[MDEV-15935] Connection Redirection Mechanism in MariaDB Client/Server Protocol - Jira

Alex Lee created issue - 2018-04-20 01:58

Alex Lee made changes - 2018-04-20 01:59

Field	Original Value	New Value
Description	h1. Why Redirection mechanism is widely used in proxy-based scenario. Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue. Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted. h1. What to Do Client/Server Protocol change needs both client and server support. The routine is # Server sets the capability flag [3] in Initial Handshake Packet, indicating server supports MARIADB_SERVER_REDIRECTION. ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ \| MARIADB_SERVER_REDIRECTION \| 1 <<31 \| Server support redirection \| ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ # Client sets the capability flag [3] in Client Handshake Response Packet, indicating client supports MARIADB_SERVER_REDIRECTION. # If client sets MARIADB_SERVER_REDIRECTION in Client Handshake Response Packet [3], server can send Redirection Information Packet other than OK_Packet / ERR_Packet / Authentication Switch Packet. # Client receives Redirection Information Packet, close the connection, and then use its information to connect the new host. # Client will use the redirection information to set up new connections until information expires or login parse error happens. The redirection information packet contains TTL of the information. The Redirection Information Packet schema is \|\|Heading 1\|\|Heading 2\|\| \|Col A1\|Col A2\| \|\|Field Type\|\|Field Name\|\|Notes\|\| \|int<1>\|Packet Header\|Packet header sets 0xFC to indicate it is a Redirection Information Packet.\| \|int<1>\|Protocol Version\|Currently protocol version is 0, specifying TCP-IPv4 protocol.\| \|int<2>\|Protocol Property\|For TCP-IPv4 protocol, it is the port number.\| \|string<NUL>\|Redirected Host\|Redirected host is a valid IPv4 address, an A name pointing to this address, or a CNAME record.\| \|int<2>\|Time to Live\|TTL is the life time of the validity of this redirected information with unit ‘second’.\| For recursion of redirection, in the scenario, the recursion is not useful. Should not recommend client side implement recursion of redirection. [1] https://developer.mozilla.org/en-US/docs/Web/HTTP/Redirections [2] https://docs.oracle.com/cd/B10501_01/network.920/a96580/concepts.htm [3] https://mariadb.com/kb/en/library/1-connecting-connecting/	h1. Why Redirection mechanism is widely used in proxy-based scenario. Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue. Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted. h1. What to Do Client/Server Protocol change needs both client and server support. The routine is # Server sets the capability flag [3] in Initial Handshake Packet, indicating server supports MARIADB_SERVER_REDIRECTION. ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ \| MARIADB_SERVER_REDIRECTION \| 1 <<31 \| Server support redirection \| ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ # Client sets the capability flag [3] in Client Handshake Response Packet, indicating client supports MARIADB_SERVER_REDIRECTION. # If client sets MARIADB_SERVER_REDIRECTION in Client Handshake Response Packet [3], server can send Redirection Information Packet other than OK_Packet / ERR_Packet / Authentication Switch Packet. # Client receives Redirection Information Packet, close the connection, and then use its information to connect the new host. # Client will use the redirection information to set up new connections until information expires or login parse error happens. The redirection information packet contains TTL of the information. The Redirection Information Packet schema is \|\|Field Type\|\|Field Name\|\|Notes\|\| \|int<1>\|Packet Header\|Packet header sets 0xFC to indicate it is a Redirection Information Packet.\| \|int<1>\|Protocol Version\|Currently protocol version is 0, specifying TCP-IPv4 protocol.\| \|int<2>\|Protocol Property\|For TCP-IPv4 protocol, it is the port number.\| \|string<NUL>\|Redirected Host\|Redirected host is a valid IPv4 address, an A name pointing to this address, or a CNAME record.\| \|int<2>\|Time to Live\|TTL is the life time of the validity of this redirected information with unit ‘second’.\| For recursion of redirection, in the scenario, the recursion is not useful. Should not recommend client side implement recursion of redirection. [1] https://developer.mozilla.org/en-US/docs/Web/HTTP/Redirections [2] https://docs.oracle.com/cd/B10501_01/network.920/a96580/concepts.htm [3] https://mariadb.com/kb/en/library/1-connecting-connecting/

Alex Lee made changes - 2018-04-20 01:59

Description

h1. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.
h1. What to Do
Client/Server Protocol change needs both client and server support. The routine is
# Server sets the capability flag [3] in Initial Handshake Packet, indicating server supports MARIADB_SERVER_REDIRECTION.
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
| MARIADB_SERVER_REDIRECTION | 1 <<31 | Server support redirection |
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
# Client sets the capability flag [3] in Client Handshake Response Packet, indicating client supports MARIADB_SERVER_REDIRECTION.
# If client sets MARIADB_SERVER_REDIRECTION in Client Handshake Response Packet [3], server can send Redirection Information Packet other than OK_Packet / ERR_Packet / Authentication Switch Packet.
# Client receives Redirection Information Packet, close the connection, and then use its information to connect the new host.
# Client will use the redirection information to set up new connections until information expires or login parse error happens. The redirection information packet contains TTL of the information.

The Redirection Information Packet schema is

||Field Type||Field Name||Notes||
|int<1>|Packet Header|Packet header sets 0xFC to indicate it is a Redirection Information Packet.|
|int<1>|Protocol Version|Currently protocol version is 0, specifying TCP-IPv4 protocol.|
|int<2>|Protocol Property|For TCP-IPv4 protocol, it is the port number.|
|string<NUL>|Redirected Host|Redirected host is a valid IPv4 address, an A name pointing to this address, or a CNAME record.|
|int<2>|Time to Live|TTL is the life time of the validity of this redirected information with unit ‘second’.|

For recursion of redirection, in the scenario, the recursion is not useful. Should not recommend client side implement recursion of redirection.

[1] https://developer.mozilla.org/en-US/docs/Web/HTTP/Redirections
[2] https://docs.oracle.com/cd/B10501_01/network.920/a96580/concepts.htm
[3] https://mariadb.com/kb/en/library/1-connecting-connecting/

h1. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.
h1. What to Do
Client/Server Protocol change needs both client and server support. The routine is
# Server sets the capability flag [3] in Initial Handshake Packet, indicating server supports MARIADB_SERVER_REDIRECTION.
| MARIADB_SERVER_REDIRECTION | 1 <<31 | Server support redirection |
# Client sets the capability flag [3] in Client Handshake Response Packet, indicating client supports MARIADB_SERVER_REDIRECTION.
# If client sets MARIADB_SERVER_REDIRECTION in Client Handshake Response Packet [3], server can send Redirection Information Packet other than OK_Packet / ERR_Packet / Authentication Switch Packet.
# Client receives Redirection Information Packet, close the connection, and then use its information to connect the new host.
# Client will use the redirection information to set up new connections until information expires or login parse error happens. The redirection information packet contains TTL of the information.

