RPC

CRIU-RPC is a remote procedure call (RPC) protocol which uses Google Protocol Buffers to encode its calls. The requests are served by CRIU service launched with criu service command. It uses a SEQPACKET Unix domain socket listening at /var/run/criu-service.socket as a transport.

The criu_req/criu_resp are two main messages for requests/responses. They are to be used for transferring messages and needed to provide compatibility with an older versions of rpc. Field type in them must be set accordingly to type of request/response that is stored. Types of request/response are defined in enum criu_req_type. See the API compliance page for information what each option might mean.

Protocol

The protocol is simple -- client connect to service, send a criu_req message to server, server responds back with criu_resp and closes the socket. There are three exceptions from this rule, see below.

Request

This is the header of the request. It defines the operation requested and options.

message criu_req {
	required criu_req_type type	= 1;
	optional criu_opts opts		= 2;
	optional notify_success         = 3; /* see Notifications below */
	optional keep_open              = 4; /* for multi-req, below */
}

There are 8 request/response types for now.

enum criu_req_type {
	EMPTY		= 0;
	DUMP		= 1; /* criu dump */
	RESTORE		= 2; /* criu restore */
	CHECK		= 3; /* criu check */
	PRE_DUMP	= 4; /* criu pre-dump */
	PAGE_SERVER	= 5; /* criu page-server */
	NOTIFY          = 6; /* see Notifications below */
	CPUINFO_DUMP	= 7; /* criu cpuinfo dump */
	CPUINFO_CHECK	= 8; /* criu cpuinfo check */
}

message criu_opts {
	required int32			images_dir_fd	= 1;
	optional int32			pid		= 2; /* if not set on dump, will dump requesting process */

	optional bool			leave_running	= 3;
	optional bool			ext_unix_sk	= 4;
	optional bool			tcp_established	= 5;
	optional bool			evasive_devices	= 6;
	optional bool			shell_job	= 7;
	optional bool			file_locks	= 8;
	optional int32			log_level	= 9 [default = 2];
	optional string			log_file	= 10; /* No subdirs are allowed. Consider using work-dir */

	optional criu_page_server_info	ps		= 11;

	optional bool			notify_scripts	= 12;

	optional string			root		= 13;
	optional string			parent_img	= 14;
	optional bool			track_mem	= 15;
	optional bool			auto_dedup	= 16;

	optional int32			work_dir_fd	= 17;
	optional bool			link_remap	= 18;
	repeated criu_veth_pair		veths		= 19;

	optional uint32			cpu_cap		= 20 [default = 0xffffffff];
	optional bool			force_irmap	= 21;
	repeated string			exec_cmd	= 22;

	repeated ext_mount_map		ext_mnt		= 23;
	optional bool			manage_cgroups	= 24;
	repeated cgroup_root		cg_root		= 25;

	optional bool			rst_sibling	= 26; /* swrk only */
}

Comments and examples

• If no pid is set and type is DUMP, CRIU will dump client process by default.
• All processes in the subtree starting with <pid> must have the same uid, as a client, or client's uid must be root (uid == 0), otherwise CRIU will return error.
• Only the images_dir_fd is required, all other fields are optional. Client must open directory for/with images by himself and set images_dir_fd to it's fd. CRIU will open /proc/<client's_pid>/fd/<images_dir_fd>.

The logic of setting request is the same as when setting options in console. Here is an example:

#criu restore -D /path/to/imgs_dir -v4 -o restore.log

is equal to:

request.type = RESTORE;

request.opts.imgs_dir_fd	= open("/path/to/imgs_dir")
request.opts.log_level		= 4
request.opts.log_file		= "restore.log"

Sub-messages for options

Info about page-server.

message criu_page_server_info {
	optional string		address	= 1; /* bind address -- if not set 0.0.0.0 is used */
	optional int32		port	= 2; /* bind port -- if not set on request, autobind is used and port is returned in response */
	optional int32		pid	= 3; /* page-server pid -- returned in response */
	optional int32		fd	= 4; /* could be used to inherit fd by page-server */
}

