There are 2 .up scripts that modify voip_allowed_ip_groups.ipnet
15702.up - modifies it to varchar(43)
15710.up - modifies it to varchar(46)
Because in 15710.up was backported into mr9.5 but is missing in mr10.5,
existing upgrades from mr9.5 -> mr10.5 apply 15702 after 15710.up,
effectively downgrading the column to varchar(46).
This .up script fixes the scenario and also the .down script modifies
the column also to varchar(46) to respect the previous 15710.up the
sets it to varchar(46), so to achieve the consistent state.
Change-Id: If14917ef437de3abf46b1900f68c9cc3a9a4b947
(cherry picked from commit 9c6ef86372)
* the ngcp.timezone table is not replicated and hence
the script must be also not replicated.
- the script is renamed to _not_replicated
- the ALTER TABLE ADD COLUMN is enhanced with
"IF NOT EXISTS" to mitigate scenarios where
somebody manually added the column on the other node,
and to ensure that the upgrades are safe
Change-Id: Ic44f76b5f3f43ae4124cebad30d559224fa90b31
(cherry picked from commit ff81df911a)
* this column stores the currently used olson database timezone
version
Change-Id: Ib26c0d731d7f378717ea399d3a9f06d5f033a596
(cherry picked from commit 55477004f8)
* due to the fact that the initial script had been created as replicated
and then renamed to _not_replicated, caused already deployed platforms on
mr11.1, mr11.1.1, mr11.2, mr11.2.1 to have the index in both tables
(because of the upgrade/ngcp-sync-db step on sp2) and a missing
record in ngcp.db_schema for 'sp2' (because the script was named
as replicated initially). To further address the issue, "CREATE INDEX"
is now adjusted to "CREATE INDEX IF NOT EXISTS"
Change-Id: Ie1554bad9300da5ff711cfe53d1cc6c14f91793d
(cherry picked from commit d23bfb856c)
* the db scripts have 'set sql_log_bin=0' but do not have the
_not_replicated filename part. they are not renamed to have
the _not_replicated required part.
Change-Id: I78d841f5ac4c5dff7c13e292746688faecd062d9
(cherry picked from commit 4fbcd5bfa5)
The preference is used only for device provisioning
to give priority to user changes over the automated
provisioning coming from the server.
Change-Id: I65bb0b90b80e11aa34f7b56019f0b57319678cc0
Activate profile preferences to all the options already
available for both subscriber and domain.
Extend 'music_on_hold' and 'clir_intrapbx' to be
domain preferences.
Add 'ignore_cf_when_hunting' preference to subscriber
role.
Fix and update 'expose_to_customer' and 'expose_to_subscriber'
columns to properly expose what is necessary for each
role and hide features that should be activated only
by system administrator.
Change-Id: Ia0aa8f9f10dc1bab47ff8b9ccfc90c2741e988e0
* ringtimeout is now internal = 1 as this preference is dynamically
changed by CFT call forwards and is not meant to be usable from the
UI/API
Change-Id: I6c2abf94e455cf50cdbc6cf7122844582ba6bbb4
* expose_to_customer is now set to 1 for 'cli' preference so that
subscriberadmin users can access it
Change-Id: Ia661a7095457bd0fbf6eab5c74cd17eb224d7c3b
* expose_to_subscriber is a new preferences flag that
indicates that the preference can be read/modified by
'subscriber' role
* expose_to_customer now only does the same as expose_to_subscriber
but for 'subscriberadmin' role
Change-Id: I4b38ac5653e7132510354de4fd8b72d010d0ff68
* ncos_sets table contains a list of ncos_sets
* ncos_set_levels table contains "many to many" mappings
for ncos_sets and ncos_levels
* add ncos_levels.time_set_id column (default null)
Change-Id: Ib46fe46bd75ef13b8375961c93f586281f28866d
This reverts commit 1631573453.
The script is actually a one-time only script, no need to re-execute
it again, instead it fails to run due to:
| Applying revision script /usr/share/ngcp-db-schema/db_scripts/diff/15743.up: ERROR 1050 (42S01) at line 4: Table 'autoprov_firmwares_data' already exists
| failed. :(
Change-Id: I75ea1a0ce86064113032844a9ff4fc8a2b0af7cd
In git rev a9c1cf4cb we had to take care of adjusting the offset
variable to support MariaDB >=10.6 in the existing diff/15301.up script.