The Redirection Information Packet schema is

||Field Type||Field Name||Notes||
|int<1>|Packet Header|Packet header sets 0xFC to indicate it is a Redirection Information Packet.|
|int<1>|Protocol Version|Currently protocol version is 0, specifying TCP-IPv4 protocol.|
|int<2>|Protocol Property|For TCP-IPv4 protocol, it is the port number.|
|string<NUL>|Redirected Host|Redirected host is a valid IPv4 address, an A name pointing to this address, or a CNAME record.|
|int<2>|Time to Live|TTL is the life time of the validity of this redirected information with unit ‘second’.|

For recursion of redirection, in the scenario, the recursion is not useful. Should not recommend client side implement recursion of redirection.

[1] https://developer.mozilla.org/en-US/docs/Web/HTTP/Redirections
[2] https://docs.oracle.com/cd/B10501_01/network.920/a96580/concepts.htm
[3] https://mariadb.com/kb/en/library/1-connecting-connecting/

Alex Lee made changes - 2018-04-20 02:02

Description

h1. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.
h1. What to Do
Client/Server Protocol change needs both client and server support. The routine is
# Server sets the capability flag [3] in Initial Handshake Packet, indicating server supports MARIADB_SERVER_REDIRECTION.
| MARIADB_SERVER_REDIRECTION | 1 <<31 | Server support redirection |
# Client sets the capability flag [3] in Client Handshake Response Packet, indicating client supports MARIADB_SERVER_REDIRECTION.
# If client sets MARIADB_SERVER_REDIRECTION in Client Handshake Response Packet [3], server can send Redirection Information Packet other than OK_Packet / ERR_Packet / Authentication Switch Packet.
# Client receives Redirection Information Packet, close the connection, and then use its information to connect the new host.
# Client will use the redirection information to set up new connections until information expires or login parse error happens. The redirection information packet contains TTL of the information.

The Redirection Information Packet schema is

||Field Type||Field Name||Notes||
|int<1>|Packet Header|Packet header sets 0xFC to indicate it is a Redirection Information Packet.|
|int<1>|Protocol Version|Currently protocol version is 0, specifying TCP-IPv4 protocol.|
|int<2>|Protocol Property|For TCP-IPv4 protocol, it is the port number.|
|string<NUL>|Redirected Host|Redirected host is a valid IPv4 address, an A name pointing to this address, or a CNAME record.|
|int<2>|Time to Live|TTL is the life time of the validity of this redirected information with unit ‘second’.|

For recursion of redirection, in the scenario, the recursion is not useful. Should not recommend client side implement recursion of redirection.

[1] https://developer.mozilla.org/en-US/docs/Web/HTTP/Redirections
[2] https://docs.oracle.com/cd/B10501_01/network.920/a96580/concepts.htm
[3] https://mariadb.com/kb/en/library/1-connecting-connecting/

h2. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.
h2. What to Do
Client/Server Protocol change needs both client and server support. The routine is
# Server sets the capability flag [3] in Initial Handshake Packet, indicating server supports MARIADB_SERVER_REDIRECTION.
| MARIADB_SERVER_REDIRECTION | 1 <<31 | Server support redirection |
# Client sets the capability flag [3] in Client Handshake Response Packet, indicating client supports MARIADB_SERVER_REDIRECTION.
# If client sets MARIADB_SERVER_REDIRECTION in Client Handshake Response Packet [3], server can send Redirection Information Packet other than OK_Packet / ERR_Packet / Authentication Switch Packet.
# Client receives Redirection Information Packet, close the connection, and then use its information to connect the new host.
# Client will use the redirection information to set up new connections until information expires or login parse error happens. The redirection information packet contains TTL of the information.

The Redirection Information Packet schema is

||Field Type||Field Name||Notes||
|int<1>|Packet Header|Packet header sets 0xFC to indicate it is a Redirection Information Packet.|
|int<1>|Protocol Version|Currently protocol version is 0, specifying TCP-IPv4 protocol.|
|int<2>|Protocol Property|For TCP-IPv4 protocol, it is the port number.|
|string<NUL>|Redirected Host|Redirected host is a valid IPv4 address, an A name pointing to this address, or a CNAME record.|
|int<2>|Time to Live|TTL is the life time of the validity of this redirected information with unit ‘second’.|

For recursion of redirection, in the scenario, the recursion is not useful. Should not recommend client side implement recursion of redirection.

[1] https://developer.mozilla.org/en-US/docs/Web/HTTP/Redirections
[2] https://docs.oracle.com/cd/B10501_01/network.920/a96580/concepts.htm
[3] https://mariadb.com/kb/en/library/1-connecting-connecting/

Alex Lee made changes - 2018-04-20 05:01

Description

h2. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.
h2. What to Do
Client/Server Protocol change needs both client and server support. The routine is
# Server sets the capability flag [3] in Initial Handshake Packet, indicating server supports MARIADB_SERVER_REDIRECTION.
| MARIADB_SERVER_REDIRECTION | 1 <<31 | Server support redirection |
# Client sets the capability flag [3] in Client Handshake Response Packet, indicating client supports MARIADB_SERVER_REDIRECTION.
# If client sets MARIADB_SERVER_REDIRECTION in Client Handshake Response Packet [3], server can send Redirection Information Packet other than OK_Packet / ERR_Packet / Authentication Switch Packet.
# Client receives Redirection Information Packet, close the connection, and then use its information to connect the new host.
# Client will use the redirection information to set up new connections until information expires or login parse error happens. The redirection information packet contains TTL of the information.

The Redirection Information Packet schema is

||Field Type||Field Name||Notes||
|int<1>|Packet Header|Packet header sets 0xFC to indicate it is a Redirection Information Packet.|
|int<1>|Protocol Version|Currently protocol version is 0, specifying TCP-IPv4 protocol.|
|int<2>|Protocol Property|For TCP-IPv4 protocol, it is the port number.|
|string<NUL>|Redirected Host|Redirected host is a valid IPv4 address, an A name pointing to this address, or a CNAME record.|
|int<2>|Time to Live|TTL is the life time of the validity of this redirected information with unit ‘second’.|

For recursion of redirection, in the scenario, the recursion is not useful. Should not recommend client side implement recursion of redirection.

[1] https://developer.mozilla.org/en-US/docs/Web/HTTP/Redirections
[2] https://docs.oracle.com/cd/B10501_01/network.920/a96580/concepts.htm
[3] https://mariadb.com/kb/en/library/1-connecting-connecting/

h2. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.
h2. What to Do
Client/Server Protocol change needs both client and server support. The routine is
# Server sets the capability flag [3] in Initial Handshake Packet, indicating server supports MARIADB_SERVER_REDIRECTION.
| MARIADB_SERVER_REDIRECTION | 1 <<31 | Server support redirection |
# Client sets the capability flag [3] in Client Handshake Response Packet, indicating client supports MARIADB_SERVER_REDIRECTION.
# If client sets MARIADB_SERVER_REDIRECTION in Client Handshake Response Packet [3], server can send Redirection Information Packet other than OK_Packet / ERR_Packet / Authentication Switch Packet.
# Client receives Redirection Information Packet, close the connection, and then use its information to connect the new host.
# Client will use the redirection information to set up new connections until information expires or login parse error happens. The redirection information packet contains TTL of the information.

