On 19-01-2006 17:43, Yeung OnTai-q16645 wrote:
When I use "#serctl add <username> <password> <email>", it does not work anymore, because it complains that subscriber table is not found.
serctl (the shell script) will be deprecated.
Whatever database table(subscriber or other) is being used now, there should be a place where I can provision a new client with password. Otherwise how can I authenticate it when the client register to SER?
The table is called credentials, there is a command called ser_cred in sip_router/tools/serctl which can be used to provision the table. Note that an interface similar to the original serctl is being worked on, it will be available soon.
It seems like in ser-0.10.99, clients can just register to ser without any provisioning....
No, but the newer SER version has much cleaner data model and the meaning of tables changed. credentials table maps digest credentials to users (identified by uid). uri table maps SIP URIs to users, and so on.
Jan.
Can it be assume that without the ser_cred, there is no reliable way of populating the credentials table? I found that www_authenticate doesn't work in the presence snapshot. It seems that it cannot find the relevant credentials for a particular 'user', even though I have manually entered (what I believe to be) the relevant entries (username, password, ha1, ha1b). www_authenticate may be failing because I entered the information incorrectly - but there's no way of entering the information reliably without ser_cred, I believe?
Chee
--- Jan Janak jan@iptel.org wrote:
On 19-01-2006 17:43, Yeung OnTai-q16645 wrote:
When I use "#serctl add <username> <password>
<email>", it does not work
anymore, because it complains that subscriber
table is not found.
serctl (the shell script) will be deprecated.
Whatever database table(subscriber or other) is
being used now, there
should be a place where I can provision a new
client with password.
Otherwise how can I authenticate it when the
client register to SER?
The table is called credentials, there is a command called ser_cred in sip_router/tools/serctl which can be used to provision the table. Note that an interface similar to the original serctl is being worked on, it will be available soon.
It seems like in ser-0.10.99, clients can just
register to ser without
any provisioning....
No, but the newer SER version has much cleaner data model and the meaning of tables changed. credentials table maps digest credentials to users (identified by uid). uri table maps SIP URIs to users, and so on.
Jan.
__________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
On 24-01-2006 07:19, C Y wrote:
Can it be assume that without the ser_cred, there is no reliable way of populating the credentials table? I found that www_authenticate doesn't work in the presence snapshot. It seems that it cannot find the relevant credentials for a particular 'user', even though I have manually entered (what I believe to be) the relevant entries (username, password, ha1, ha1b). www_authenticate may be failing because I entered the information incorrectly - but there's no way of entering the information reliably without ser_cred, I believe?
Yes, there is, you can enter the information "by hand", corresponding hashes (ha1, ha1b) can be calculated using gen_ha1.
Credentials table is in fact very similar to subscriber table -- instead of username, domain, pair you have uid column.
If www_authenticate could not find the credentials then check whether you use correct realm (the first parameter of www_authenticate).
Jan.
Chee
--- Jan Janak jan@iptel.org wrote:
On 19-01-2006 17:43, Yeung OnTai-q16645 wrote:
When I use "#serctl add <username> <password>
<email>", it does not work
anymore, because it complains that subscriber
table is not found.
serctl (the shell script) will be deprecated.
Whatever database table(subscriber or other) is
being used now, there
should be a place where I can provision a new
client with password.
Otherwise how can I authenticate it when the
client register to SER?
The table is called credentials, there is a command called ser_cred in sip_router/tools/serctl which can be used to provision the table. Note that an interface similar to the original serctl is being worked on, it will be available soon.
It seems like in ser-0.10.99, clients can just
register to ser without
any provisioning....
No, but the newer SER version has much cleaner data model and the meaning of tables changed. credentials table maps digest credentials to users (identified by uid). uri table maps SIP URIs to users, and so on.
Jan.
Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com
On Tue, 24 Jan 2006 16:31:23 +0100, Jan Janak wrote
On 24-01-2006 07:19, C Y wrote:
Can it be assume that without the ser_cred, there is no reliable way of populating the credentials table? I found that www_authenticate doesn't work in the presence snapshot. It seems that it cannot find the relevant credentials for a particular 'user', even though I have manually entered (what I believe to be) the relevant entries (username, password, ha1, ha1b). www_authenticate may be failing because I entered the information incorrectly - but there's no way of entering the information reliably without ser_cred, I believe?
Yes, there is, you can enter the information "by hand", corresponding hashes (ha1, ha1b) can be calculated using gen_ha1.
Credentials table is in fact very similar to subscriber table -- instead of username, domain, pair you have uid column.
If www_authenticate could not find the credentials then check whether you use correct realm (the first parameter of www_authenticate).
Jan.
What does the UID column (replacing username and domain columns) look like? Is it going to be username@domain??
Is there any particular reason that = searchable tables would be deprecated by tables that require LIKE pattern matching for referencing?
If username/domain are removed in favour of UID, how does that affect other tables which use username/domain (such as AVPs) ?
It all sounds so mysterious! :) I'm looking forward to a full rewrite of the docs so that it will all make sense again.