Chirp User's Manual

Overview

Chirp is a system for performing input and output across the Internet. Using Chirp, an ordinary user can share storage space and data with friends and colleagues without requiring any sort of administrator privileges anywhere.

Chirp is like a distributed filesystem (such as NFS) except that it can be run over wide area networks and requires no special privileges on either the client or the server end. Chirp allows the end user to set up fine-grained access control so that data can be shared (or not shared) with the right people.

Chirp is also like a file transfer system (such as FTP) that provides streaming point-to-point data transfer over the Internet. However, Chirp also provides fine-grained Unix-like data access suitable for direct access by ordinary programs.

Chirp also includes several advanced features for authentication tickets, space allocation, and more. However, each of these features must be explicitly enabled, so you don't have to worry about them if all you want is simple storage access. Read on below for more details.

Getting Started

Installing

See the Installation Instructions for the Cooperative Computing Tools package. Then, Make sure to set your PATH appropriately.

Running a Chirp Server

Running a Chirp server is easy. You may run a Chirp server as any ordinary user, and you do not need to install the software or even run the programs as root. To run a Chirp server, you must do three things: pick a storage directory, run the server, and then adjust the access control.

• Pick a storage directory. The Chirp server will only allow access to the directory that you choose. It could be a scratch directory, your home directory, or even your filesystem root. For now, let's store everything in a temporary directory: /tmp/mydata

• Run the server. Run chirp_server and direct it to your storage directory:

$ chirp_server -r /tmp/mydata &

• Adjust the access control.* When first started, the Chirp server will allow access only to YOU from the same host. You will probably want to change this to allow access to other people and hosts. To adjust the access control, use the chirp tool and the setacl command to set the access control list. For example, to also allow other hosts in your domain to read and write the server:

$ chirp localhost
chirp:localhost:/>
chirp:localhost:/> setacl . hostname:*.mydomain.edu write

Now that you have a server running on one machine, let's use some tools to move data to and from your server.

Accessing Chirp Servers

The easiest way to access Chirp servers is by using a tool called Parrot. Parrot is a personal virtual filesystem: it "speaks" remote I/O operations on behalf of ordinary programs. For example, you can use Parrot with your regular shell to list and access Chirp servers like so:

$ parrot_run bash
$ cd /chirp
$ ls angband.somewhere.edu:9094 dustpuppy.somewhere.edu:9094 peanuts.somewhere.edu:9094 ...
$ cd /chirp/peanuts.somewhere.edu
$ cp /tmp/bigfile .
$ ls -la

total 804
drwx------ 2 fred users 4096      Sep 10 12:40 .
drwx------ 2 fred users 4096      Sep 10 12:40 ..
-rw-r--r-- 1 fred users 104857600 Sep 10 12:57 bigfile
-rw-r--r-- 1 fred users 147       Sep 10 12:39 hosts

$ parrot_getacl unix:fred rwlda hostname:hedwig rl ...

If you are having difficulting accessing your server, have a look at debugging hints below.

Parrot is certainly the most convenient way to access storage, but it has some limitations: it only works on Linux, and it imposes some performance penalty.

You can also attach to Chirp filesystems by using the FUSE package to attach Chirp as a kernel filesystem module. Unlike Parrot, this requires superuser privileges to install the FUSE package, but will likely work more reliably on a larger number of programs. You can do this with either Linux FUSE or MacFuse. Once you have downloaded and installed FUSE, simply run chirp_fuse with the name of a directory on which the filesystem should be mounted. For example:

$ mkdir /tmp/chirp
$ chirp_fuse /tmp/chirp
$ cd /tmp/chirp
$ ls -la
total 9742
dr-xr-xr-x 0 fred users 6697  Feb 22 13:54 albus.cse.nd.edu:9094
dr-xr-xr-x 0 fred users 6780  Feb 22 13:54 aluminum.helios.nd.edu:9094
dr-xr-xr-x 0 fred users 27956 Feb 22 13:54 angband.cse.nd.edu:9094
dr-xr-xr-x 0 fred users 6466  Feb 22 13:54 angelico.helios.nd.edu:9094

For more portable, explicit control of a Chirp server, use the Chirp command line tool. This allows you to connect to a server, copy files, and manage directories, much like an FTP client:

$ chirp
chirp::> open myhost.somewhere.edu
chirp:myhost.somewhere.edu:/> put /tmp/bigfile file
/tmp/bigfile -> /bigfile (11.01 MB/s)
chirp:myhost.somewhere.edu:/> ls -la
dir  4096      .       Fri Sep 10 12:40:27 2018
dir  4096      ..      Fri Sep 10 12:40:27 2018
file 147       hosts   Fri Sep 10 12:39:54 2018
file 104857600 bigfile Fri Sep 10 12:53:21 2018
chirp:myhost.somewhere.edu:/>