The Redirection Information Packet schema is

||Field Type||Field Name||Notes||
|int<1>|Packet Header|Packet header sets 0xFC to indicate it is a Redirection Information Packet.|
|int<1>|Protocol Version|Currently protocol version is 0, specifying TCP-IPv4 protocol.|
|int<2>|Protocol Property|For TCP-IPv4 protocol, it is the port number.|
|int<2>|Time to Live|TTL is the life time of the validity of this redirected information with unit ‘second’.|
|string<NUL>|Redirected Host|Redirected host is a valid IPv4 address, an A name pointing to this address, or a CNAME record.|

For recursion of redirection, in the scenario, the recursion is not useful. Should not recommend client side implement recursion of redirection.

[1] https://developer.mozilla.org/en-US/docs/Web/HTTP/Redirections
[2] https://docs.oracle.com/cd/B10501_01/network.920/a96580/concepts.htm
[3] https://mariadb.com/kb/en/library/1-connecting-connecting/

Alex Lee made changes - 2018-04-20 08:17

Description

h2. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.
h2. What to Do
Client/Server Protocol change needs both client and server support. The routine is
# Server sets the capability flag [3] in Initial Handshake Packet, indicating server supports MARIADB_SERVER_REDIRECTION.
| MARIADB_SERVER_REDIRECTION | 1 <<31 | Server support redirection |
# Client sets the capability flag [3] in Client Handshake Response Packet, indicating client supports MARIADB_SERVER_REDIRECTION.
# If client sets MARIADB_SERVER_REDIRECTION in Client Handshake Response Packet [3], server can send Redirection Information Packet other than OK_Packet / ERR_Packet / Authentication Switch Packet.
# Client receives Redirection Information Packet, close the connection, and then use its information to connect the new host.
# Client will use the redirection information to set up new connections until information expires or login parse error happens. The redirection information packet contains TTL of the information.

The Redirection Information Packet schema is

||Field Type||Field Name||Notes||
|int<1>|Packet Header|Packet header sets 0xFC to indicate it is a Redirection Information Packet.|
|int<1>|Protocol Version|Currently protocol version is 0, specifying TCP-IPv4 protocol.|
|int<2>|Protocol Property|For TCP-IPv4 protocol, it is the port number.|
|int<2>|Time to Live|TTL is the life time of the validity of this redirected information with unit ‘second’.|
|string<NUL>|Redirected Host|Redirected host is a valid IPv4 address, an A name pointing to this address, or a CNAME record.|

For recursion of redirection, in the scenario, the recursion is not useful. Should not recommend client side implement recursion of redirection.

[1] https://developer.mozilla.org/en-US/docs/Web/HTTP/Redirections
[2] https://docs.oracle.com/cd/B10501_01/network.920/a96580/concepts.htm
[3] https://mariadb.com/kb/en/library/1-connecting-connecting/

h2. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.
h2. What to Do
Client/Server Protocol change needs both client and server support. The routine is
# Server sets the capability flag [3] in Initial Handshake Packet, indicating server supports MARIADB_SERVER_REDIRECTION.
| MARIADB_SERVER_REDIRECTION | 1 <<31 | Server support redirection |
# Client sets the capability flag [3] in Client Handshake Response Packet, indicating client supports MARIADB_SERVER_REDIRECTION.
# If client sets MARIADB_SERVER_REDIRECTION in Client Handshake Response Packet [3], server can send Redirection Information Packet other than OK_Packet / ERR_Packet / Authentication Switch Packet.
# Client receives Redirection Information Packet, close the connection, and then use its information to connect the new host.
# Client will use the redirection information to set up new connections until information expires or login parse error happens. The redirection information packet contains TTL of the information.

The Redirection Information Packet schema is

||Field Type||Field Name||Notes||
|int<1>|Packet Header|Packet header sets 0xFC to indicate it is a Redirection Information Packet.|
|int<1>|Protocol Version|Currently protocol version is 0, specifying TCP-IPv4 protocol.|
|int<2>|Protocol Property|For TCP-IPv4 protocol, it is the port number.|
|int<2>|Time to Live|TTL is the life time of the validity of this redirected information with unit ‘second’.|
|string<NUL>|Redirected Host|Redirected host is a valid IPv4 address, an A name pointing to this address, or a CNAME record.|
|string<NUL>|Redirected User|Default will be the original user in connection string, but can used for proxy user.|

For recursion of redirection, in the scenario, the recursion is not useful. Should not recommend client side implement recursion of redirection.

[1] https://developer.mozilla.org/en-US/docs/Web/HTTP/Redirections
[2] https://docs.oracle.com/cd/B10501_01/network.920/a96580/concepts.htm
[3] https://mariadb.com/kb/en/library/1-connecting-connecting/

xiangyhu made changes - 2018-05-09 05:38

Attachment		redirection.pcapng [ 45610 ]
Attachment		redirection.png [ 45611 ]

libin made changes - 2018-07-02 08:13

Attachment

screenshot-1.png [ 45838 ]

libin made changes - 2018-07-02 08:16

Attachment

screenshot-2.png [ 45839 ]

libin made changes - 2018-07-02 12:38

Attachment

screenshot-3.png [ 45842 ]

libin made changes - 2018-07-02 12:39

Attachment

screenshot-4.png [ 45843 ]

Diego Dupin made changes - 2019-09-12 09:14

Link

This issue blocks ~~MDEV-20567~~ [ ~~MDEV-20567~~ ]

Diego Dupin made changes - 2020-04-24 09:31

Description

h2. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.
h2. What to Do
Client/Server Protocol change needs both client and server support. The routine is
# Server sets the capability flag [3] in Initial Handshake Packet, indicating server supports MARIADB_SERVER_REDIRECTION.
| MARIADB_SERVER_REDIRECTION | 1 <<31 | Server support redirection |
# Client sets the capability flag [3] in Client Handshake Response Packet, indicating client supports MARIADB_SERVER_REDIRECTION.
# If client sets MARIADB_SERVER_REDIRECTION in Client Handshake Response Packet [3], server can send Redirection Information Packet other than OK_Packet / ERR_Packet / Authentication Switch Packet.
# Client receives Redirection Information Packet, close the connection, and then use its information to connect the new host.
# Client will use the redirection information to set up new connections until information expires or login parse error happens. The redirection information packet contains TTL of the information.

