Occasional zone crash in EQNetworkConnection::IncomingARSP

[Derision]

Does anyone else get the occasional zone crash like this ?

#0 0x08077dd6 in EQNetworkConnection::IncomingARSP(unsigned short) (
this=0x8506818, arsp=23006) at ../common/EQNetwork.cpp:1278
1278 if (arsp >= iterator.GetData()->dwARQ) {
(gdb) bt
#0 0x08077dd6 in EQNetworkConnection::IncomingARSP(unsigned short) (
this=0x8506818, arsp=23006) at ../common/EQNetwork.cpp:1278
#1 0x08077aa9 in EQNetworkConnection::ProcessPacket(EQNetworkPacket *, bool) (
this=0x8506818, pack=0x85b3e50, from_buffer=false)
at ../common/EQNetwork.cpp:1069
#2 0x08077999 in EQNetworkConnection::RecvData(unsigned char*, unsigned) (
this=0x8506818, data=0x44764480 "", size=10)
at ../common/EQNetwork.cpp:1014
#3 0x0807505b in EQNetworkServer::RecvData(unsigned char*, unsigned, unsigned,
unsigned short) (this=0x81d7300, data=0x44764480 "", size=10, irIP=2570232856,
irPort=51972) at ../common/EQNetwork.cpp:280
#4 0x08074e15 in EQNetworkServer::Process() (this=0x81d7300)
at ../common/EQNetwork.cpp:222
#5 0x08075223 in EQNetworkServerLoop(void*) (tmp=0x81d7300)
at ../common/EQNetwork.cpp:324
#6 0x402cc7d3 in start_thread () from /lib/tls/libpthread.so.0
#7 0x40248b4a in clone () from /lib/tls/libc.so.6
(gdb)

Also, whenever a single zoneserver crashes it seems that although I get one core dump, all the zones eventually go bye-bye with messages in world.log like this:

Removing zoneserver from ip:0.0.0.0 port:0 (xxx.yyyy.co.uk:8000)
Hold Zones mode is ON - rebooting lost zone

Linux version is Mandrake 10, Kernel 2.4.25, GCC version 3.3.2

(Using the 05-09 source from biosprite).

[kathgar]

That backtrace is not horribly helpful.. was it processing a truly NULL buffer of size 10, or did it just happen to begin with a \0? If the crash reoccures please reply.

[wize_one]

happened to us.. turn hold zones off

[Derision]

happened to us.. turn hold zones off

Thanks Wize_One, I turned hold zones off and although I have had a couple of zone crashes since, they don't bring the world to it's knees as it did when I had hold zones on. I am not using the latest TCS DB, but if you haven't already, it seems it would be a good idea to set holdzones to 0 in the DB's you release.

Thanks again.