In scripts, you may find it easier to use the standalone commands chirp_get and chirp_put, which move single files to and from a Chirp server. These commands also allow for streaming data, which can be helpful in a shell pipeline. Also, the -f option to both commands allows you to follow a file, much like the Unix tail command:

$ tar cvzf archive.tar.gz ~/mydata
$ chirp_put archive.tar.gz myhost.somewhere.edu archive.tar.gz

$ chirp_get myhost.somewhere.edu archive.tar.gz - | tar xv

$ chirp_get -f myhost.somewhere.edu logfile - |& less

You can also write programs that access the Chirp C interface directly. This interface is relatively self explanatory: programs written to use this library may perform explicit I/O operations in a manner very similar to Unix. For more information, see the HOWTO: Write Code that Uses Chirp

Finding Chirp Servers

Now that you know how to run and use Chirp servers, you will need a way to keep track of all of the servers that are available for use. For this purpose, consult the Chirp storage catalog. This web page is a list of all known Chirp servers and their locations. Note that this same list appears if you use Parrot to perform an ls on /chirp

The storage catalog is highly dynamic. By default, each Chirp server makes itself known to the storage catalog every five minutes. The catalog server records and reports all Chirp servers that it knows about, but will discard servers that have not reported for fifteen minutes.

If you do not want your servers to report to a catalog, then run them setting the option -u to -:

$ chirp_server -u -

Alternatively, you may establish your own catalog server. See Catalog Servers for details.

Security

Different sites require different levels of security and different technological methods of enforcing security. For these reasons, Chirp has a very flexible security system that allows for a range of tools and policies from simple address checks to Kerberos authentiation.

Security really has two aspects: authentication and authorization. Authentication deals with the question Who are you? Once your identity has been established, then authorization deals with the question What are you allowed to do?:

Authentication

Chirp supports the following authentication schemes:

The Chirp tools will attempt all of the authentication types that are known and available in the order above until one works. For example, if you have Kerberos installed in your system, Chirp will try that first. If not, Chirp attempts the others.

Once an authentication scheme has succeeded, Chirp assigns the incoming user a subject that describes both the authentication method and the user name within that method. For example, a user that authenticates via Kerberos might have the subject:

kerberos:dthain@nd.edu

A user authenticating with Globus credentials might be:
(Note that Chirp substitutes underscores for spaces.)

globus:/O=Cooperative_Computing_Lab/CN=Douglas_L_Thain

While another user authenticating by local unix ids might be:

unix:dthain

While a user authenticating by simple hostnames might be:

hostname:pigwidgeon.cse.nd.edu

Take note that Chirp considers all of the subjects as different identities, although some of them might correspond to the same person in varying circumstances.

Authorization

Once Chirp has authenticated your identity, you are logged into a server. However, when you attempt to read or manipulate files on a server, Chirp checks to see whether you are authorized to do so. This is determined by access control lists or ACLs.

Every directory in a Chirp server has an ACL, much like filesystems such as as AFS or NTFS. To see the ACL for a directory, use the Chirp tool and the getacl command:

chirp:host.mydomain.edu:/> getacl
unix:dthain rwlda hostname:*.mydomain.edu rwl

Or, if you are using Parrot, you can use parrot_getacl to examine ACLs in the same way:

$ parrot_run bash
$ cd /chirp/host.mydomain.edu
$ parrot_getacl
unix:dthain rwlda hostname:*.mydomain.edu rwl

This ACL indicates that the subject unix:dthain has five access rights, while the subject pattern hostname:*.mydomain.edu has only three access rights. The access rights are as follows:

Access rights often come in combinations, so there are a few aliases for your convenience:

To change an access control list on a directory, use the setacl command in the Chirp command line tool:

chirp:host.mydomain.edu:/> setacl / kerberos:dthain@nd.edu write
chirp:host.mydomain.edu:/> getacl
unix:dthain rwlda hostname:*.mydomain.edu rwl kerberos:dthain@nd.edu rwld

Note that for subject names that contain spaces, you should simply substitute underscores. For example, if your subject name is /O=Univ of Somewhere/CN=Fred Flint, then you might issue a setacl command like this:

chirp:host.mydomain.edu:/> setacl / /O=Univ_of_Somewhere/CN=Fred_Flint rwlda

Or, you can accomplish the same thing using parrot_setacl inside of Parrot:

$ parrot_run bash
$ cd /chirp/host.mydomain.edu $ parrot_setacl . /O=Univ_of_Somewhere/CN=Fred_Flint rwlda

The meaning of ACLs is fairly obvious, but there are few subtleties you should know:

• Rights are generally inherited. When a new directory is created, it automatically gets the ACL of its parent. Exception: read about the reserve right below.