The Redirection Information Packet schema is

||Field Type||Field Name||Notes||
|int<1>|Packet Header|Packet header sets 0xFC to indicate it is a Redirection Information Packet.|
|int<1>|Protocol Version|Currently protocol version is 0, specifying TCP-IPv4 protocol.|
|int<2>|Protocol Property|For TCP-IPv4 protocol, it is the port number.|
|int<2>|Time to Live|TTL is the life time of the validity of this redirected information with unit ‘second’.|
|string<NUL>|Redirected Host|Redirected host is a valid IPv4 address, an A name pointing to this address, or a CNAME record.|
|string<NUL>|Redirected User|Default will be the original user in connection string, but can used for proxy user.|

For recursion of redirection, in the scenario, the recursion is not useful. Should not recommend client side implement recursion of redirection.

[1] https://developer.mozilla.org/en-US/docs/Web/HTTP/Redirections
[2] https://docs.oracle.com/cd/B10501_01/network.920/a96580/concepts.htm
[3] https://mariadb.com/kb/en/library/1-connecting-connecting/

h2. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to Do
Client/Server Protocol change needs both client and server support.
Ending [OK_Packet |https://mariadb.com/kb/en/ok_packet/#fields] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.

info field format is then:

The routine is
# Server sets the capability flag [3] in Initial Handshake Packet, indicating server supports MARIADB_SERVER_REDIRECTION.
| MARIADB_SERVER_REDIRECTION | 1 <<31 | Server support redirection |
# Client sets the capability flag [3] in Client Handshake Response Packet, indicating client supports MARIADB_SERVER_REDIRECTION.
# If client sets MARIADB_SERVER_REDIRECTION in Client Handshake Response Packet [3], server can send Redirection Information Packet other than OK_Packet / ERR_Packet / Authentication Switch Packet.
# Client receives Redirection Information Packet, close the connection, and then use its information to connect the new host.
# Client will use the redirection information to set up new connections until information expires or login parse error happens. The redirection information packet contains TTL of the information.

The Redirection Information Packet schema is

||Field Type||Field Name||Notes||
|int<1>|Packet Header|Packet header sets 0xFC to indicate it is a Redirection Information Packet.|
|int<1>|Protocol Version|Currently protocol version is 0, specifying TCP-IPv4 protocol.|
|int<2>|Protocol Property|For TCP-IPv4 protocol, it is the port number.|
|int<2>|Time to Live|TTL is the life time of the validity of this redirected information with unit ‘second’.|
|string<NUL>|Redirected Host|Redirected host is a valid IPv4 address, an A name pointing to this address, or a CNAME record.|
|string<NUL>|Redirected User|Default will be the original user in connection string, but can used for proxy user.|

For recursion of redirection, in the scenario, the recursion is not useful. Should not recommend client side implement recursion of redirection.

[1] https://developer.mozilla.org/en-US/docs/Web/HTTP/Redirections
[2] https://docs.oracle.com/cd/B10501_01/network.920/a96580/concepts.htm
[3] https://mariadb.com/kb/en/library/1-connecting-connecting/

Diego Dupin made changes - 2020-04-24 09:44

Description

h2. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to Do
Client/Server Protocol change needs both client and server support.
Ending [OK_Packet |https://mariadb.com/kb/en/ok_packet/#fields] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.

info field format is then:

The routine is
# Server sets the capability flag [3] in Initial Handshake Packet, indicating server supports MARIADB_SERVER_REDIRECTION.
| MARIADB_SERVER_REDIRECTION | 1 <<31 | Server support redirection |
# Client sets the capability flag [3] in Client Handshake Response Packet, indicating client supports MARIADB_SERVER_REDIRECTION.
# If client sets MARIADB_SERVER_REDIRECTION in Client Handshake Response Packet [3], server can send Redirection Information Packet other than OK_Packet / ERR_Packet / Authentication Switch Packet.
# Client receives Redirection Information Packet, close the connection, and then use its information to connect the new host.
# Client will use the redirection information to set up new connections until information expires or login parse error happens. The redirection information packet contains TTL of the information.

The Redirection Information Packet schema is

||Field Type||Field Name||Notes||
|int<1>|Packet Header|Packet header sets 0xFC to indicate it is a Redirection Information Packet.|
|int<1>|Protocol Version|Currently protocol version is 0, specifying TCP-IPv4 protocol.|
|int<2>|Protocol Property|For TCP-IPv4 protocol, it is the port number.|
|int<2>|Time to Live|TTL is the life time of the validity of this redirected information with unit ‘second’.|
|string<NUL>|Redirected Host|Redirected host is a valid IPv4 address, an A name pointing to this address, or a CNAME record.|
|string<NUL>|Redirected User|Default will be the original user in connection string, but can used for proxy user.|

For recursion of redirection, in the scenario, the recursion is not useful. Should not recommend client side implement recursion of redirection.

[1] https://developer.mozilla.org/en-US/docs/Web/HTTP/Redirections
[2] https://docs.oracle.com/cd/B10501_01/network.920/a96580/concepts.htm
[3] https://mariadb.com/kb/en/library/1-connecting-connecting/

h2. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to Do
Client/Server Protocol change needs both client and server support.
Ending [OK_Packet |https://mariadb.com/kb/en/ok_packet/#fields] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.
There is no recursion, meaning that connector that have been redirect will not take another redirection in account.

info field format is then:

{code:java}
Location: mysql:[replication:|loadbalance:|sequential:]//<hostDescription>[,<hostDescription>...]/[database][?<key1>=<value1>[&<key2>=<value2>]]\n
{code}

HostDescription:

{code:java}
<host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
{code}

specific redirection option:

||option||description||
|user|if set, must superseed connection user.|
|ttl|validity timeout in second. This permit caching redirection client side. 0 means no caching 'default'|

Diego Dupin made changes - 2020-04-24 09:50

Link

This issue relates to CONC-468 [ CONC-468 ]

Diego Dupin made changes - 2020-04-24 09:51

Link

This issue relates to CONJS-135 [ CONJS-135 ]

Diego Dupin made changes - 2020-04-24 09:52

Link

This issue relates to CONPY-60 [ CONPY-60 ]

Diego Dupin made changes - 2020-04-24 09:53

Description

h2. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to Do
Client/Server Protocol change needs both client and server support.
Ending [OK_Packet |https://mariadb.com/kb/en/ok_packet/#fields] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.
There is no recursion, meaning that connector that have been redirect will not take another redirection in account.

info field format is then:

{code:java}
Location: mysql:[replication:|loadbalance:|sequential:]//<hostDescription>[,<hostDescription>...]/[database][?<key1>=<value1>[&<key2>=<value2>]]\n
{code}

HostDescription:

{code:java}
<host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
{code}

specific redirection option:

||option||description||
|user|if set, must superseed connection user.|
|ttl|validity timeout in second. This permit caching redirection client side. 0 means no caching 'default'|

h2. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to Do
Client/Server Protocol change needs both client and server support.
Ending connection [OK_Packet |https://mariadb.com/kb/en/ok_packet/#fields] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.
There is no recursion, meaning that connector that have been redirect will not take another redirection in account.

