5 - RHSA-2010:0773

Problem Description:

Updated packages that fix two security issues, several bugs, and add multiple enhancements are now available as part of the ongoing support and maintenance of Red Hat Enterprise MRG Messaging and Grid for Red Hat Enterprise Linux 5.

The Red Hat Security Response Team has rated this update as having moderate security impact. Common Vulnerability Scoring System (CVSS) base scores, which give detailed severity ratings, are available for each vulnerability from the CVE links in the References section.

2. Relevant releases/architectures:

MRG Grid Execute Node for RHEL 5 Server - i386, noarch, x86_64 MRG Grid for RHEL 5 Server - i386, noarch, x86_64 MRG Management for RHEL 5 Server - i386, noarch, x86_64 MRG Realtime for RHEL 5 Server - noarch Red Hat MRG Messaging Base for RHEL 5 Server - i386, noarch, x86_64 Red Hat MRG Messaging for RHEL 5 Server - i386, noarch, x86_64

3. Description:

Red Hat Enterprise MRG (Messaging, Realtime, and Grid) is a realtime IT infrastructure for enterprise computing. MRG Messaging uses Apache Qpid to implement the Advanced Message Queuing Protocol (AMQP) standard, adding persistence options, kernel optimizations, and operating system services.

This update moves Red Hat Enterprise MRG to version 1.3.

A flaw was found in the way Apache Qpid handled the receipt of invalid AMQP data. A remote user could send invalid AMQP data to the server, causing it to crash, resulting in the cluster shutting down. (CVE-2009-5005)

A flaw was found in the way Apache Qpid handled a request to redeclare an existing exchange while adding a new alternate exchange. If a remote, authenticated user issued such a request, the server would crash, resulting in the cluster shutting down. (CVE-2009-5006)

This update also adds the following enhancements:

* This update introduces a protocol-independent C++ API. The extra layer of indirection will make it easy to support new versions of the AMQP protocol, as well as multiple versions simultaneously. (BZ#497747)

* The management component is now capable of working in a cluster. (BZ#501015)

* The Messaging Client Python API is now protocol-independent. (BZ#497748)

* This update allows a JMS client to subscribe to the failover exchange to retrieve cluster membership information and subsequently to receive updates. (BZ#483753)

* With this update, the qpidd service can be run without additional authentication options. (BZ#515513)

* This update adds an OpenMPI wrapper script to Condor. It adds support for OpenMPI jobs. (BZ#537232)

* The Messaging Client Python API now provides a failover mechanism for clustered brokers. (BZ#495718)

* The Python Messaging API now includes support for Simple Authentication and Security Layer (SASL), which allows authentication support to be added to connection-based protocols. (BZ#548493)

* The qpid-tool is now able to determine which session a queue consumer belongs to. (BZ#504325)

* This update handles backward/forward compatibility for QMF and its components. (BZ#506698)

* Both Secure Sockets Layer (SSL) and Remote Direct Memory Access (RDMA) entries can now appear in the list of known URLs. (BZ#471632)

* This update allows for the scheduler daemon to run without swap. (BZ#548090)

* This update introduces a mechanism that specifies the queue size of a queue that is setup via the Java API. (BZ#534008)

* Previously, a collector could not be remotely restarted. With this update, the restart is possible and works as expected. (BZ#543021)

* The usage information for the qpid-config utility (that is, the output of the "qpid-config -h" command) has been updated to include a brief explanation of the exchange type. (BZ#506420)

These updated packages include many other bug fixes and enhancements. Users are directed to the Red Hat Enterprise MRG 1.3 Technical Notes for information on these changes:

http://docs.redhat.com/docs/en-US/Red_Hat_Enterprise_MRG/1/html/Technical_N otes/index.html

All Red Hat Enterprise MRG users are advised to upgrade to these updated packages, which resolve these issues and add these enhancements, as well as resolving the issues and adding the enhancements noted in the Red Hat Enterprise MRG 1.3 Technical Notes. After installing the updated packages, the qpidd service must be restarted ("service qpidd restart") for this update to take effect.

4. Solution:

Before applying this update, make sure all previously-released errata relevant to your system have been applied.