• Rights are generally not hierarchical. In order to access a directory, you only need the appropriate permissions on that directory. For example, if you have permission to write to /data/x/y/z, you do not need any other permissions on /data, /data/x and so forth. Of course, it may be difficult to discover a deep directory without rights on the parents, but you can still access it.

• The delete right is absolute. If you have permission to delete a directory, then you are able to delete the entire subtree that it contains, regardless of any other ACLs underneath.

Reservation

The v - reserve right is an important concept that deserves its own discussion.

A shared-storage environment such as Chirp aims to allow many people to read and write common storage space. Of course, with many people reading and writing, we need some mechanism to make sure that everybody does not step on each other's toes.

The reserve right allows a user to create what is essentially a fresh workspace for their own use. When a user creates a new directory and has the v right (but not the w right), Chirp will create a new directory with a fresh ACL that gives the creating user restricted rights.

A good way to use the reserve right is with a wildcard at the top directory. Here's an example. Suppose that Fred creates a new Chirp server on the host bigwig. Initially, no-one except Fred can access the server. The first time it starts, the Chirp server initializes its root directory with the following ACL:

unix:fred rwla

Now, Fred wants other users in his organization to be able to use this storage, but doesn't want them messing up his existing data. So, Fred uses the Chirp tool to give the list ( l ) and reserve ( v ) rights to anyone calling from any machine in his organization:

chirp:bigwig:/> setacl / hostname:*.somewhere.edu lv(rwlda)
chirp:bigwig:/> getacl /
unix:fred rwlda hostname:*.somewhere.edu lv(rwlda)

Now, any user calling from anywhere in somewhere.edu can access this server. But, all that any user can do is issue ls or mkdir in the root directory. For example, suppose that Betty logs into this server from ws1.somewhere.edu. She can not modify the root directory, but she can create her own directory:

chirp:bigwig:/> mkdir /mydata

And, in the new directory, ws1.somewhere.edu can do anything, including edit the access control. Here is the new ACL for /mydata:

chirp:bigwig:/> getacl /mydata
hostname:ws1.somewhere.edu rwlda

If Betty wants to authenticate with Globus credentials from here on, she can change the access control as follows:

chirp:bigwig:/> setacl /mydata globus:/O=Univ_of_Somewhere/CN=Betty rwla

And, the new acl will look as follows:

chirp:bigwig:/> getacl /mydata
hostname:ws1.somewhere.edu rwlda globus:/O=Univ_of_Somewhere/CN=Betty rwla

Simple Group Management

Chirp currently supports a simple group management system based on files. Create a directory on your local filesystem in which to store the groups. Each file in the directory will have the name of the desired groups, and contain a list of the members of the group in plain text format. Then, give your Chirp server the -G argument to indicate the URL of the group directory. Once the groups are defined, you can refer to them in access control lists using the group: prefix.

For example, suppose you wish to have two groups named group:students and group:faculty. You could define the groups in the /data/groups directory as follows:

/data/groups/students: unix:astudent unix:bstudent
/data/groups/faculty: unix:aprof unix:bprof

Then, run the chirp server with the option -G file:///data/groups. (Notice the URL syntax.) Then, to make a directory /homework that is readable by students and writable by faculty, do this:

chirp:bigwig:/> mkdir /homework
chirp:bigwig:/> setacl /homework group:students rl
chirp:bigwig:/> setacl /homework group:faculty rwld

If the groups are to be shared among many Chirp servers, place the group directory on a web server and refer to it via an http URL.

Notes on Authentication

Each of the authentication types has a few things you should know:

• Kerberos: The server will attempt to use the Kerberos identity of the host it is run on. (i.e. host/coral.cs.wisc.edu@CS.WISC.EDU) Thus, it must be run as the superuser in order to access its certificates. Once authentication is complete, there is no need for the server to keep its root access, so it will change to any unprivileged user that you like. Use the -i option to select the userid.

• Globus: The server and client will attempt to perform client authentication using the Grid Security Infrastructure (GSI)> Both sides will load either user or host credentials, depending on what is available. If the server is running as an ordinary user, then you must give a it a proxy certificate with grid-proxy-init. Or, the server can be run as root and will use host certificates in the usual place.