info field format is then:

{code:java}
Location: mysql:[replication:|loadbalance:|sequential:]//<hostDescription>[,<hostDescription>...]/[database][?<key1>=<value1>[&<key2>=<value2>]]\n
{code}

HostDescription:

{code:java}
<host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
{code}

specific redirection option:

||option||description||
|user|if set, must superseed connection user.|
|ttl|validity timeout in second. This permit caching redirection client side. 0 means no caching 'default'|

Diego Dupin made changes - 2020-04-24 10:00

Link

This issue relates to ODBC-281 [ ODBC-281 ]

Sergei Golubchik made changes - 2021-12-06 21:21

Workflow

MariaDB v3 [ 86662 ]

MariaDB v4 [ 130819 ]

markus makela made changes - 2022-06-10 06:20

Link

This issue blocks MXS-4153 [ MXS-4153 ]

Julien Fritsch made changes - 2022-06-10 09:35

Assignee

Julien Fritsch [ julien.fritsch ]

Julien Fritsch made changes - 2022-06-10 09:36

Assignee

Julien Fritsch [ julien.fritsch ]

Julien Fritsch made changes - 2022-06-10 09:39

Assignee

Julien Fritsch [ julien.fritsch ]

Julien Fritsch made changes - 2022-06-10 09:40

Assignee

Julien Fritsch [ julien.fritsch ]

Sergei Golubchik made changes - 2022-06-28 08:52

Link

This issue blocks CONC-599 [ CONC-599 ]

Sergei Golubchik made changes - 2022-06-28 08:53

Link

This issue blocks ~~CONJ-981~~ [ ~~CONJ-981~~ ]

Sergei Golubchik made changes - 2022-06-28 08:54

Link

This issue blocks ~~CONJS-207~~ [ ~~CONJS-207~~ ]

Sergei Golubchik made changes - 2022-06-28 08:54

Link

This issue blocks ~~R2DBC-66~~ [ ~~R2DBC-66~~ ]

Sergei Golubchik made changes - 2022-06-28 08:54

Link

This issue blocks CONTRIB-3 [ CONTRIB-3 ]

AirFocus made changes - 2022-08-09 16:11

Description

h2. Why
Redirection mechanism is widely used in proxy-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy-based scenario. It is more like HTTP redirects [1] or Oracle redirected connections [2]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to Do
Client/Server Protocol change needs both client and server support.
Ending connection [OK_Packet |https://mariadb.com/kb/en/ok_packet/#fields] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.
There is no recursion, meaning that connector that have been redirect will not take another redirection in account.

info field format is then:

{code:java}
Location: mysql:[replication:|loadbalance:|sequential:]//<hostDescription>[,<hostDescription>...]/[database][?<key1>=<value1>[&<key2>=<value2>]]\n
{code}

HostDescription:

{code:java}
<host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
{code}

specific redirection option:

||option||description||
|user|if set, must superseed connection user.|
|ttl|validity timeout in second. This permit caching redirection client side. 0 means no caching 'default'|

h2. Why

Redirection mechanism is widely used in proxy\-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy\-based scenario. It is more like HTTP redirects \[1\] or Oracle redirected connections \[2\]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to Do

Client/Server Protocol change needs both client and server support.
Ending connection \[OK_Packet \|https://mariadb.com/kb/en/ok_packet/#fields\] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.
There is no recursion, meaning that connector that have been redirect will not take another redirection in account.

info field format is then:

{code:java}
Location: mysql:[replication:|loadbalance:|sequential:]//<hostDescription>[,<hostDescription>...]/[database][?<key1>=<value1>[&<key2>=<value2>]]\n
{code}

HostDescription:

{code:java}
<host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
{code}

specific redirection option:

||option||description||
|user|if set, must superseed connection user.|
|ttl|validity timeout in second. This permit caching redirection client side. 0 means no caching 'default'|

Sergei Golubchik made changes - 2023-04-21 11:09

Description

h2. Why

Redirection mechanism is widely used in proxy\-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy\-based scenario. It is more like HTTP redirects \[1\] or Oracle redirected connections \[2\]. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to Do

Client/Server Protocol change needs both client and server support.
Ending connection \[OK_Packet \|https://mariadb.com/kb/en/ok_packet/#fields\] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.
There is no recursion, meaning that connector that have been redirect will not take another redirection in account.

info field format is then:

{code:java}
Location: mysql:[replication:|loadbalance:|sequential:]//<hostDescription>[,<hostDescription>...]/[database][?<key1>=<value1>[&<key2>=<value2>]]\n
{code}

HostDescription:

{code:java}
<host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
{code}

specific redirection option:

||option||description||
|user|if set, must superseed connection user.|
|ttl|validity timeout in second. This permit caching redirection client side. 0 means no caching 'default'|

h2. Why

Redirection mechanism is widely used in proxy\-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy\-based scenario. It is more like HTTP redirects or Oracle redirected connections. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to do

We'll add a new global/session variable, say, {{redirect_url}}. The value should be an empty string or a connection string in the conventional format (in the style of a connection url of C/J or C/NodeJS or C/C or FederatedX).

Internally the server might have a check to ensure that the value of this variable is a valid url.

This variable is appended to the default value of {{session_track_system_variables}} variable.

When this variable is changed, the existing {{SESSION_TRACK_SYSTEM_VARIABLES}} implementation will send its new value to the client as a part of the OK packet. It's up to the client or to the connector to use this value or to ignore it.

h3. Thoughts

* perhaps the server should always send this variable if it's set, even if it was not just changed. This will require some tweaking of the sysvar tracker code.
* the server can disallow changing of this variable in an active transaction. Or it can delegate this check to connectors, perhaps a connector would want to reconnect and replay the transaction?

h3. Possible use cases:

* always redirect all clients to a new location
** set {{@@global.redirect_url}} or start the server with {{--redurect-url=}} or put it in my.cnf
* redirect to a group of servers randomly
** create a table with connection urls, one per row.
** use an sql script that selects a random row from it and sets {{@@redirect_url}} to this value
** specify this script in the {{--init-connect}} server parameter
* dynamically redirect from the master to one of the slaves
** same as above, but use {{INFORMATION_SCHEMA.PROCESSLIST}} to get the list of active slaves.

and so on. An ability to redirect any time using SQL to calculate a new location opens limitless possibilities.

