CVE-2018-7262

In Ceph before 12.2.3 and 13.x through 13.0.1, the rgw_civetweb.cc RGWCivetWeb::init_env function in radosgw doesn't handle malformed HTTP headers properly, allowing for denial of service.

In Ceph before 12.2.3 and 13.x through 13.0.1, the rgw_civetweb.cc RGWCivetWeb::init_env function in radosgw doesn't handle malformed HTTP headers properly, allowing for denial of service.

In Ceph before 12.2.3 and 13.x through 13.0.1, the rgw_civetweb.cc RGWCivetWeb::init_env function in radosgw doesn't handle malformed HTTP headers properly, allowing for denial of service.

In Ceph before 12.2.3 and 13.x through 13.0.1, the rgw_civetweb.cc RGWCivetWeb::init_env function in radosgw doesn't handle malformed HTTP headers properly, allowing for denial of service.

In Ceph before 12.2.3 and 13.x through 13.0.1, the rgw_civetweb.cc RGWCivetWeb::init_env function in radosgw doesn't handle malformed HTTP headers properly, allowing for denial of service.

This update for ceph fixes the following issues:

Security issues fixed:

• CVE-2018-7262: rgw: malformed http headers can crash rgw (bsc#1081379).
• CVE-2017-16818: User reachable asserts allow for DoS (bsc#1063014).

Bug fixes:

• bsc#1061461: OSDs keep generating coredumps after adding new OSD node to cluster.
• bsc#1079076: RGW openssl fixes.
• bsc#1067088: Upgrade to SES5 restarted all nodes, majority of OSDs aborts during start.
• bsc#1056125: Some OSDs are down when doing performance testing on rbd image in EC Pool.
• bsc#1087269: allow_ec_overwrites option not in command options list.
• bsc#1051598: Fix mountpoint check for systemctl enable --runtime.
• bsc#1070357: Zabbix mgr module doesn't recover from HEALTH_ERR.
• bsc#1066502: After upgrading a single OSD from SES 4 to SES 5 the OSDs do not rejoin the cluster.
• bsc#1067119: Crushtool decompile creates wrong device entries (device 20 device20) for not existing / deleted OSDs.
• bsc#1060904: Loglevel misleading during keystone authentication.
• bsc#1056967: Monitors goes down after pool creation on cluster with 120 OSDs.
• bsc#1067705: Issues with RGW Multi-Site Federation between SES5 and RH Ceph Storage 2.
• bsc#1059458: Stopping / restarting rados gateway as part of deepsea stage.4 executions causes core-dump of radosgw.
• bsc#1087493: Commvault cannot reconnect to storage after restarting haproxy.
• bsc#1066182: Container synchronization between two Ceph clusters failed.
• bsc#1081600: Crash in civetweb/RGW.
• bsc#1054061: NFS-GANESHA service failing while trying to list mountpoint on client.
• bsc#1074301: OSDs keep aborting: SnapMapper failed asserts.
• bsc#1086340: XFS metadata corruption on rbd-nbd mapped image with journaling feature enabled.
• bsc#1080788: fsid mismatch when creating additional OSDs.
• bsc#1071386: Metadata spill onto block.slow.

This update for ceph to 12.2.5-407-g5e7ea8cf03 fixes the following issues:

Security issue fixed:

• CVE-2018-7262: The rgw_civetweb.cc RGWCivetWeb::init_env function in radosgw doesn't handle malformed HTTP headers properly, allowing for denial of service. rgw: make init env methods return an error (bsc#1081379)

Other issues fixed:

• osd: do not crash on empty snapset (bsc#1074301)
• mon: add 'ceph osd pool get erasure allow_ec_overwrites' command (bsc#1087269)
• journal: limit number of appends sent in one librados op (bsc#1086340)
• RGW user stats fixes (bsc#1087493)
• rgw openssl fixes (bsc#1079076, bsc#1081379)
• rocksdb: fixes early metadata spill over to slow device in bluefs (bsc#1071386)
• mon: reenable timer to send digest when paxos is temporarily inactive (bsc#1070357)
• fsid mismatch when creating additional OSDs (bsc#1080788)
• crash in civetweb/RGW (bsc#1081600)

This update for ceph fixes the following issues:

Security issues fixed:

• CVE-2018-10861: Fix ceph-mon authorization on OSD pool ops (bsc#1099162).
• CVE-2018-1128: Fix cephx signature check bypass (bsc#1096748).
• CVE-2018-1129: Fix cephx protocol vulnerability to replay attack (bsc#1096748).
• CVE-2018-7262: Fix malformed http headers that can crash rgw (bsc#1081379).

Bug fixes:

• bsc#1072512: multipart uploads are broken if the bucket has been resharded
• bsc#1080112: rgw: user stats increased after bucket reshard
• bsc#1086340: SES5: XFS metadata corruption on rbd-nbd mapped image with journaling feature enabled