• If port is 0 server will auto-bind and the port will be returned back in responce.

message criu_veth_pair {
	required string		if_in	= 1; /* inside veth device name */
	required string		if_out	= 2; /* outside veth device name */
};

Info about veth mappings (--ext-mount-map analogue)

message ext_mount_map {
	required string		key	= 1;
	required string		val	= 2;
};

Specifying where cgroup root should be (--cgroup-root analogue)

message cgroup_root {
	optional string		ctrl	= 1;
	required string		path	= 2;
};

Response

This message is sent after (un)successful execution of the request.

message criu_resp {
	required criu_req_type		type		= 1;
	required bool			success		= 2;

	optional criu_dump_resp		dump		= 3;
	optional criu_restore_resp	restore		= 4;
	optional criu_notify		notify		= 5;
	optional criu_page_server_info	ps		= 6;

        optional int32			cr_errno	= 7;
}

Field "success" reports result of processing request, while criu_***_resp store some request-specific information. The response type is set to the corresponding request type or to EMPTY to report a "generic" error. If success == false one should check cr_errno field to get a more detailed error code(see include/cr-errno.h).

The criu_dump_resp is used to store response from DUMP request.

message criu_dump_resp {
	optional bool restored		= 1;
}

This message can be sent twice -- one for the process that calls DUMP and the other for the same process again if it requested a self-dump. In the latter case the restored field would be true.

The response on RESTORE request.

message criu_restore_resp {
	required int32 pid		= 1;
}

Field "pid" is set to the PID of the newly restored process.

Info about page server

The criu_page_server_info from requests will be sent back on PAGE_SERVER request. The port field will contain the port to which server is bound.

Notifications

If the opts.notify_scripts in the request is set to TRUE CRIU would report back resp messages with type set to NOTIFY and this field present. The notifications are the way action scripts work for RPC mode.

message criu_notify {
	optional string script		= 1;
	optional int32	pid		= 2;
}

After handling the notification the client must response with the request again with the type set to NOTIFY and the notify_success set to the successfulness of the notification. In case of successful notification acknowledge the server doesn't close the socket and continues to work.

Pre-dumps

Before issuing a DUMP request client may send one or more PRE_DUMP requests. Once the PRE_DUMP is sent and response is received, client may send one more PRE_DUMP or DUMP request. The server would only close the socket after the DUMP one.

Multi-request mode

If the req.keep_open flag is set to true server will not close the socket after response, but will wait for more requests. This mode is supported only for the following request types:

• PRE_DUMP (automatically)
• PAGE_SERVER
• CPUINFO_DUMP and CPUINFO_CHECK

Run

Server

On a server side, CRIU creates SOCK_SEQPACKET Unix socket and listens for connections on it. After receiving criu_req, CRIU processes it, do what is requested and sends criu_resp with set request-specific criu_***_resp field back. If CRIU gets unknown type of request, it will return criu_resp with type == EMPTY and success == false.

To launch service server, run:

#criu service [options]

Options accepted by service are

--address <path>

is where to put listening socket

--pid-file <path>

is where to write pid of service process

--daemon

tells service to daemonize

-o <file>

says where to write logs

-v[N]

sets the log-level

systemd

If you are running systemd you can make service start and operate automatically. First do

make install

to make files criu.service and criu.socket appear in systemd configs (/lib/systemd/system/). Then

systemctl start criu.socket

to get /var/run/criu-service.socket, and

systemctl enable criu.socket

to make /var/run/criu-service.socket available at boot.

Client

Client, in its turn, must connect to service socket, send criu_req with request in it, and wait for a criu_resp with response. You can find examples of client programs in C and Python in test/rpc/.

With RPC facilities one can perform a self dump.

There's a library that implements simple wrappers on top of RPC.

SWRK mode

Note: This mode is implemented by not yet standard. The API can change at any time

This mode turns on when you fork() + exec() CRIU with the swrk action. With this CRIU works as service worker task accepting standard RPC requests.