{panel:title=Original specs for comparison:|bgColor=#EEEEEE}
h2. What to Do

Client/Server Protocol change needs both client and server support.
Ending connection \[OK_Packet \|https://mariadb.com/kb/en/ok_packet/#fields\] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.
There is no recursion, meaning that connector that have been redirect will not take another redirection in account.

info field format is then:

{code:java}
Location: mysql:[replication:|loadbalance:|sequential:]//<hostDescription>[,<hostDescription>...]/[database][?<key1>=<value1>[&<key2>=<value2>]]\n
{code}

HostDescription:

{code:java}
<host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
{code}

specific redirection option:

||option||description||
|user|if set, must superseed connection user.|
|ttl|validity timeout in second. This permit caching redirection client side. 0 means no caching 'default'|
{panel}

markus makela made changes - 2023-06-09 05:26

Link

This issue relates to ~~MXS-4635~~ [ ~~MXS-4635~~ ]

Sergei Golubchik made changes - 2023-06-21 17:06

Fix Version/s

11.3 [ 28565 ]

Sergei Golubchik made changes - 2023-06-22 07:38

Description

h2. Why

Redirection mechanism is widely used in proxy\-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy\-based scenario. It is more like HTTP redirects or Oracle redirected connections. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to do

We'll add a new global/session variable, say, {{redirect_url}}. The value should be an empty string or a connection string in the conventional format (in the style of a connection url of C/J or C/NodeJS or C/C or FederatedX).

Internally the server might have a check to ensure that the value of this variable is a valid url.

This variable is appended to the default value of {{session_track_system_variables}} variable.

When this variable is changed, the existing {{SESSION_TRACK_SYSTEM_VARIABLES}} implementation will send its new value to the client as a part of the OK packet. It's up to the client or to the connector to use this value or to ignore it.

h3. Thoughts

* perhaps the server should always send this variable if it's set, even if it was not just changed. This will require some tweaking of the sysvar tracker code.
* the server can disallow changing of this variable in an active transaction. Or it can delegate this check to connectors, perhaps a connector would want to reconnect and replay the transaction?

h3. Possible use cases:

* always redirect all clients to a new location
** set {{@@global.redirect_url}} or start the server with {{--redurect-url=}} or put it in my.cnf
* redirect to a group of servers randomly
** create a table with connection urls, one per row.
** use an sql script that selects a random row from it and sets {{@@redirect_url}} to this value
** specify this script in the {{--init-connect}} server parameter
* dynamically redirect from the master to one of the slaves
** same as above, but use {{INFORMATION_SCHEMA.PROCESSLIST}} to get the list of active slaves.

and so on. An ability to redirect any time using SQL to calculate a new location opens limitless possibilities.

{panel:title=Original specs for comparison:|bgColor=#EEEEEE}
h2. What to Do

Client/Server Protocol change needs both client and server support.
Ending connection \[OK_Packet \|https://mariadb.com/kb/en/ok_packet/#fields\] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.
There is no recursion, meaning that connector that have been redirect will not take another redirection in account.

info field format is then:

{code:java}
Location: mysql:[replication:|loadbalance:|sequential:]//<hostDescription>[,<hostDescription>...]/[database][?<key1>=<value1>[&<key2>=<value2>]]\n
{code}

HostDescription:

{code:java}
<host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
{code}

specific redirection option:

||option||description||
|user|if set, must superseed connection user.|
|ttl|validity timeout in second. This permit caching redirection client side. 0 means no caching 'default'|
{panel}

h2. Why

Redirection mechanism is widely used in proxy\-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy\-based scenario. It is more like HTTP redirects or Oracle redirected connections. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to do

We'll add a new global/session variable, say, {{redirect_url}}. The value should be an empty string or a connection string in the conventional format (in the style of a connection url of C/J or C/NodeJS or C/C or FederatedX).

Internally the server might have a check to ensure that the value of this variable is a valid url.

This variable is appended to the default value of {{session_track_system_variables}} variable.

When this variable is changed, the existing {{SESSION_TRACK_SYSTEM_VARIABLES}} implementation will send its new value to the client as a part of the OK packet. It's up to the client or to the connector to use this value or to ignore it.

h3. Thoughts

* perhaps the server should always send this variable if it's set, even if it was not just changed. This will require some tweaking of the sysvar tracker code.
* the server should not disallow changing of this variable in an active transaction. This is up to connectors, perhaps a connector would want to reconnect and replay the transaction.

h3. Possible use cases:

* always redirect all clients to a new location
** set {{@@global.redirect_url}} or start the server with {{--redurect-url=}} or put it in my.cnf
* redirect to a group of servers randomly
** create a table with connection urls, one per row.
** use an sql script that selects a random row from it and sets {{@@redirect_url}} to this value
** specify this script in the {{--init-connect}} server parameter
* dynamically redirect from the master to one of the slaves
** same as above, but use {{INFORMATION_SCHEMA.PROCESSLIST}} to get the list of active slaves.

and so on. An ability to redirect any time using SQL to calculate a new location opens limitless possibilities.

{panel:title=Original specs for comparison:|bgColor=#EEEEEE}
h2. What to Do

Client/Server Protocol change needs both client and server support.
Ending connection \[OK_Packet \|https://mariadb.com/kb/en/ok_packet/#fields\] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.
There is no recursion, meaning that connector that have been redirect will not take another redirection in account.

info field format is then:

{code:java}
Location: mysql:[replication:|loadbalance:|sequential:]//<hostDescription>[,<hostDescription>...]/[database][?<key1>=<value1>[&<key2>=<value2>]]\n
{code}

HostDescription:

{code:java}
<host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
{code}

specific redirection option:

||option||description||
|user|if set, must superseed connection user.|
|ttl|validity timeout in second. This permit caching redirection client side. 0 means no caching 'default'|
{panel}

Sergei Golubchik made changes - 2023-06-22 07:38

Description

h2. Why

Redirection mechanism is widely used in proxy\-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy\-based scenario. It is more like HTTP redirects or Oracle redirected connections. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to do

We'll add a new global/session variable, say, {{redirect_url}}. The value should be an empty string or a connection string in the conventional format (in the style of a connection url of C/J or C/NodeJS or C/C or FederatedX).

Internally the server might have a check to ensure that the value of this variable is a valid url.

This variable is appended to the default value of {{session_track_system_variables}} variable.

When this variable is changed, the existing {{SESSION_TRACK_SYSTEM_VARIABLES}} implementation will send its new value to the client as a part of the OK packet. It's up to the client or to the connector to use this value or to ignore it.

h3. Thoughts

* perhaps the server should always send this variable if it's set, even if it was not just changed. This will require some tweaking of the sysvar tracker code.
* the server should not disallow changing of this variable in an active transaction. This is up to connectors, perhaps a connector would want to reconnect and replay the transaction.

h3. Possible use cases:

* always redirect all clients to a new location
** set {{@@global.redirect_url}} or start the server with {{--redurect-url=}} or put it in my.cnf
* redirect to a group of servers randomly
** create a table with connection urls, one per row.
** use an sql script that selects a random row from it and sets {{@@redirect_url}} to this value
** specify this script in the {{--init-connect}} server parameter
* dynamically redirect from the master to one of the slaves
** same as above, but use {{INFORMATION_SCHEMA.PROCESSLIST}} to get the list of active slaves.

and so on. An ability to redirect any time using SQL to calculate a new location opens limitless possibilities.

