[Gluster-devel] Re: TLA 579 seems broken

[Brent A Nelson]

FYI, here's the backtrace output:

#0  0xb7f276bc in sys_poll_register () from /usr/lib/libglusterfs.so.0
(gdb) bt
#0  0xb7f276bc in sys_poll_register () from /usr/lib/libglusterfs.so.0
#1  0xb7f26e99 in poll_register () from /usr/lib/libglusterfs.so.0
#2  0xb658e0be in gf_transport_init ()
   from /usr/lib/glusterfs/1.3.8/transport/tcp/server.so
#3  0xb7f270d9 in transport_load () from /usr/lib/libglusterfs.so.0
#4  0xb759d09a in init ()
   from /usr/lib/glusterfs/1.3.8/xlator/protocol/server.so
#5  0xb7f21234 in ?? () from /usr/lib/libglusterfs.so.0
#6  0x08051290 in ?? ()
#7  0x0000002a in ?? ()
#8  0x00000000 in ?? ()

Thanks,

Brent

On Tue, 27 Nov 2007, Brent A Nelson wrote:

> I just tried TLA 579, but glusterfsd core dumped on me.  This is an 
> epoll-related patch, and strace shows it dies immediately after epoll-related 
> activity on startup.
>
> Strace snippet:
> ...
> open("/usr/lib/glusterfs/1.3.8/transport/tcp/server.so", O_RDONLY) = 5
> read(5, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0\3\0\1\0\0\0P\r\0\000"..., 512) = 
> 512
> fstat64(5, {st_mode=S_IFREG|0755, st_size=9128, ...}) = 0
> mmap2(NULL, 12196, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 5, 0) = 
> 0xb65c1000
> mmap2(0xb65c3000, 4096, PROT_READ|PROT_WRITE, 
> MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 5, 0x1) = 0xb65c3000
> close(5)                                = 0
> socket(PF_INET, SOCK_STREAM, IPPROTO_TCP) = 5
> setsockopt(5, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
> bind(5, {sa_family=AF_INET, sin_port=htons(6996), 
> sin_addr=inet_addr("0.0.0.0")}, 16) = 0
> listen(5, 10)                           = 0
> epoll_create(-1)                        = -1 EINVAL (Invalid argument)
> epoll_create(1024)                      = 6
> epoll_ctl(6, EPOLL_CTL_ADD, 5, {EPOLLIN|EPOLLPRI|EPOLLERR|EPOLLHUP, 
> {u32=134552776, u64=134552776}}) = 0
> --- SIGSEGV (Segmentation fault) @ 0 (0) ---
> +++ killed by SIGSEGV (core dumped) +++
>
> Thanks,
>
> Brent