[389-users] 389 DS is reseting connections

[Diego Woitasen]

On Mon, Feb 7, 2011 at 1:48 PM, Rich Megginson <rmeggins at redhat.com> wrote:
> On 02/07/2011 09:41 AM, Diego Woitasen wrote:
>>
>> Hi,
>>  I have 389 DS 1.2.7.5 running on Debian Squeeze. It was working fine
>> but the last days the process started to hang very often. I restart
>> the service, works fine for a few minutes and hangs again. The process
>> is running, accept connections but reset them.
>>
>> The only error message that I see is from ldapsearch:
>>
>> ldap_start_tls: Can't contact LDAP server (-1)
>> ldap_sasl_bind(SIMPLE): Can't contact LDAP server (-1)
>
> Could this be https://bugzilla.redhat.com/show_bug.cgi?id=668619 ?
>>
>> I ran ldapsearch with strace, the last lines:
>>
>> socket(PF_INET, SOCK_STREAM, IPPROTO_IP) = 3
>> fcntl(3, F_SETFD, FD_CLOEXEC)           = 0
>> setsockopt(3, SOL_SOCKET, SO_KEEPALIVE, [1], 4) = 0
>> setsockopt(3, SOL_TCP, TCP_NODELAY, [1], 4) = 0
>> connect(3, {sa_family=AF_INET, sin_port=htons(389),
>> sin_addr=inet_addr("140.191.48.138")}, 16) = 0
>> write(3, "0\35\2\1\1w\30\200\0261.3.6.1.4.1.1466.20037", 31) = 31
>> poll([{fd=3, events=POLLIN|POLLPRI|POLLERR|POLLHUP}], 1, -1) = 1
>> ([{fd=3, revents=POLLIN|POLLERR|POLLHUP}])
>> read(3, 0x11ed85f, 8)                   = -1 ECONNRESET (Connection
>> reset by peer)
>> write(2, "ldap_start_tls: Can't contact LD"..., 47ldap_start_tls:
>> Can't contact LDAP server (-1)
>> ) = 47
>> write(2, "ldap_sasl_bind(SIMPLE): Can't co"...,
>> 55ldap_sasl_bind(SIMPLE): Can't contact LDAP server (-1)
>> ) = 55
>> exit_group(-1)
>>
>> I tried to trace ns-slapd, but I don't see anything special (except
>> the ENOTCONN error in getpeername() but it's on a different FD):
>>
>> 2007  accept(6, {sa_family=AF_INET, sin_port=htons(53395),
>> sin_addr=inet_addr("140.191.48.
>> 138")}, [16]) = 34
>> 2007  fcntl(34, F_GETFL)                = 0x2 (flags O_RDWR)
>> 2007  fcntl(34, F_SETFL, O_RDWR|O_NONBLOCK) = 0
>> 2007  fcntl(34, F_DUPFD, 64)            = 64
>> 2007  close(34)                         = 0
>> 2007  setsockopt(64, SOL_TCP, TCP_NODELAY, [0], 4) = 0
>> 2007  getpeername(64, {sa_family=AF_INET, sin_port=htons(53395),
>> sin_addr=inet_addr("140.1
>> 91.48.138")}, [16]) = 0
>> 2007  getsockname(64, {sa_family=AF_INET, sin_port=htons(389),
>> sin_addr=inet_addr("140.191
>> .48.138")}, [16]) = 0
>> 2007  getpeername(7, 0x7fff1acd6e90, [112]) = -1 ENOTCONN (Transport
>> endpoint is not conne
>> cted)
>> 2007  poll([{fd=22, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7,
>> events=POLLIN}, {fd=-1},
>>  {fd=64, events=POLLIN}], 5, 250) = 1 ([{fd=64, revents=POLLIN}])
>> 2007  close(64)                         = 0
>> 2007  getpeername(7, 0x7fff1acd6e90, [112]) = -1 ENOTCONN (Transport
>> endpoint is not conne
>> cted)
>> 2007  poll([{fd=22, events=POLLIN}, {fd=6, events=POLLIN}, {fd=7,
>> events=POLLIN}, {fd=-1}]
>> , 4, 250<unfinished ...>
>> 2010<... select resumed>  )            = 0 (Timeout)
>> 2010  select(0, NULL, NULL, NULL, {0, 100000}<unfinished ...>
>> 2012<... select resumed>  )            = 0 (Timeout)
>>
>>
>> Any hint to help to find the problem? I tried with different slapd log
>> levels but i don't see anything special. I don't except a magical
>> solution, only a hint to discover what's happening.
>>
>> Regards,
>>  Diego
>>
>
>

Yes, looks the same problem. Thanks!

-- 
Diego Woitasen