This update is available via the Red Hat Network. Details on how to use the Red Hat Network to apply this update are available at http://kbase.redhat.com/faq/docs/DOC-11259

5. Bugs fixed (http://bugzilla.redhat.com/):

445749 - [python client] kerberos based authentication 452546 - No way to determine if session/connection is established 455318 - A tx commit fails without a proper error message when a queue runs out of capacity 456482 - submit -spool and transfer_executable = false 458344 - Messages are not released on rollback 462461 - Clustering broker fail-over must replicate federation links 469919 - qpidd init script over-rides user option settings. 470080 - Cluster integration with security. 471054 - focus linking of gsoap, X11 and pq into daemons and tools 471286 - Grid Statistics Job Activity Graphics doesn't update correctly 471315 - Grid, Parse error on Hold a job reason entry. 471326 - Grid: It appears that the default for jobs is to show up as held in the boxed graphic 471632 - Add support for SSL/RDMA URLs in cluster's know urls list 479031 - Cluster member can't be added while management session open 479326 - cluster broker crashes with race condition in DispatchHandle 482944 - Management messages can get staged - which breaks management 483666 - Dynamic Slots and STARTD_JOB_EXPRS, invalid attribute name 483753 - Add failover exchange support for the java client 484048 - qpidd+store flush() failed: jexception 0x0106 slock::slo ck() threw JERR__PTHREAD: pthread failure. (pthread_mutex_lock failed: errno=22 (Invalid argument)) (MessageStoreImpl.cpp:1331) 485091 - "Unknown Publisher" when installing Windows grid client 485429 - qpidd stopped by critical Broker start-up failed: Cannot lock ... Resource temporarily unavailable 486595 - condor_configuration_node input validation 486779 - [RFE] configurable sesame publish rate (sesame publishing too often) 487023 - UID&FILESYSTEM_DOMAIN mis-configuration causing unintended side-effects 488942 - c++ client aborts when session and connection not closed 489315 - perftest shutdown seems to be not clean 'Error in shutdown: Connection closed' 489537 - Cluster - Bogus(?) messages in log file when a new broker joins a cluster 489540 - Memory leak in SASL client code. 490170 - qpidd init script does not implement condrestart though the rpm has a script to call it 490855 - clustered qpidd segfaults in qpid::broker::Exchange::propagateFedOp 491203 - "Timed out waiting for daemon" if recovery from journal takes a long time 491305 - clustered qpidd - replicating non-acked messages is not made visible for managent tools qpid-tool/cumin 491313 - Subscribing sessions should be terminated with exception if the queue they are subscribed to is deleted 492334 - qpidd+store startup crash in mrg::msgstore::MessageStoreImpl::init() 493710 - condor_configure_node: delete not inverse of add 494393 - First two nodes join 'simultaneously'; no node can reach the 'ready' state. 494399 - Bindings from durable queues to the default exchange are not shown after restore 494651 - sesame README points to old apache SVN location 495718 - Python client needs to have failover for clustering 497747 - Feature: Protocol independent API for c++ 497748 - Feature: Protocol independent API for python 498056 - SASL/GSSAPI - Connection hangs when GSSAPI context expires 498247 - CLI utilities display Python back-traces in some error cases 500712 - QMF queries to the broker may return records for deleted objects 500779 - Feature: Provide access to the Connection a Session corresponds to 501015 - Management and cluster do not work together. 501305 - Cluster node gets stuck as updatee and 'hangs' cluster 501749 - If an XML exchange is declared durable, the broker crashes on recovery 504000 - qpid-config's altern-ex option doesn't work 504325 - Enhancement: it should be possible to determine through qpid-tool which sessions a queues consumers belong to 504691 - alternate-exchange proprty of exchange and queue are not persisted 505287 - Messages with no content that 'flow to disk' result in protocol errors on delivery 505314 - qpid-tool crashes down after input wrong command list query 505923 - dedicated scheduler may be inappropriately reusing claims 506420 - qpid-config -h does not explain exchange type 506553 - sesame - memory bloat over time 506556 - c++ client may not timeout accurately where multiple connections exist in the process 506698 - Handle backward/forward compatibility for QMF 507363 - clustered qpidd fails to start - gather loop causes openais_dispatch_recv() to block 507413 - Broker with single IO thread gets stuck looping if it runs out of file handles 507421 - Cluster flow control does not appear to be working properly. 