Why deactivating ONU before activating it?

[wdl2015]

Hi Igor,

If BAL calls bbf_xpon_onu_discovered() callback it means this ONU is going through O1~O5 process for activation, it must be currently in inactive state, why we need deactivate this ONU before activating it in this callback?

Thanks a lot in advance for explanation.

/ Handle ONU_DISCOVERED indication / void bbf_xpon_onu_discovered(bcmolt_oltid olt, bcmolt_msg msg) { bcmolt_pon_interface_onu_discovered od = (bcmolt_pon_interface_onu_discovered *)msg; bcmos_bool auto_activated = BCMOS_FALSE; bcmos_errno err;

NC_LOG_DBG("ONU_discovered: onu_id=%u serial_number=%c%c%c%c-%02x%02x%02x%02x\n",
    od->data.onu_id,
    od->data.serial_number.vendor_id.arr[0], od->data.serial_number.vendor_id.arr[1],
    od->data.serial_number.vendor_id.arr[2], od->data.serial_number.vendor_id.arr[3],
    od->data.serial_number.vendor_specific.arr[0], od->data.serial_number.vendor_specific.arr[1],
    od->data.serial_number.vendor_specific.arr[2], od->data.serial_number.vendor_specific.arr[3]);

/* If onu_id is already assigned, that means that ONU is already provisioned.
   In this case activate it */
if (BCMOLT_FIELD_IS_SET(&od->data, pon_interface_onu_discovered_data, onu_id) &&
    od->data.onu_id != BCMOLT_ONU_ID_INVALID)
{
    const xpon_v_ani *onu_info = xpon_v_ani_get_by_id(od->key.pon_ni, od->data.onu_id);
    if (onu_info != NULL && onu_info->admin_state == XPON_ADMIN_STATE_ENABLED)
    {
        bcmolt_onu_set_onu_state set_state;
        bcmolt_onu_key key = { .pon_ni = od->key.pon_ni, .onu_id = od->data.onu_id };

        /* Deactivate first in case it is already active */
        err = _onu_deactivate(od->key.pon_ni, od->data.onu_id, BCMOS_FALSE);
        if (err != BCM_ERR_OK)
        {
            NC_LOG_ERR("ONU.set_onu_state(INACTIVE) failed for (pon-id:%u, onu-id:%u). Error %s\n",
                    key.pon_ni, key.onu_id, bcmos_strerror(err));
        }

Best regards, George Deng

[igortern]

Hi George, The idea is to start ONU provisioning from a well known state. It is possible to query ONU state and deactivate only if active, but I opted not to do it here because deactivating ONU doesn't do any harm if ONU was inactive. ONU discovery handling might need to be revised once someone implements logic for a sync with an active ONU. Regards, Igor

[wdl2015]

Hi Igor,

Thanks for quick response for this question. Here is a scenario that the deactivation might harm.

• Channel-termination and ONU/v-ani is provisioned and saved in sysrepo startup DB so that after OLT reboots the configuration recovers.
• Reboot the OLT and reload the sysrepo DB from startup. The sysrepo callback functions are triggered.
• When the channel-termination callback executes the PON port is open and ONU start activating. BAL calls the onu_discover() callback.
• At the same time the v-ani sysrepo callback executes and assign ONU-ID.
• In this case the bbf_xpon_onu_discovered() function does ONU deactivation while the ONU is doing activation. The deactivation fails due to the ONU is in "processing" state. It schedules a timer to disable again in 500ms later.
• The bbf_xpon_onu_discovered() moves forward activating the ONU and starts the FSM for MIB reset/upload. Now the 500ms timer expires and deactivates this ONU. The MIB reset/upload fails.

Could you review if this is a problem? I removed the ONU deactivation for now and this problem is solved.