• Unix: This method makes use of a challenge-response in the local Unix filesystem to determine the client's Unix identity. It assumes that both machines share the same conception of the user database and have a common directory which they can read and write. By default, the server will pick a filename in /tmp, and challenge the client to create that file. If it can, then the server will examine the owner of the file to determine the client's username. Naturally, /tmp will only be available to clients on the same machine. However, if a shared filesystem directory is available, give that to the chirp server via the -c option. Then, any authorized client of the filesystem can authenticate to the server. For example, at Notre Dame, we use -c /afs/group/ccl/software/rendezvous to authenticate via our AFS distributed file system.

• Hostname: The server will rely on a reverse DNS lookup to establish the fully-qualified hostname of the calling client. The second field gives the hostname to be accepted. It may contain an asterisk as a wildcard. The third field is ignored. The fourth field is then used to select an appropriate local username.

• Address: Like "hostname" authentication, except the server simply looks at the client's IP address.

By default, Chirp and/or Parrot will attempt every authentication type knows until one succeeds. If you wish to restrict or re-order the authentication types used, give one or more -a options to the client, naming the authentication types to be used, in order. For example, to attempt only hostname and kerberos authentication, in that order:

$ chirp -a hostname -a kerberos

Advanced Topics

Cluster Management

Several tools are available for managing a large cluster of Chirp servers.

First, a Java visual display applet gives a graphical view of all servers in a cluster, as well as active network connections between each client and server. This tool can be used to quickly view whether storage is free or used, whether CPUs are idle or busy, and whether the network is idle or in use. Clicking on individual nodes shows the same detailed data as is avaliable in the catalog page.

Next, it can be helpful to give a single 'superuser' limited access to all servers and directories in a cluster, allowing them to fix broken access controls and solve other problems. To allow this, the -P user argument can be given to a Chirp server, and will implicitly give the named user the L and A rights on any directory on that server.

When managing a large system with many users, it is important to keep track of what users are employing the cluster, and how much space they have consumed. We refer to this as auditing the cluster. To audit a single node, use the audit command of the Chirp tool. This produces a listing of all users of a single host. (You must have the A right in the root directory of the server to run this command.) For example:

$ chirp ccl01.cse.nd.edu audit
FILES DIRS      DATA    OWNER
82842   27    5.0 GB    globus:/O=UnivNowhere/CN=Fred
 6153  607  503.4 MB    unix:fred 2 2 200.3 MB hostname:laptop.nowhere.edu 12 2 3.9 MB unix:betty

To audit an entire cluster, run the chirp_audit_cluster tool. This will extract the current list of hosts from your catalog, run an audit on all hosts in parallel, and then produce several reports in text files: audit.users.txt, audit.hosts.txt, audit.users.hosts.txt, and audit.hosts.users.txt.

Often, users of a cluster will wish to replicate commonly used data across all disks in the system, perhaps to provide fast access to relatively static data. The chirp_distribute tool can be used to rapidly move data from one node to all others. Given a source host and path, chirp_distribute will create a spanning tree and then move data directly from host to host in parallel. This is much faster than running cp or chirp put directly. For example, this will copy the /database directory from host server.nd.edu to all hosts in your cluster:

# First we get a list of all the chirp hosts in the cluster:
ALL_CHIRP_HOSTS=$(chirp_status -s)

# Then we use chirp_distribute to copy /database to all the hosts found:
$ chirp_distribute server.nd.edu /database $ALL_CHIRP_HOSTS

Another common pattern is cleaning up data that has been copied this way. To delete, simply run chirp_distribute again with the -X option and the same arguments, that is:

$ chirp_distribute -X server.nd.edu /database $ALL_CHIRP_HOSTS

Space Management

When multiple users share a common storage space, there is the danger that one aggressive user can accidentally (or deliberately) consume all available storage and prevent other work from happening. Chirp has two mechanisms available to deal with this problem.

The simpler tool is just a free space limit. If run with the -F option, a Chirp server will stop consuming space when the free space on the disk falls below this limit. External users will see a "No space left on device." error. For example, -F 100MB will leave a minimum of 100MB free on the local disk. This mechanism imposes little or no performance penalty on the server.

The more complex tool is a user-level quota and allocation system. If run with the -Q option, a Chirp server will establish a software quota for all external users. That is, -Q 2GB will limit external users to consuming a total of 2 GB of storage within a single Chirp server. This mechanism imposes some run-time performance penalty, and also delays server startup somewhere: the Chirp server must traverse its storage directory to count up the available space.

With the -Q option enabled, external users can allocate space before consuming it. Using the Chirp tools, users may use the mkalloc command to create new directories with an attached space allocation. For example, mkalloc /mydata 1GB will create a new directory /mydata with an allocation of 1GB. This allocation is a limit that prevents files in that directory from consuming more than 1GB; it is also a guarantee that other users of the server will not be able to steal the space. Such allocations may also be subdivided by using mkalloc to create sub-directories.

