[
https://issues.apache.org/jira/browse/TS-3942?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Brian Geffon reassigned TS-3942:
--------------------------------
Assignee: Brian Geffon
> Crash when enumerating HostDB records in HostDB UI
> --------------------------------------------------
>
> Key: TS-3942
> URL: https://issues.apache.org/jira/browse/TS-3942
> Project: Traffic Server
> Issue Type: Bug
> Components: HostDB
> Reporter: Brian Geffon
> Assignee: Brian Geffon
>
> When using the show all feature of hostdb ui it will periodically crash when
> enumerating round robin dns records. This is because of a new permanent
> hostname field that I added a while back so that we have the hostname of all
> hostdb records. The problem is that when allocating each HostDBInfo structure
> for the round robin entry the code doesn't currently memset() the entire
> structure (it does in the SRV record case but for some reason not the RR
> case). This means that in certain situations if we were unlucky the memory
> would be initialized to random memory and at some point down the road it
> would result in a crash in UnsunkPtrRegistry::ptr as seen in the stack trace
> below.
> #0 0x00000000006d91f3 in UnsunkPtrRegistry::ptr (this=..., i=...) at
> MultiCache.cc:1314
> #1 0x00000000006d922d in UnsunkPtrRegistry::ptr (this=..., i=...) at
> MultiCache.cc:1318
> #2 0x00000000006d9380 in MultiCacheBase::ptr (this=..., poffset=...,
> partition=...) at MultiCache.cc:1341
> #3 0x00000000006d000a in HostDBInfo::perm_hostname (this=...) at
> HostDB.cc:2297
> #4 0x00000000006d346e in ShowHostDB::showOne (this=..., r=..., rr=...,
> event=..., e=..., hostdb_rr=...) at HostDB.cc:2505
> #5 0x00000000006d2982 in ShowHostDB::showAllEvent (this=..., event=...,
> e=...) at HostDB.cc:2407
> #6 0x000000000050d0a6 in Continuation::handleEvent (this=..., event=...,
> data=...) at ../iocore/eventsystem/I_Continuation.h:146
> #7 0x00000000006cb8fe in HostDBContinuation::iterateEvent (this=...,
> event=..., e=...) at HostDB.cc:1087
> #8 0x000000000050d0a6 in Continuation::handleEvent (this=..., event=...,
> data=...) at ../iocore/eventsystem/I_Continuation.h:146
> #9 0x000000000078c3ca in EThread::process_event (this=..., e=...,
> calling_code=...) at UnixEThread.cc:144
> #10 0x000000000078c6e5 in EThread::execute (this=...) at UnixEThread.cc:223
> #11 0x000000000053e6a2 in main (argv=...) at Main.cc:1765
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