So that also existing installations receive that change during upgrade,
let's provide the according script which re-creates the
split_autoprov_firmware_data stored procedure.
Change-Id: Iae9634b39bda9f1ad63b2b6093f2e6095f68fc45
Thanks: Kirill Solomko <ksolomko@sipwise.com>
Fixes:
| Applying revision script /usr/share/ngcp-db-schema/db_scripts/diff/15301.up: ERROR 1064 (42000) at line 14: You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near 'offset int;
| DECLARE chunk longblob;
| DECLARE done int DEFAULT FALSE;
|
| ...' at line 4
| failed. :(
This is caused by MariaDB having "offset" as reserved word, quoting from
https://mariadb.com/kb/en/upgrading-from-mariadb-10-5-to-mariadb-10-6/:
| New reserved word: OFFSET. This can no longer be used as an identifier without being quoted.
Now that Debian/bookworm includes mariadb-server v10.6.10-1
we need to rename the variable.
Change-Id: I9db699ef60259411b8b79305f08179d4989edcd2
* the new table is to contain data for customer speed dials
(that affect all subscribers under the customer)
* triggers that replicate the data into kamailio.speed_dial
Change-Id: I0bc175e0ede7fa79b296c83a7f0d512b03c74aab
rtcengine product has been discontinued already since mr10.5.1
so it is useless to keep the database in the system
Change-Id: Icdf06c8d9de9a0236a5fc87b645d698952db3b44
This reverts commit 5f34f84b04.
Reason for revert: upstream provide us with a solution that doesn't need a kamailio change
Change-Id: Ifdb7f8c55f27498a880cc35a104e20472e6e69f7
This modification enables the get uri related to a speed dial number for a group (with 3 digits - *999).
It gets the new_uri from the table speed_dial using the group_id, the sd_username and sd_domain.
Change-Id: I629c2b5e35b53ee21a1e3a67df7a66f095da9db4
We have to concern slow selections like this:
"SELECT a.callid FROM acc a WHERE a.method = 'INVITE' GROUP BY a.callid LIMIT 0,200000;"
which can take procesisng time over 10 sec (as reported in the WF ticket).
As a simple solution, we can give an additional index to the 'acc' table:
KEY `method_callid_idx` (`method`,`callid`)
Which will give us an indexation on selecitons with the WHERE clause on 'method' column.
Without the index (explain of the selection):
*************************** 1. row ***************************
id: 1
select_type: SIMPLE
table: a
type: range
possible_keys: NULL
key: callid_method_idx
key_len: 817
ref: NULL
rows: 2
Extra: Using where; Using index for group-by
With the index (explain of the selection):
*************************** 1. row ***************************
id: 1
select_type: SIMPLE
table: a
type: ref
possible_keys: method_callid_idx
key: method_callid_idx
key_len: 50
ref: const
rows: 1
Extra: Using where; Using index
The newer index is not to be mixed with the previously existing one:
KEY `callid_method_idx` (`callid`,`method`)
Change-Id: Ie8a4ab102adb8788d9e273bd5a9edf1ddbb702bf
In order to introduce the new 'fax=' prefix, we have to update
the already configured CF destinations to fax2mail, and prefix them
with 'fax='.
Change-Id: Id09886932ba3246b7e10b36fb4f3f50c61139c7a
New index for watchers table for columns "inserted_time" and "status".
This is a little performance boost since the function
ps_watchers_db_timer_clean() cleans pending subscriptions
using this columns inside the "where" clause.
Change-Id: I9cb2220f6fe89aa2070d9a739c4a3a0417e3de71
Remove 'serial' from the label and description of
'cloud_pbx_hunt_timeout' as this preference can now also
be used for parallel groups.
Change-Id: I66a51a5dddd9b872d97b2ec3de36ce79524abce6
Improve description of ignore_cf_when_hunting
explicitly stating that this preference
only affects members of hunt groups when called in a hunt group context.
Change-Id: I0cfb0f8187e18f2771055866879363ee0dd4a050
see RQRAINB-6610
Add the new domain/subscriber preference 'busy_hg_member_mode' to
determine if a huntgroup call can be distributed or not to a busy member.
