[TS-644] http_ui cache lookup interface will crash traffic_server in full cluster mode

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.1.5
• Fix Version/s: 2.1.6
• Component/s: Cache, Clustering

Environment

full cluster mode and http_ui enabled

• Assignee: Zhao Yongming
• Reporter: Zhao Yongming
• Created: 16/Jan/11 07:43
• Updated: 03/Mar/11 17:24
• Resolved: 11/Feb/11 20:58

Description

when I use http_ui cache lookup interface to lookup objects in cache, with "http://cdn.zymlinux.net/cache/lookup_url?url=http%3A%2F%2Fsource.zymlinux.net%2Ftrafficserver%2Fts75.png", it will crash traffic_server process:

[Jan 16 14:40:46.089] Server

{47005498951424} DEBUG: (http) [3] State Transition: HTTP_API_POST_REMAP -> PROXY_INTERNAL_REQUEST
[Jan 16 14:40:46.089] Server {47005498951424}

DEBUG: (cache_inspector) query params: url=http%3A%2F%2Fsource.zymlinux.net%2Ftrafficserver%2Fts75.png [unescaped]
[Jan 16 14:40:46.089] Server

{47005498951424} DEBUG: (cache_inspector) query params: url=http://source.zymlinux.net/trafficserver/ts75.png [escaped]
[Jan 16 14:40:46.089] Server {47005498951424}

DEBUG: (cache_inspector) there were 1 url(s) passed in
[Jan 16 14:40:46.089] Server

{47005498951424} DEBUG: (cache_inspector) URL 1: http://source.zymlinux.net/trafficserver/ts75.png
[Jan 16 14:40:46.117] Server {47005498951424}

DEBUG: (cache_msg) do_op opcode=1 seqno=7 Machine=0x11fac90 data=0x52113b10 datalen=75 mio=0x0
[Jan 16 14:40:46.117] Server

{47005498951424} DEBUG: (cache_proto) open_local-s (R) seqno=7
[Jan 16 14:40:46.117] Server {47005498951424}

DEBUG: (cluster_trace) alloc_channel local chan=7 VC=0x54d66620
[Jan 16 14:40:46.117] Server

{47005498951424} DEBUG: (cluster_trace) VC start alloc local chan=7 VC=0x54d66620
[Jan 16 14:40:46.117] Server {47005498951424}

DEBUG: (cache_proto) 0open_local-s (R) success, seqno=7 chan=7 token=-308068854,3 VC=0x54d66620
[Jan 16 14:40:46.140] Server

{47005657089792}

DEBUG: (cluster_trace) complete_channel_read chan=7 len=9520
NOTE: Traffic Server received Sig 11: Segmentation fault
/usr/bin/traffic_server - STACK TRACE:
[Jan 16 14:40:46.158] Manager

{139779918772000} FATAL: [LocalManager::pollMgmtProcessServer] Error in read (errno: 104)
[Jan 16 14:40:46.158] Manager {139779918772000}

FATAL: (last system error 104: Connection reset by peer)
[Jan 16 14:40:46.158] Manager

{139779918772000} NOTE: [LocalManager::mgmtShutdown] Executing shutdown request.
[Jan 16 14:40:46.159] Manager {139779918772000}

NOTE: [LocalManager::processShutdown] Executing process shutdown request.
[Jan 16 14:40:46.159] Manager

{139779918772000} ERROR: [LocalManager::sendMgmtMsgToProcesses] Error writing message
[Jan 16 14:40:46.159] Manager {139779918772000}

ERROR: (last system error 32: Broken pipe)
[E. Mgmt] log ==> [TrafficManager] using root directory '/usr'
[Jan 16 14:40:46.179]

{139723850565408} NOTE: updated diags config
[Jan 16 14:40:46.182] Manager {139723850565408}

NOTE: [ClusterCom::ClusterCom] Node running on OS: 'Linux' Release: '2.6.36-gentoo-zym'
[Jan 16 14:40:46.183] Manager

{139723850565408} NOTE: [LocalManager::listenForProxy] Listening on port: 80
[Jan 16 14:40:46.183] Manager {139723850565408}

NOTE: [TrafficManager] Setup complete
[Jan 16 14:40:47.183] Manager

{139723786659584}

NOTE: adding node 10.62.163.238 to the cluster
[Jan 16 14:40:47.199] Manager

{139723850565408}

NOTE: [LocalManager::startProxy] Launching ts process