To examine an allocation, use the lsalloc command.

To destroy an allocation, simply delete the corresponding directory.

Ticket Authentication

Often a user will want to access a Chirp server storing files for cluster computing jobs but will have difficulty accessing it securely without transferring their credentials with the jobs dispatched to the cluster. To facilitate ease-of-use, users typically solve this by giving rights to a hostname mask (e.g. *.cse.nd.edu ) on the Chirp server. However, this level of access can be innappropriate due to sensitive data. Instead, these users are forced to use difficult authentication methods such as Globus or Kerberos for running the Chirp server. They may also use a virtual network solution but users typically lack this amount of control on clusters. To provide an easy solution to this problem, Chirp offers its own ticket based authentication system which is convenient and simple to setup.

To start, users may create a ticket for authentication using:

$ chirp <host:port> ticket_create -output myticket.ticket -subject unix:USER -bits 1024 -duration 86400 / rl /foo rwl

This command performs multiple tasks in three stages:

First, it creates a ticket which is composed of an RSA Private Key with a key (modulus) size of 1024 bits. When we refer to the ticket, we are speaking of this Private Key. By default, the ticket file generated is named ticket.MD5SUM where MD5SUM is the MD5 digest of the Public Key of the ticket.

Once the ticket is created, it is registered with the Chirp server with a validity period in seconds defined by the duration option (86400, or a day). The -subject unix:USER switch allows the user to set the ticket for another user with unix id USER; however, only the chirp_server superuser (-P) may set tickets for any subject. For regular users, the -subject option is unnecessary as it is by default the subject you possess when registering the ticket. Users who authenticate using this ticket in the future will become this subject with certain masked rights.

Once the ticket is created and registered, we give the ticket a set of ACL masks. The ACL mask will mask the rights of the ticket-authenticated user with the rights of the subject that registered the ticket. For example, if a user named foo (subject is unix:foo) has rights rwl in the root directory of the Chirp server and if a ticket is registered for foo with the ACL mask / rlx, the effective rights of the ticket-authenticated user is rl in the root directory.

ACL masks are also inherited from parent directories. So, in the above example, the root directory has the ACL mask rl while the foo directory has the ACL mask rwl. Other nested directories within the root directory also inherit the rl mask. Similarly, nested directories of the foo directory inherit the rwl mask. We emphasize that the ACL mask does not give rights but limits them. If the user that registers a ticket has no rights in a directory, then neither will the ticket authenticated user.

Authenticating with a ticket

To authenticate using a ticket, it can be as simple as including the ticket file with your job. Tickets that follow the ticket.MD5SUM template are automatically added to the list of tickets to try when authenticating. You can also give specific tickets to authenticate with using a comma-delimited list of ticket filenames in either the CHIRP_CLIENT_TICKETS environment variable or via the -i option. Tickets are tried in the order they are specified.

$ chirp <host:port>

The above command will try ticket authentication as a last resort but will use tickets it finds in the current directory following the template.

$ chirp -a ticket -i file.ticket <host:port>

The above command forces ticket authentication and only uses the file.ticket ticket to authenticate.

Authenticating is this simple. It is important to note that tickets are obviously not protected in any way from theft when you distribute the ticket with jobs in a distributed computing environment (no ticket system can give this guarantee). Users may want to protect their tickets in basic ways by setting a restrictive file mode and by giving tickets a limited duration on the server.

Finally, users should be careful to experiment with small key sizes for a balance of quick authentication and security. Smaller key sizes may be rejected outright by openssl when given a 64 byte challenge to sign. Chirp will not authenticate or use smaller challenge sizes if openssl rejects the ticket.

Manually Registering a Ticket

A ticket is only useful when registered with a server. The ticket_create command does this for you automatically but you may also wish to register the ticket with multiple servers. To do this, you can manually register a ticket that is already created by using the ticket_register command:

$ chirp <host:port> ticket_register myticket.ticket unix:user 86400

The first argument to ticket_register is the name of the ticket, followed by the subject, and finally the ticket duration. The second option (the subject) is optional. As described earlier, specifying the subject allows you to register a ticket with a user other than yourself. This is only possible if you are authenticated with the server as the super user.

Modifying the Rights of a Ticket

You may use the ticket_modify command to change the rights a ticket has in a directory. You are restricted to giving rights to a ticket you already possess. Recall, however, that the rights are actually a mask that are logically ANDed with the rights the user has at the time.

$ chirp <host:port> ticket_modify myticket.ticket / rl

The above command changes the ACL mask of myticket.ticket to rl in the root directory.

A ticket identifier as returned by ticket_list may also be used instead of a ticket filename.