The new preference accepts the following values:
- 'ring' the call is always distributed to the member whatever is
his/her busy status (note: default value for domain preference)
- 'skip_totaluser' or 'skip_activeuser' the call is not distributed
to a member considered as busy depending of the counter 'totaluser'
or 'activeuser'.
Change-Id: I83973d7bdae2420d8c6836ddcbc542eb7c6f9fd2
By default calls that come from a peer group and are looped back
to the same peer group due to a subscriber location mappings 'forward'
entry are blocked to avoid loop creation.
This preference allows to override the default behavior and permit
the loop over the same peering group.
It has to be used with caution and activated only if necessary.
Change-Id: I5040204805c3b2a740250cec11a8e9914333f10b
New subscriber location mapping mode "forward" is now part of the
enumeration inside provisioning.voip_subscriber_location_mappings.
Change-Id: I93bbecd1e88452ddbc5249f344c3ad66ddea8739
Set peer flag for the following user preferences to allow peer sdp filtering:
- codecs_list
- codecs_filter
- codecs_id_list
- codecs_id_filter
Change-Id: I619cd3c3f6f7d87b91502ecf7563556e991d5089
This value will be the default that is assigned to 'normal' subscribers
The HuntGroup subscribers will contain instead one of the other possible
values: 'serial','parallel', 'random', 'circular'.
Change-Id: I5400e2ccfa5af5905777f1bc608054e0743ea1ef
there is an issue with a cornercase of resolving
domain preferences allowed_ips overloads. the cursor
loop of the proc therefor now gets proper statemachine
logic.
Change-Id: Ieeacc4d7c8a931c6e9f0f8087b2a05112f78c675
rationale:
-ipv6 allowed_ip adresses were not yet supported in kamailio
-the current ipv4 allowed_ip check has bad performance
this change introduces a new UDF to check the UA ip against
allowed_ip and man_allowed_ip preferences in a fast and
convenient way, ie.:
SELECT provisioning.ip_is_allowed(
<subscriber_uuid>,
<ipv4_or_ipv6_address>
);
the check is based on range queries supported by indexes,
following the proven approach we already use for roaming/
billing_networks.
it requires to transfrom IPnet values of existing data into
numerical representation of the network start (network) and
end (broadcast) addresses.
mariadb does not support arithmetic or bitwise operation
for varbinary or numerical values larger than bigint, which
is required to calculate network and broadcast adresses.
the change therefore provides an own implementation for
large integer operations. it works with hex string
representations, which fits the existing mariadb "hex()",
"unhex()" and "inet6_aton/ntoa()" functions.
Change-Id: I9dbaa769252babf168243167c061f035f0f43242
All "transcode" preferences are listed as type "0" in the DB except for
DTMF and CN which are listed as "1". The LUA/Kamailio scripts fail to
include tyoe "1" preferences in the list of flags given to rtpengine.
Change the type to "0" to fix.
Change-Id: I38f0d9f1394e07b190f1bd196ac85f456017ae6f
play_announce_before_recording cant have 2 default values
and "never" should not be default for subscriber
Change-Id: I0ecfb76ac6d9f9c7e431f4bc4b28d2619a06e076
* fileshare is not replicated by default therefore the script
breaks replication and it is converted to not_replicated to
address that.
Change-Id: I52f59630f92f4b9c901a84a296cd9e53442f5c0c
* there can be more than one journal record created per a
transaction (e.g. a subscriber and a customer are created
at once, therefore the tx_id field cannot be unique)
Change-Id: I0046f41ec52412dbad319610d7b1b87988798586
- create table acl_roles that contains information about a role
- inserted existing roles (system, admin, reseller, ccareadmin, ccare,
lintercept)
- create table acl_role_mappings that contains information about which
role has access to other roles
- remove field role from journals table and add FK role_id
- add FK role_id and set the right one accordingly
Change-Id: Id92726e9aaf7c6f6c126b18a76f86fc133aea5ae
Kirill Solomko
Marco Capetta
Michael Prokop
Víctor Seva
Fabricio Santolin da Silva
dzenichev
Daniel Grotti
Rene Krenn
Richard Fuchs
Alessio Garzi
Antoine Lingat
Oleksandr Duts
Michael Berger