507538 - method exchange_declare is missing in ruby qpid session class 507586 - qpid-config ends with failure 508137 - C++ QMF agent not connecting to broker under valgrind 508144 - A broker stopped and restarted does not remember 'redelivered' status correctly 508675 - Unresponsive qpidd process hangs the cluster 508959 - Attempt to propagate binding info over dynamic link can crash broker if link is concurrently destroyed 509395 - The JMS Client does not default to the correct priority as specified in the spec 509437 - Failure in failover_soak 509449 - JMS client releases messages in an unpredictable order on recover 509454 - [RFE] Add validation for the '--cluster-url' qpidd option 509800 - If journal capacity is exceeded as a result of cluster-durable mode being invoked, last man standing exits 509892 - byte credit calculation inconsistent for messages transfered to new joiner 510241 - clustered qpidd crash in qpid::sys::Poller::run() 510475 - clustered qpidd startup - abort because of unhandled exception 510583 - Unhandled exception when running qpid-cluster against a standalone broker. 510747 - Out of Bounds exception when sending large QMF response 511066 - Replication exchange type should record the usual management stats 511292 - Unexpected connection shows up for qpid-stat -c 513426 - string to double conversion results in questionable precision 513641 - qpid-config gives error "Failed: ()" when creating persistent queue 514054 - [store] Journal can fill under some conditions, and recover from full condition not possible 514751 - QMF agent logging to file, no stdout 515513 - Make cluster update work out of the box without special authentication options 517836 - exclusive parameter ignored in JMS url binding, if durable attribute is present 518291 - Python management tools must handle SystemExit exception properly 518394 - Creating durable and cluster-durable queue which has bad --file-count and/or --file-size parameter causes an exception only for first time 518872 - [FEATURE] exchange flag auto-delete is not recognizable 519183 - Matchmaker code doesn't implement fair share correctly 519476 - Invalid accept data sent by Java client after failover. 519505 - Broker strips domain from userID, causes mismatch on GSSAPI id checking 520600 - Intermittent leak in client library, connector thread not joined. 522267 - Windows: Qpid C++ pid_t and ssize_t 3rd-party compat 526299 - the clustered broker seems to sometimes not send a close-ok before shutting down the socket 526680 - Exchanges named "amq." are declarable, but amqp spec 527233 - shadow process bloat 529670 - qpid-config - inappropriate error message if trying to authenticate with non-existing user 530594 - restart of libvirtd causes condor_vm-gahp to hang. 531561 - alternate exchange not visible on a queue via QMF 531833 - FailoverExchangeMethod getNextBrokerDetails() loops infinitely after a total cluster failure or if the inital connect node is down 531837 - Java client should set the process ID in the client properties during Connection open 531842 - When kerberos auth is used, Java client should use the kerberos user_id & domain when setting the user_id in messages 533045 - Feature Request: support for SASL EXTERNAL with TLS/SSL 533173 - --max-connectoins has no effect 534008 - Need mechanism to specify the queue size of a queue that is setup via the Java API. 537232 - PU: need OpenMPI wrapper script 537481 - qpid-stat needs option to link sesion to queue via subscription object 538188 - connection.start() hangs if connection is not accepted 540545 - WANT_SUSPEND evaluating to UNDEFIEND causes condor_startd exception 541927 - Persistent cluster problems after reboot -f 543021 - Unable to restart collector with condor_restart on remote node 543524 - Cluster with --cluster-size should not hold up init scripts. 543560 - VM Universe libvirt script issues 544092 - message store should not delete backups when qpidd starts 544306 - clustered broker does not retry CPG calls that return TRY_AGAIN 545436 - Cluster node shutsdown with inconsistent error 546736 - Schedd performs unnecessary file operations on SPOOL, targeting mpp.X.Y files 546770 - condor_schedd performance, job removal fsync for each job 547295 - qpid-stat -b threading exception during shutdown 'exceptions.TypeError: 'NoneType' object is not callable' 547397 - Compile with -O2 547769 - clustered qpidd: qpid-cluster/qpid-stat -b reports different widths on different nodes while replication is working well on all nodes 548090 - RESERVED_SWAP doesn't default to 0 as stated in docs 548137 - TIMEOUT_MULTIPLIER only available in _ form 548493 - SASL support missing for Python messaging API 549389 - condor_master -pidfile will stomp pidfile of running master 549432 - Parallel Universe jobs require job spool directory 549443 - qpid-config cannot create bindings for the XML or Headers exchange types 549956 - Clustered broker crashes with inconsistency error 552330 - qpid-config from trunk causes exception in broker 552407 - classad debug() function doesn't work with IfThenElse 554980 - [qpidd+store] broker rarely aborts when stressed by perftest 555716 - [qpidd+store] broker rarely segfaults when stressed by perftest 556351 - clustered qpidd - durable exchanges do not survive cluster restart. 