Deleting a Ticket

Deleting a ticket unregisters the ticket with the server. Additionally, the ticket on the client is deleted.

$ chirp <host:port> ticket_delete myticket.ticket

A ticket identifier as returned by ticket_list may also be used instead of a ticket filename.

Listing the Registered Tickets on the Server

To list the tickets registered on a server, use the ticket_list command:

$ chirp <host:port> ticket_list unix:user

The subject argument instructs the command to fetch all the tickets belonging to the user. You may also use ticket_list all to list all the tickets of all users on the server. The latter command is only executable by the Chirp super user. The output is a list of tickets identifiers. You can query information about a ticket using these identifiers with the ticket_get command.

Getting the Information of a Registered Ticket from the Server

To check the status of a ticket on a server, you may use the ticket_get command:

$ chirp <host:port> ticket_get myticket.ticket

So long as you own the ticket or are authenticated as the super user, the server will return to you information associated with the ticket. The ticket must also exist and must also not have expired. ticket_get takes a client side ticket filename as an argument or a ticket identifier as returned by the ticket_list command.

ticket_get prints the subject that owns the ticket, the base64 encoded public key of the ticket, the time left until the ticket expires in seconds, and a variable number of directory and ACL masks. For example, we might have the following output:

$ chirp host:port ticket_get myticket.ticket
unix:pdonnel3
LS0tLS1CRUdJTiBQVUJMSUMgS0VZLS0tLS0KTUlHZk1BMEdDU3FHU0liM0RRRUJBUVVBQTRHTkFEQ0Jp
UUtCZ1FEZVoyZWxKYXdlcHBHK0J4SFlaMmlmWFIzNAovU3RhUElta0lmeit4TDZxN21wS2lnMDJQZ2Z5
emdKRWFjMk50NzJrUlBpOEJWYWdkOHdvSGhWc25YZ1YvNjFPCjVkaG13STNLYWRlYjNUbkZXUUo3bFhh
anhmVTZZR1hXb2VNY1BsdjVQUWloWm8yWmFXTUUvQVA4WUtnVVphdXcKelI2RkdZWGd6N2RGZzR6Yk9R
SURBUUFCCi0tLS0tRU5EIFBVQkxJQyBLRVktLS0tLQo=
5993
/ rl /foo rwl

Note that the base64 encoded public key above is wrapped to fit an 80 character width for this manual. In the actual output, the public key is on one line. All of the information is new-line-delimited.

HDFS Backend Storage for Chirp

The Chirp server is able to bind to backend filesystems besides the local filesystem. In particular, it is able to act as a frontend for the Hadoop HDFS filesystem. When used on top of HDFS, Chirp gives you the benefit of a robust system of ACLs, simple userspace access and POSIX semantics (with some limitations, discussed below). Perhaps best of all, client jobs will no longer have any Hadoop or Java (version) dependencies.

To run a Chirp server as a frontend to your HDFS filesystem, you will need to install the libhdfs-devel package and then set several environment variables which describe your Hadoop installation. JAVA_HOME and HADOOP_HOME LIBHDFS_PATH should be set to indicate the location of the libhdfs library. Common values for the Cloudera Hadoop installation would be this:

setenv JAVA_HOME /usr/lib/jvm/java-openjdk setenv HADOOP_HOME /usr/lib/hadoop setenv LIBHDFS_PATH /usr/lib64/libhdfs.so Then, start chirp_server and indicate the root storage directory in HDFS with -r like this:

$ chirp_server -r hdfs://headnode.hadoop.domain.edu/mydata ...other arguments...

By default, chirp will use whatever default replication factor is defined by HDFS (typically 3). To change the replication factor of a single file, use the chirp setrep or parrot_setrep commands. A path of &&& will set the replication factor for all new files created in that session.

Temporary Local Storage

Chirp allows you to setup a location to place temporary files such as those for caching groups, and other items. You can set this using the -y path. This allows for faster access, POSIX semantics, and less load on HDFS. By default, Chirp assumes the current directory for temporary storage.

Limitations

Chirp tries to preserve POSIX filesystem semantics where possible despite HDFS violating certain assumptions. For example, random writes are not possible for Chirp on HDFS. When the user requests to open a file for writing, Chirp assumes an implicit O_APPEND flag was added. In addition, HDFS does not maintain an execute permission bit for regular files. Chirp assumes all files have the execute bit set.

Chirp also does not allow using the thirdput command or user space management (-F) when using HDFS as a backend.

Job Execution on Chirp

As of version 4.2.0, Chirp supports job execution. Jobs run using executables and files located on the Chirp server. Each job description sent to the Chirp server provides a binding of each file the job requires to a local namespace (a sandbox).