{panel:title=Original specs for comparison:|bgColor=#EEEEEE}
h2. What to Do

Client/Server Protocol change needs both client and server support.
Ending connection \[OK_Packet \|https://mariadb.com/kb/en/ok_packet/#fields\] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.
There is no recursion, meaning that connector that have been redirect will not take another redirection in account.

info field format is then:

{code:java}
Location: mysql:[replication:|loadbalance:|sequential:]//<hostDescription>[,<hostDescription>...]/[database][?<key1>=<value1>[&<key2>=<value2>]]\n
{code}

HostDescription:

{code:java}
<host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
{code}

specific redirection option:

||option||description||
|user|if set, must superseed connection user.|
|ttl|validity timeout in second. This permit caching redirection client side. 0 means no caching 'default'|
{panel}

h2. Why

Redirection mechanism is widely used in proxy\-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy\-based scenario. It is more like HTTP redirects or Oracle redirected connections. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to do

We'll add a new global/session variable, say, {{redirect_url}}. The value should be an empty string or a connection string in the conventional format (in the style of a connection url of C/J or C/NodeJS or C/C or FederatedX).

Internally the server might have a check to ensure that the value of this variable is a valid url.

This variable is appended to the default value of {{session_track_system_variables}} variable.

When this variable is changed, the existing {{SESSION_TRACK_SYSTEM_VARIABLES}} implementation will send its new value to the client as a part of the OK packet. It's up to the client or to the connector to use this value or to ignore it.

Note that the server should not disallow changing of this variable in an active transaction. This is up to connectors, perhaps a connector would want to reconnect and replay the transaction.

h3. Thoughts

* perhaps the server should always send this variable if it's set, even if it was not just changed. This will require some tweaking of the sysvar tracker code.

h3. Possible use cases:

* always redirect all clients to a new location
** set {{@@global.redirect_url}} or start the server with {{--redurect-url=}} or put it in my.cnf
* redirect to a group of servers randomly
** create a table with connection urls, one per row.
** use an sql script that selects a random row from it and sets {{@@redirect_url}} to this value
** specify this script in the {{--init-connect}} server parameter
* dynamically redirect from the master to one of the slaves
** same as above, but use {{INFORMATION_SCHEMA.PROCESSLIST}} to get the list of active slaves.

and so on. An ability to redirect any time using SQL to calculate a new location opens limitless possibilities.

{panel:title=Original specs for comparison:|bgColor=#EEEEEE}
h2. What to Do

Client/Server Protocol change needs both client and server support.
Ending connection \[OK_Packet \|https://mariadb.com/kb/en/ok_packet/#fields\] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.
There is no recursion, meaning that connector that have been redirect will not take another redirection in account.

info field format is then:

{code:java}
Location: mysql:[replication:|loadbalance:|sequential:]//<hostDescription>[,<hostDescription>...]/[database][?<key1>=<value1>[&<key2>=<value2>]]\n
{code}

HostDescription:

{code:java}
<host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
{code}

specific redirection option:

||option||description||
|user|if set, must superseed connection user.|
|ttl|validity timeout in second. This permit caching redirection client side. 0 means no caching 'default'|
{panel}

Yuchen Pei made changes - 2023-06-22 07:53

Assignee

Yuchen Pei [ JIRAUSER52627 ]

Sergei Golubchik made changes - 2023-07-06 10:58

Link

This issue relates to ~~MDEV-31609~~ [ ~~MDEV-31609~~ ]

Yuchen Pei made changes - 2023-08-08 06:50

Status

Open [ 1 ]

In Progress [ 3 ]

Yuchen Pei made changes - 2023-08-08 08:08

Priority

Major [ 3 ]

Critical [ 2 ]

Yuchen Pei made changes - 2023-08-15 08:14

Assignee	Yuchen Pei [ JIRAUSER52627 ]	Sergei Golubchik [ serg ]
Status	In Progress [ 3 ]	In Review [ 10002 ]

Sergei Golubchik made changes - 2023-08-16 08:16

Description

h2. Why

Redirection mechanism is widely used in proxy\-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy\-based scenario. It is more like HTTP redirects or Oracle redirected connections. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to do

We'll add a new global/session variable, say, {{redirect_url}}. The value should be an empty string or a connection string in the conventional format (in the style of a connection url of C/J or C/NodeJS or C/C or FederatedX).

Internally the server might have a check to ensure that the value of this variable is a valid url.

This variable is appended to the default value of {{session_track_system_variables}} variable.

When this variable is changed, the existing {{SESSION_TRACK_SYSTEM_VARIABLES}} implementation will send its new value to the client as a part of the OK packet. It's up to the client or to the connector to use this value or to ignore it.

Note that the server should not disallow changing of this variable in an active transaction. This is up to connectors, perhaps a connector would want to reconnect and replay the transaction.

h3. Thoughts

* perhaps the server should always send this variable if it's set, even if it was not just changed. This will require some tweaking of the sysvar tracker code.

h3. Possible use cases:

* always redirect all clients to a new location
** set {{@@global.redirect_url}} or start the server with {{--redurect-url=}} or put it in my.cnf
* redirect to a group of servers randomly
** create a table with connection urls, one per row.
** use an sql script that selects a random row from it and sets {{@@redirect_url}} to this value
** specify this script in the {{--init-connect}} server parameter
* dynamically redirect from the master to one of the slaves
** same as above, but use {{INFORMATION_SCHEMA.PROCESSLIST}} to get the list of active slaves.

and so on. An ability to redirect any time using SQL to calculate a new location opens limitless possibilities.

{panel:title=Original specs for comparison:|bgColor=#EEEEEE}
h2. What to Do

Client/Server Protocol change needs both client and server support.
Ending connection \[OK_Packet \|https://mariadb.com/kb/en/ok_packet/#fields\] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.
There is no recursion, meaning that connector that have been redirect will not take another redirection in account.

info field format is then:

{code:java}
Location: mysql:[replication:|loadbalance:|sequential:]//<hostDescription>[,<hostDescription>...]/[database][?<key1>=<value1>[&<key2>=<value2>]]\n
{code}

HostDescription:

{code:java}
<host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
{code}

specific redirection option:

||option||description||
|user|if set, must superseed connection user.|
|ttl|validity timeout in second. This permit caching redirection client side. 0 means no caching 'default'|
{panel}

h2. Why

Redirection mechanism is widely used in proxy\-based scenario.
Currently when multiple servers share one proxy, proxy forwards all the packets between servers and clients. Thus, the proxy adds latency, take many computing resources and impacts the overall performance. Especially for short connections scenarios like WordPress, the latency can be a critical issue.
Supporting redirection mechanism is helpful for proxy\-based scenario. It is more like HTTP redirects or Oracle redirected connections. Clients get the servers’ address from proxies and connect to servers transparently without latency and computing resource wasted.

h2. What to do

We'll add a new global/session variable, say, {{redirect_url}}. The value should be an empty string or a connection string in the conventional format (in the style of a connection url of C/J or C/NodeJS or C/C or FederatedX).