557159 - Queue-Purge does not send messages to alternate-exchange 557896 - The ttl of messages is not adjusted when forwarding on to other brokers in a federation. 558526 - clustered qpidd shutdowns during start-up with 'Authentication failed: SASL(-1): generic failure: Unable to find a callback: 32775' 558864 - JMS_QPID_DESTTYPE is not set making getJMSDestination unusable. 558968 - initscript lsb compliance 559014 - clustered qpid: durable exchange state not replicated to broker joining cluster 559071 - VM_MEMORY handled inconsistently between Startd and VMGahp 559625 - Segfault if FailoverManager is closed before being opened. 560005 - Broker options "--auth" and "--require-encryption" can fail when used with SSL/TLS 561955 - PREPARE hook invoked as condor, not as user. cannot access $PWD. 561958 - PREPARE hook invocation failure does not abort job execution 565618 - condor_submit fsync()s UserLog for each job 566825 - Grid with no slots throws exception in MRG Management Console 568502 - Collector should advertise itself immediately 568661 - JMS client does not verify that the hostname connected to matches that specified in the servers certificate 568718 - Is acl reload safe to use? 568838 - Dynamic federation duplicates messages 568863 - Dynamic federation tears links down incorrectly 570756 - DtxSetTimeout sent after XID has already been committed 572574 - Error reported from execute node incomplete for IWD access failure 572668 - Potential shadow/schedd protocol error 575147 - condor_master can't start additional schedd's without a restart 575150 - Need to be able to configure maximum cluster id 575177 - Messages set with a TTL expire immediately when sent on qpid queues with LVQ ordering 575748 - broker exits with "critical Broker start-up failed: St9bad_alloc" when ran with --worker-threads 0|-1 575777 - scheduler universe jobs can start during schedd shutdown 575784 - improper RELEASE_CLAIM after REQUEST_CLAIM rejection 576693 - qpid-cluster -d does not close the client connection 578216 - condor_schedd reuses claims to partitioned slots inappropriately 578600 - Dyanamic Slot INVALIDATE_STARTD_ADS causes collector pegging 579681 - Topic exchange duplicates messages 582366 - When reloading a large acl file , the broker core dumps 583131 - Fix Java Client logging 583526 - Management methods disallowed in Clusters must be re-enabled 584089 - ACL module core dumps if management is disabled 591292 - MRG-M Heartbeat causes core 592861 - Recovered messages larger than 65523 bytes result in framing violation 597362 - Sporadic failure of check-long in cluster_tests.py test_failover 601828 - QMF Agent returning STATUS_USER returns error 7 to QMF Console 603201 - condor-7.4.3-0.17.el5 postuninstall uses invalid init script option 603839 - Concurrent tagging of message with trace id while message is delivered from another queue causes segfault 605311 - condor_schedd double free on SOAP transaction timeout 606824 - Acquired but Not Accepted Messages Not Sent to Alterntate Exchange 614993 - Using Memory or RequestMemory in job requirements drops both default RequestMemory and Memory requirements 615313 - condor_chirp fails when querying the value of a non-existing attribute 615492 - starter hooks, HOOK_UPDATE_JOB_INFO and HOOK_JOB_EXIT not run as job owner 615504 - condor_chirp relies on getenv("_CONDOR_SCRATCH_DIR") 615510 - Job hooks environment does not contain _CONDOR_SCRATCH_DIR and the like 615633 - condor_chirp get_job_attr can return garbage 617709 - fix hfs accountant stats 619552 - negotiator hfs incorrect remaining and infinite loop 621902 - Permissions not set correctly on key pair file 623684 - condor_userlog core dumps when unable to open log file r/o 625205 - shadows create a spool directory per job 628034 - negotiator core on quota_dynamic =0 628086 - GROUP_DYNAMIC_MACH_CONSTRAINT unused with HFS 642373 - CVE-2009-5005 qpid: crash on receipt of invalid AMQP data 642377 - CVE-2009-5006 qpid: crash when redeclaring the exchange with specified alternate_exchange