To support the new job interface, Chirp has the following new RPCs:

                <integer result> = **job_create** <JSON-encoded job description>
                <integer result> = **job_commit** <JSON-encoded array of job IDs>
<JSON-encoded array of statuses> = **job_status** <JSON-encoded array of job IDs>
<JSON-encoded array of statuses> = **job_wait** <job ID> <timeout>
                <integer result> = **job_reap** <JSON-encoded array of job IDs>
                <integer result> = **job_kill** <JSON-encoded array of job IDs>

As usual, these RPCs may be sent through the Chirp client command line tool or through the C API.

Creating a Job

To create a job, you need the usual attributes of an executable to run, the arguments to pass to the executable, any environment variables to add, and any files to bind into the job's namespace.

In Chirp's Job execution framework, files are bound into the job's namespace. The name of the file in the task's name space is labeled task_path while the name in the server namespace is labeled serv_path. Files are bound in the task namespace at job start and job end, for inputs and outputs, respectively.

Example 1

my-first-job.json

my-first-job.json:
{
    "executable": "/bin/sh",
    "arguments": [ "sh", "-c", "echo Hello, world! > my.output" ],
    "files": [
                {
                    "task_path": "my.output",
                    "serv_path": "/some/directory/my.0.output",
                    "type": "OUTPUT"
                }
            ]
}

This job can be created from the command line as follows:

# We make sure the appropiate directories exist:
$ chirp <host:port> mkdir -p /some/directory

# Create the job. It prints the job-id when the job is succesfully created:
# <integer result> = **job_create** <JSON-encoded job description>

$ chirp <host:port> job_create "$(cat my-first-job.json)"
1

Example 2 -- Two Inputs

job-with-two-inputs.json

{
    "executable": "/bin/tar",
    "arguments": [ "tar", "-cf", "archive.tar", "a", "b" ],
    "files": [
                {
                    "task_path": "a",
                    "serv_path": "/users/btovar/a.txt",
                    "type": "INPUT",
                    "binding": "LINK"
                },
                {
                    "task_path": "b",
                    "serv_path": "/users/btovar/b.txt",
                    "type": "INPUT",
                    "binding": "LINK"
                },
                { "task_path": "archive.tar",
                  "serv_path": "/users/btovar/archive.tar",
                  "type": "OUTPUT",
                  "binding": "LINK"
                }
            ]
}

Here, each file is bound using hard links to the file located on the server. This type of access is fast as the server does not need make a copy. You may also bind files as COPY if necessary. LINK is the default.

# Create the job:
$ chirp <host:port> job_create "$(cat job-with-two-inputs.json)"
2

Example 3 -- Using custom executable

Often, you will have a script or executable which is present on the Chirp server which you want to execute directly. To do this, bind the executable as you would any other file and give a relative (task) path for the executable job attribute:

In this example my-custom-exec.sh, takes as first argument the name of a file to print its output.

job-with-custom-exec.json

{
    "executable": "./my-custom-exec.sh",
    "arguments": [ "my-custom-exec.sh", "output.txt" ],
    "files": [
                {
                    "task_path": "my-custom-exec.sh",
                    "serv_path": "/some/directory/my-custom-exec.sh",
                    "type": "INPUT",
                    "binding": "LINK" },
                {
                    "task_path": "output.txt",
                    "serv_path": "/some/directory/output.txt",
                    "type": "OUTPUT",
                    "binding": "LINK"
                }
            ]
}

# Make sure that myscript.sh is in the correct location:
$ chirp <host:port> put my-custom-exec.sh /some/directory/my-custom-exec.sh

$ chirp <host:port> job_create "$(cat job-with-custom-exec.sh)"
4

Committing (to Start) a Job

Chirp uses two-phase commit for creating a job. This serves to protect against orphan jobs which become lost because a client or the server lose a connection.

To commit a job, pass a JSON-encoded array of job identifiers to the job_commit RPC. For example:

# <integer result> = **job_commit** <JSON-encoded array of job IDs>

$ chirp host:port job_commit '[1, 2, 4]'

will commit jobs 1, 2, and 4.

Once a job is committed, the Chirp server is free to schedule and execute the job. You may query the status of the job to see if it has begun executing or wait for the job to finish.

Querying the Status of a Job

At any point in a job's lifetime, you may query its status. Status information is JSON-encoded and holds all job metadata.