Br George Deng

[igortern]

Hi George, I suspect that without ONU deactivation you will run into problem if restart NETCONF server after ONU was activated. Can you try adding nc_config_lock()/nc_config_unlock() calls in bbf_xpon_onu_discovered()? Thanks, Igor

[wdl2015]

Hi Igor,

1) By "NETCONF server" you mean bcmolt_netconf_server process or netopeer2-server process?

2) Could you detail the scenario "restart NETCONF server after ONU was activated"?

3) Tried adding nc_config_lock()/nc_config_unlock() as follows. Have the same problem for the scenario mentioned previously.

nc_config_logck() and nc_config_unlock are added in bbf_xpon_onu_discovered() like this.

        nc_config_lock(); // george
        /* Deactivate first in case it is already active */
        err = _onu_deactivate(od->key.pon_ni, od->data.onu_id, BCMOS_FALSE);
        if (err != BCM_ERR_OK)
        {
            NC_LOG_ERR("ONU.set_onu_state(INACTIVE) failed for (pon-id:%u, onu-id:%u). Error %s\n",
                    key.pon_ni, key.onu_id, bcmos_strerror(err));
        }

        /* Now activate */
        BCMOLT_OPER_INIT(&set_state, onu, set_onu_state, key);
        BCMOLT_MSG_FIELD_SET(&set_state, onu_state, BCMOLT_ONU_OPERATION_ACTIVE);
        err = bcmolt_oper_submit(netconf_agent_olt_id(), &set_state.hdr);
        if (err == BCM_ERR_OK)
        {
            auto_activated = BCMOS_TRUE;
        }
        else if (!(err == BCM_ERR_STATE && strstr(set_state.hdr.hdr.err_text, "processing") != NULL))
        {
            NC_LOG_ERR("ONU activation failed for %s. Error %s-%s\n",
                onu_info->hdr.name, bcmos_strerror(err), set_state.hdr.hdr.err_text);
        }
        nc_config_unlock(); // george

Repeat the scenario mentioned previously, see the same error in log.

Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91137: I OMCI_TRANSPORT ] omci_transport.c 206| Retransmission: OLT 0 PON 0 ONU 2 Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91137: D OMCI_TRANSPORT ] omci_transport.c 128| Resending (after ACK timeout) OMCI message to OLT 0 PON 0 ONU ID 2}, (buf = 0x7fa684001240, Msg len = 44, TCI = 0x1), retry msg counter = 1, ar=1 Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91137: D api ] /olt 0;^[[1m^[[32m/api/oper object=onu sub=cpu_packets pon_ni=0 onu_id=2 packet_type=omci calc_crc=yes number_of_packets=1 packet_size=44 buffer=00014f0a00020000000000000000000000000000000 Oct 10 18:57:51 bcmolt_netconf_server[5360]: 000000000000028
^[[0m^[#5
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: E OMCI_SVC ] omci_svc_remote_proxy_if.c 56| OMCI message send Failed for (PON=0, ONU=2)
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: E OMCI_SVC ] omci_svc_remote_proxy_if.c 110| omci_svc_omci_data_req failed: PON=0, ONU=2, rc=Object is in bad state
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: E OMCI_TRANSPORT ] omci_transport.c 144| Failed to send OMCI message to OLT 0 PON 0 ONU ID 2, rc=Error in parameters (Msg val = 0x7fa684001240, len = 44) Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: D OMCI_SVC ] omci_svc_onu.c 2467| key={pon_ni=0, onu_id=2}, op_ref=1: MIB reset request completed successfully
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: D OMCI_SVC ] omci_svc_onu.c 2342| key={pon_ni=0, onu_id=2}, op_ref=1: ONU SM state='mib_reset', event='mib_reset_success'
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: D OMCI_SVC ] omci_svc_onu.c 2342| key={pon_ni=0, onu_id=2}, op_ref=1: ONU SM state='mib_reset', event='start'
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: D OMCI_ME_LAYER ] omci_stack_enc_dec.c 8729| {olt=0 pon_if=0, onu_id=2, cookie=2}: Dump ME: onu_data (2), Entity: 0, Action: MIB_UPLOAD [
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: D OMCI_ME_LAYER ] omci_stack_enc_dec.c 8738| ]
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: D OMCI_TRANSPORT ] omci_transport.c 871| Send OMCI message {OLT:0 PON:0, ONU: 2, class:onu_data, instance:0, msg type:MIB_UPLOAD TCI=2 (0x7fa670000cb0)}, retry msg counter = 0 Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: D OMCI_TRANSPORT ] omci_stack_rx.c 913| 00 02 4D 0A 00 02 00 00 00 00 00 00 00 00 00 00
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: D OMCI_TRANSPORT ] omci_stack_rx.c 913| 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: D OMCI_TRANSPORT ] omci_stack_rx.c 913| 00 00 00 00 00 00 00 00 00 00 00 28
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: D OMCI_TRANSPORT ] omci_transport.c 128| Sending OMCI message to OLT 0 PON 0 ONU ID 2}, (buf = 0x7fa670000cb0, Msg len = 44, TCI = 0x2), retry msg counter = 0, ar=1
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: D api ] /olt 0;^[[1m^[[32m/api/oper object=onu sub=cpu_packets pon_ni=0 onu_id=2 packet_type=omci calc_crc=yes number_of_packets=1 packet_size=44 buffer=00024d0a00020000000000000000000000000000000 Oct 10 18:57:51 bcmolt_netconf_server[5360]: 000000000000028
^[[0m^[#5
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: E OMCI_SVC ] omci_svc_remote_proxy_if.c 56| OMCI message send Failed for (PON=0, ONU=2)
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: E OMCI_SVC ] omci_svc_remote_proxy_if.c 110| omci_svc_omci_data_req failed: PON=0, ONU=2, rc=Object is in bad state
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: E OMCI_TRANSPORT ] omci_transport.c 144| Failed to send OMCI message to OLT 0 PON 0 ONU ID 2, rc=Error in parameters (Msg val = 0x7fa670000cb0, len = 44) Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: E OMCI_SVC ] omci_svc_onu.c 944| key={pon_ni=0, onu_id=2}, op_ref=2: omci_svc_omci_mib_upload_req() failed, rc=Error in parameters
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: D OMCI_SVC ] omci_svc_onu.c 2420| key={pon_ni=0, onu_id=2}, op_ref=2: Starting transaction rollback from SM state='inactive'
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: D OMCI_TRANSPORT ] omci_transport.c 645| ONU Init: OLT 0 PON 0 ONU 2
Oct 10 18:57:51 bcmolt_netconf_server[5360]: [91138: E OMCI_SVC ] omci_svc_onu.c 861| key={pon_ni=0, onu_id=2}, op_ref=2: omci_svc_omci_activate_req() failed, rc=Entry already exists/already in requested state

Br George Deng

[igortern]

Hi George,

By "NETCONF server" you mean bcmolt_netconf_server process or netopeer2-server process? Could you detail the scenario "restart NETCONF server after ONU was activated"? Usually I terminate bcmolt_netconf_server and then use start_netconf_server.sh again. It restarts netopeer2-server as well. The scenario is simple: 1) apply configuration that activates an ONU 2) restart NETCONF server 3) apply the same configuration again

If nc_config_lock()/.unlock() doesn't help, another option is:

• remove _onu_deactivate() call (as you did)
• if activation fails with error BCM_ERR_STATE, query ONU state. If the state is active, invoke the same logic as invoked in _olt_onu_indication_cb:BCMOLT_ONU_AUTO_SUBGROUP_ONU_ACTIVATION_COMPLETED

Thanks, Igor

[wdl2015]

Hi Igor,

Thanks a lot. Clear now. Trying your proposal.

Do you think below is an option?

• Check ONU state. If it is "processing" don't do _onu_deactivate() because "processing" means this ONU is going from O2 to O5 for activation. If it is "active" do _onu_deactivate().

