[SR-Users] Dialog is not cleaned up after failed calls and BYEs.

[Aft nix]

Hi,

I'm trying to limit concurrent calls through kamailio using following script:
#!ifdef WITH_CALL_LIMIT
        if (is_method("INVITE")) {
                dlg_manage();
                $var(100) = 0;
                get_profile_size("callquota", "$fd", "$var(100)");
                if ($var(100) >= 100 ) {
                        xdbg("DEBUG: Simultaneous calls limit reached");
                        sl_send_reply("503","Simultaneous calls limit reached");
                        exit;
                }

                set_dlg_profile("callquota","$fd");

}
#!endif

But the dialogs are not cleaned up after the call ends or a failed call happens.

For example after two failed calls :

[root at new tls]# kamctl fifo profile_get_size callquota
profile::  name=callquota value= count=4

and the dialog list :

[root at new tls]# kamctl fifo dlg_list
dialog::  hash=896:11654
        state:: 1
        ref_count:: 1
        timestart:: 0
        timeout:: 0
        callid:: iSGRqQR0C4suBpQD0rHAEgV42vw0Sbpe
        from_uri:: sip:123 at 205.164.40.150
        from_tag:: d8x2LDBptN9FRJWtqFTj4DIpuQlKq.KN
        caller_contact:: sip:123 at 180.149.7.45:33852;ob
        caller_cseq:: 11755
        caller_route_set::
        caller_bind_addr:: udp:185.8.105.70:9547
        callee_bind_addr::
        to_uri:: sip:8801684077496 at 205.164.40.150
        to_tag::
        callee_contact::
        callee_cseq::
        callee_route_set::
dialog::  hash=896:11655
        state:: 1
        ref_count:: 1
        timestart:: 0
        timeout:: 0
        callid:: iSGRqQR0C4suBpQD0rHAEgV42vw0Sbpe
        from_uri:: sip:123 at 205.164.40.150
        from_tag:: d8x2LDBptN9FRJWtqFTj4DIpuQlKq.KN
        caller_contact:: sip:123 at 180.149.7.45:33852;ob
        caller_cseq:: 11755
        caller_route_set::
        caller_bind_addr:: udp:185.8.105.70:9547
        callee_bind_addr::
        to_uri:: sip:8801684077496 at 205.164.40.150
        to_tag::
        callee_contact::
        callee_cseq::
        callee_route_set::
dialog::  hash=896:11657
        state:: 1
        ref_count:: 1
        timestart:: 0
        timeout:: 0
        callid:: iSGRqQR0C4suBpQD0rHAEgV42vw0Sbpe
        from_uri:: sip:123 at 205.164.40.150
        from_tag:: d8x2LDBptN9FRJWtqFTj4DIpuQlKq.KN
        caller_contact:: sip:123 at 180.149.7.45:33852;ob
        caller_cseq:: 11756
        caller_route_set::
        caller_bind_addr:: udp:185.8.105.70:9547
        callee_bind_addr::
        to_uri:: sip:8801684077496 at 205.164.40.150
        to_tag::
        callee_contact::
        callee_cseq::
        callee_route_set::
dialog::  hash=1532:8089
        state:: 1
        ref_count:: 1
        timestart:: 0
        timeout:: 0
        callid:: povqQaC.sRhKvmUaLzzR83bRXxouoNy8
        from_uri:: sip:123 at 205.164.40.150
        from_tag:: qXQll3QXrff.t27FHY2hucPfJ7K8Pfrs
        caller_contact:: sip:123 at 180.149.7.45:33852;ob
        caller_cseq:: 9232
        caller_route_set::
        caller_bind_addr:: udp:185.8.105.70:9547
        callee_bind_addr::
        to_uri:: sip:8801684077496 at 205.164.40.150
        to_tag::
        callee_contact::
        callee_cseq::
        callee_route_set::


You can see there are multiple entries. And these dialog does not get
cleaned up. The data is collected when there was no call(pending or
connected).


My kamailio version is :

 [root at new tls]# kamailio -V
version: kamailio 4.0.0-pre0 (x86_64/linux) 83eafc
flags: STATS: Off, USE_IPV6, USE_TCP, USE_TLS, TLS_HOOKS,
USE_RAW_SOCKS, DISABLE_NAGLE, USE_MCAST, DNS_IP_HACK, SHM_MEM,
SHM_MMAP, PKG_MALLOC, USE_FUTEX, FAST_LOCK-ADAPTIVE_WAIT,
USE_DNS_CACHE, USE_DNS_FAILOVER, USE_NAPTR, USE_DST_BLACKLIST,
HAVE_RESOLV_RES
ADAPTIVE_WAIT_LOOPS=1024, MAX_RECV_BUFFER_SIZE 262144, MAX_LISTEN 16,
MAX_URI_SIZE 1024, BUF_SIZE 65535, DEFAULT PKG_SIZE 4MB
poll method support: poll, epoll_lt, epoll_et, sigio_rt, select.
id: 83eafc
compiled on 18:34:58 Jan 17 2013 with gcc 4.4.6

Any suggestion?(It seems dialog module is not working as it should be.
I've also tried calling unset_profile() forcefully in failure route
with no luck)
--
-aft