Example 1 -- Status of Example 1

# <JSON-encoded array of statuses> = **job_status** <JSON-encoded array of job IDs>

$ chirp host:port job_status '[1]'
[
  {
    "id": 1,
    "error": null,
    "executable": "/bin/sh",
    "exit_code": 0,
    "exit_status": "EXITED",
    "exit_signal": null,
    "priority": 1,
    "status": "FINISHED",
    "subject": "unix:btovar",
    "tag": "(unknown)",
    "time_commit": 1565793785,
    "time_create": 1565793785,
    "time_error": null,
    "time_finish": 1565793785,
    "time_kill": null,
    "time_start": 1565793785,
    "time_reap": null,
    "arguments": [
      "sh",
      "-c",
      "echo Hello, world! > my.output"
    ],
    "environment": {},
    "files": [
      {
        "binding": "LINK",
        "serv_path": "/some/directory/my.0.output",
        "size": 14,
        "tag": null,
        "task_path": "my.output",
        "type": "OUTPUT"
      }
    ]
  }
]

You can get the status of a job at any time, that is, before commit, during execution, and on completion. However, this RPC does not help with waiting for one or more jobs to finish. For that, we use the job_wait RPC discussed next.

Waiting for a Job to Terminate

Use job_wait to wait for a job to finish. This will give you the status information of jobs which have completed and have a status of FINISHED, KILLED, or ERRORED.

job_wait takes a job identifier argument which matches jobs in the following way:

job_wait is essentially job_status except the RPC blocks until a job matches the above condition or the timeout is exceeded:

# <JSON-encoded array of statuses> = **job_wait** <job ID> <timeout>


# wait 10 seconds for any job to finish:
$ chirp host:port job_wait 0 10

# wait indefinitely for job_wait 1 to finish:
$ chirp host:port job_wait 1
[{"id":1,"error":null,"executable":"\/bin\/sh","exit_code":0,"exit_status":"EXITED",...]

# wait 10 seconds for any job with id greater than 500 to finish:
$ chirp host:port job_wait -500 10
[]

# Note the empty array above, which indicates that no such job finished in the
# given timeout.

Reaping a Finished Job

Similar in intent to job_commit, job_reap notifies the Chirp server that your application has logged the termination of the job. This allows the Chirp server to reap the job. The side-effect of this operation is future calls to job_wait will not include the reaped jobs.

# <integer result> = **job_reap** <JSON-encoded array of job IDs>

# Reap jobs 1, 2, and 4:
$ chirp host:port job_reap '[1, 2, 4]'

Killing a Job

job_kill informs the Chirp server to kill a job. Any job which has not reached a terminal state (FINISHED, KILLED, or ERRORED) will immediately be moved to the KILLED state. If the job is running, the internal Chirp scheduler will also terminate the job at its convenience.

#<integer result> = **job_kill** <JSON-encoded array of job IDs>`

# Kill jobs 1 and 2
$ chirp host:port job_kill '[1, 2]'

Chirp Jobs on AFS

On the AFS file system, Chirp job execution will not work with LINK file bindings. This is due to limitations in AFS preventing hard links across directories. For this reason we recommend against using AFS as the backing storage for Chirp (--root). If you must use AFS, the COPY binding should work.

Debugging Advice

Debugging a distributed system can be quite difficult because of the sheer number of hosts involved and the mass of information to be collected. If you are having difficulty with Chirp, we recommend that you make good use of the debugging traces built into the tools.

In all of the Chirp and Parrot tools, the -d option allows you to turn on selected debugging messages. The simplest option is -d all which will show every event that occurs in the system.

To best debug a problem, we recommend that you turn on the debugging options on both the client and server that you are operating. For example, if you are having trouble getting Parrot to connect to a Chirp server, then run both as follows:

$ chirp_server -d all [more options] ...
$ parrot_run -d all bash

Of course, this is likely to show way more information than you will be able to process. Instead, turn on a debugging flags selectively. For example, if you are having a problem with authentication, just show those messages with -d auth on both sides.

When debugging problems with Chirp and Parrot, we recommend selectively using -d chirp, -d tcp, -d auth, and -d libcall as needed.

Further Information

The Chirp Protocol

The Chirp Protocol

Confuga

Confuga is an active storage cluster file system harnessing Chirp. To learn more about it, please see the Confuga manual.

Citation

Please use the following citation for Chirp in a scientific publication:

• Douglas Thain, Christopher Moretti, and Jeffrey Hemmes,Chirp: A Practical Global Filesystem for Cluster and Grid Computing, Journal of Grid Computing , Springer, 2008. DOI: 10.1007/s10723-008-9100-5. (The original is available on http://www.springerlink.com.)

Further Information

For more information, please see Getting Help or visit the Cooperative Computing Lab website.

Copyright

CCTools is Copyright (C) 2022 The University of Notre Dame. This software is distributed under the GNU General Public License Version 2. See the file COPYING for details.