Internally the server might have a check to ensure that the value of this variable is a valid url.

This variable is appended to the default value of {{session_track_system_variables}} variable.

When this variable is changed, the existing {{SESSION_TRACK_SYSTEM_VARIABLES}} implementation will send its new value to the client as a part of the OK packet. It's up to the client or to the connector to use this value or to ignore it.

Note that the server should not disallow changing of this variable in an active transaction. This is up to connectors, perhaps a connector would want to reconnect and replay the transaction.

h3. Thoughts

* perhaps the server should always send this variable if it's set, even if it was not just changed. This will require some tweaking of the sysvar tracker code.

h3. Possible use cases:

* always redirect all clients to a new location
** set {{@@global.redirect_url}} or start the server with {{--redurect-url=}} or put it in my.cnf
* redirect to a group of servers randomly
** create a table with connection urls, one per row.
** use an sql script that selects a random row from it and sets {{@@redirect_url}} to this value
** specify this script in the {{--init-connect}} server parameter
* dynamically redirect from the master to one of the slaves
** same as above, but use {{INFORMATION_SCHEMA.PROCESSLIST}} to get the list of active slaves.

and so on. An ability to redirect any time using SQL to calculate a new location opens limitless possibilities.

{panel:title=Original specs for comparison:|bgColor=#EEEEEE}
h2. What to Do

Client/Server Protocol change needs both client and server support.
Ending connection [OK_Packet|https://mariadb.com/kb/en/ok_packet/#fields] info field might contain redirection information.

Either connector support redirection mechanism and will close current connection and redirect to indicated value, or if not supporting redirection (or redirection is disable) then driver will continue using current connection.
There is no recursion, meaning that connector that have been redirect will not take another redirection in account.

info field format is then:

{code:java}
Location: mysql:[replication:|loadbalance:|sequential:]//<hostDescription>[,<hostDescription>...]/[database][?<key1>=<value1>[&<key2>=<value2>]]\n
{code}

HostDescription:

{code:java}
<host>[:<portnumber>] or address=(host=<host>)[(port=<portnumber>)][(type=(master|slave))]
{code}

specific redirection option:

||option||description||
|user|if set, must superseed connection user.|
|ttl|validity timeout in second. This permit caching redirection client side. 0 means no caching 'default'|
{panel}

Sergei Golubchik made changes - 2023-09-01 17:22

Assignee	Sergei Golubchik [ serg ]	Yuchen Pei [ JIRAUSER52627 ]
Status	In Review [ 10002 ]	Stalled [ 10000 ]

Yuchen Pei made changes - 2023-09-08 04:51

Assignee	Yuchen Pei [ JIRAUSER52627 ]	Sergei Golubchik [ serg ]
Status	Stalled [ 10000 ]	In Review [ 10002 ]

Sergei Golubchik made changes - 2023-09-08 19:33

Assignee	Sergei Golubchik [ serg ]	Yuchen Pei [ JIRAUSER52627 ]
Status	In Review [ 10002 ]	Stalled [ 10000 ]

Yuchen Pei made changes - 2023-09-12 06:57

Assignee	Yuchen Pei [ JIRAUSER52627 ]	Sergei Golubchik [ serg ]
Status	Stalled [ 10000 ]	In Review [ 10002 ]

Sergei Golubchik made changes - 2023-09-12 20:38

Link

This issue blocks TODO-4253 [ TODO-4253 ]

Sergei Golubchik made changes - 2023-09-13 21:01

Assignee	Sergei Golubchik [ serg ]	Yuchen Pei [ JIRAUSER52627 ]
Status	In Review [ 10002 ]	Stalled [ 10000 ]

Yuchen Pei made changes - 2023-09-14 00:46

Link

This issue blocks TODO-4253 [ TODO-4253 ]

Yuchen Pei made changes - 2023-09-14 00:46

Link

This issue is part of TODO-4253 [ TODO-4253 ]

Yuchen Pei made changes - 2023-09-14 03:11

Status

Stalled [ 10000 ]

In Testing [ 10301 ]

Yuchen Pei made changes - 2023-09-14 03:11

Assignee

Yuchen Pei [ JIRAUSER52627 ]

Ramesh Sivaraman [ JIRAUSER48189 ]

Ralf Gebhardt made changes - 2023-09-21 11:56

Labels

client protocol proxy

Preview_11.3 client protocol proxy

Ramesh Sivaraman made changes - 2023-09-26 13:23

Link

This issue causes ~~MDEV-32254~~ [ ~~MDEV-32254~~ ]

Ramesh Sivaraman made changes - 2023-10-18 11:47

Assignee	Ramesh Sivaraman [ JIRAUSER48189 ]	Yuchen Pei [ JIRAUSER52627 ]
Status	In Testing [ 10301 ]	Stalled [ 10000 ]

Yuchen Pei made changes - 2023-10-20 05:36

Link

This issue causes ~~MDEV-32525~~ [ ~~MDEV-32525~~ ]

Yuchen Pei made changes - 2023-10-20 05:37

Link

This issue is blocked by ~~MDEV-32525~~ [ ~~MDEV-32525~~ ]

Yuchen Pei made changes - 2023-10-25 23:17

Status

Stalled [ 10000 ]

In Testing [ 10301 ]

Yuchen Pei made changes - 2023-10-25 23:17

Assignee

Yuchen Pei [ JIRAUSER52627 ]

Ramesh Sivaraman [ JIRAUSER48189 ]

Ramesh Sivaraman made changes - 2023-10-26 06:29

Assignee	Ramesh Sivaraman [ JIRAUSER48189 ]	Yuchen Pei [ JIRAUSER52627 ]
Status	In Testing [ 10301 ]	Stalled [ 10000 ]

Ramesh Sivaraman made changes - 2023-10-26 06:31

Comment

[ 1h 30m ]

Yuchen Pei made changes - 2023-10-26 07:30

Fix Version/s		11.3.1 [ 29416 ]
Fix Version/s	11.3 [ 28565 ]
Resolution		Fixed [ 1 ]
Status	Stalled [ 10000 ]	Closed [ 6 ]

Sergei Golubchik made changes - 2023-12-21 17:25

Component/s		Protocol [ 14604 ]
Component/s	Server [ 13907 ]

Daniel Lenski (Inactive) made changes - 2024-04-26 20:57

Link

This issue relates to ~~MDEV-34009~~ [ ~~MDEV-34009~~ ]

Jira Automation (IT) made changes - 2024-07-04 07:39

Zendesk Related Tickets

202147 119111

markus makela made changes - 2024-09-04 14:51

Link

This issue relates to ~~MXS-5238~~ [ ~~MXS-5238~~ ]

MariaDB Server

Connection Redirection Mechanism in MariaDB Client/Server Protocol

Details

Description

Why

What to do

Thoughts

Possible use cases:

What to Do

Attachments

Attachments

Issue Links

Activity

People

Dates

Git Integration