Br George Deng

[igortern]

Hi George,

In case of embedded OMCI ONU activation is a lengthy process that involves O2->O5 transition followed by the initial OMCI provisioning. I think instead of checking onu_state=="processing", need to add some activation_in_progress flag in xpon_v_ani structure.

• set it before starting ONU activation
• clear when setting omci_ready flag

Thanks, Igor

On Tue, Oct 11, 2022 at 4:55 PM dengguochen2003 @.***> wrote:

Hi Igor,

Thanks a lot. Clear now. Trying your proposal.

Do you think below is an option?

• Check ONU state. If it is "processing" don't do _onu_deactivate() because "processing" means this ONU is going from O2 to O5 for activation. If it is "active" do _onu_deactivate().

Br George Deng

— Reply to this email directly, view it on GitHub https://github.com/balapi/netconf-polt/issues/13#issuecomment-1274123282, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEX4KW57YQ6VJTW7PDQ5HODWCT6MZANCNFSM6AAAAAARBAMSOA . You are receiving this because you commented.Message ID: @.***>

-- This electronic communication and the information and any files transmitted with it, or attached to it, are confidential and are intended solely for the use of the individual or entity to whom it is addressed and may contain information that is confidential, legally privileged, protected by privacy laws, or otherwise restricted from disclosure to anyone else. If you are not the intended recipient or the person responsible for delivering the e-mail to the intended recipient, you are hereby notified that any use, copying, distributing, dissemination, forwarding, printing, or copying of this e-mail is strictly prohibited. If you received this e-mail in error, please return the e-mail to the sender, delete it from your computer, and destroy any printed copy of it.

[wdl2015]

Hi Igor,

It is a nice idea to add one more flag activation_in_progress which indicates the ONU&OMCI is ready.

BTW: What we are considering are the "ONU reboot" and "OLT reboot" scenarios. Met the issue in this ticket for the OLT reboot scenario.

• For ONU reboot the vlan_subifs are available in OLT, we add a function vlan_subif_reapply_to_onu_flows() in omci_svc_state_up_sequence_end_event_start() to reapply the subifs to OMCI to ONU. Code follows.
• For OLT reboot the vlan_subifs are not available in OLT, they are reloaded from persistent sysrepo and apply to OMCI/ONU. But the sysrepo callback and onu_discovered() indication may conflict because they don't have order control.

static void omci_svc_state_up_sequence_end_event_start(bcmonu_mgmt_onu_key key, omci_svc_onu onu_context, void context) { bcmos_errno rc; bcmonu_mgmt_onu_cfg onu_cfg; bcmolt_oltid olt_id = onu_context->onu_cfg->hdr.hdr.olt_id; onu_context->state = OMCI_SVC_ONU_STATE_ID_ACTIVE_WORKING; onu_context->oper_status = BCMONU_MGMT_STATUS_UP; OMCI_SVC_LOG(INFO, olt_id, key, NULL, "ONU added successfully (is_up=true)\n");

/* update cfg DB first, since notify calls does a GET on the onu */
if (BCM_ERR_OK == onu_context->last_err)
{
    /* update sm cfg DB */
    onu_cfg = onu_context->onu_cfg;
    rc = omci_svc_sm_cfg_db_update_entry(onu_cfg);
    if (BCM_ERR_OK != rc)
        OMCI_SVC_LOG(ERROR, olt_id, key, &onu_cfg->hdr.hdr, "ONU cfg DB update Failed\n");
}

if (onu_context->cb)
    onu_context->cb(onu_context->context, onu_context->last_err);

rc = vlan_subif_reapply_to_onu_flows(key->pon_ni, key->onu_id);
if (BCM_ERR_OK != rc)
{
    OMCI_SVC_LOG(ERROR, olt_id, key, NULL, "Reapply vlan subif to ONU failed.\n");
}

Br, George Deng

[venkatkakumani]

Can you provide the patch for this issue?