, glite-SE_dpm_mysql, glite-SGE_utils, glite-TORQUE_client, glite-TORQUE_server, glite-TORQUE_utils, glite-UI, glite-VOBOX, glite-VOMS_mysql, glite-VOMS_oracle, glite-WMS, glite-WN, lcg-CE
The LB patch introduces a new version of glite-version.
glite-SCAS
New version of glite-security-scas
This is a drop-in replacement for glite-security-scas-0.2.6-1
fixing bug #52648, which prevents user banning for the SCAS server.
And the stability of the service has been improved significantly
when probed by network analysis tools.
LCAS
- Solved segmentation faults when a malformed proxy was
provided by the calling library or application.
- When using the lcas_pem interface (used by gLExec, SCAS
and third parties) a wrong individual certificate was
selected. It was first delegation that was selected and not
the final delegation of the certificate chain.
- The extraction of the user's subject DN has been
replaced. The Globus code calculates the RDN count of the
individual certificate and strips of the amount of RDNs equal
to the amount of delegations. This process is error prone,
causes segmentation.faults when used in a wrong way, overly complex.
It's replaced by a safer approach which has been used in
LCMAPS for years.
SCAS
The saml2-xacml2-c-lib code implements both the socket setup,
socket handling and the protocol handling (HTTP, SOAP, SAML2,
XACML2). It usually also does the TCP/IP, but by using callbacks
in the library this is extended to implement TCP/IP with SSL. The
saml2-xacml2-c-lib setups up the master socket in the default
blocking mode. This means that the multi-threaded nature of the
service boils down to the semaphore on the blocking socket. This
will change in a non-blocking socket with many worker process in
the next release.
Fix the following situations in SCAS:
- Failures in the TCP/IP are handled better
- 500ms wait time between a successful TCP/IP connection (meaning fully
setup and functional) and the initiation of the SSL handshake
initiation.
- nmap -sV $hostname is fuzzing with the SSL to
identify the service. This causes a crash, because the callbacks
made a proper cleanup of the client connection at the service, but
the saml2-xacml2-c-lib also seems to want to perform the close(2)
on the socket in the error handling deep in SOAP layers.
- The previous situation also occurs when a proxy of the pilot job
framework executor (WN scenario) or the certificate of the Grid
service is blacklisted in SSL+LCAS. (note: This is another LCAS
call that authorizes the pilot job (real) users).
- The SCAS log lines are prefixed with the time that an event has taken place. The
time stamp is set per thread, so a complete thread handling will
all share the timecode. This will ease the automated search
ability.
Platform detailed dependency:
For Glite 3.1:
- sl4 32 bit: depends on saml2-xacml2-c-lib 0.0.14_2 (providing the drop-in replacement for Production)
- sl4 64 bit: depends on saml2-xacml2-c-lib 0.0.15_2
For Glite 3.2:
- sl5 64 bit: depends on saml2-xacml2-c-lib 0.0.15_2
This update fixes various bugs. For the full list of bugs, please see list below.
Fixed bugs
Number | Description |
#35770 |
Wrong parsing of the LB events which contains the string ending with '' in field DG.USERTAG.VALUE |
#48163 |
interlogger files may grow indefinitely |
#52648 |
[SCAS] Userban fails due to incorrect construction of pemstring |
#53524 |
SCAS: Denial of Service on SCAS daemon |
#55482 |
LB server version not published correctly |
Updated rpms
The RPMs can be updated using yum via
Service reconfiguration after update
Not needed.
Service restart after update
Service must be restarted.
How to apply the fix
- Update the RPMs (see above)
- Update configuration (see above)
- Restart the service if necessary (see above)
|