Index of values

Throws changes away, and closes the db.

Aborts a transaction, and abandons the changes of the transaction

Closes the descriptors to remote services so far possible, but does not permanently shut down the client functionality.

Same as async function

Called when we can run transactions to enable the housekeeping timers.

let s2 = add x s1: Adds the element to the underlying store, and returns a new sequence where the new element is visible at the end.

Add the datastore record as-is (for initializing this module from db).

Add the input files, and create map tasks.

Add these files as output of this map/emap task identified by this ID

Adds the Dn_io service directly to the controller.

adds several elements

Just return whether the key exists

same, using an ro transaction

Gets the contents of the admin table key

same, using an ro transaction

Overwrite this admin table

Allocates a wait_entry

allocate_balanced_strategy inode bmaps td owner repl index number bs pref: Allocates blocks for the index range index to index+number-1.

allocate_trivial_strategy bmaps td owner repl index number: Allocates blocks for the index range index to index+number-1.

Retrieves the complete allocinfo for this inode

same, using an ro transaction

let s3 = append s1 s2: Adds all elements visible via s2 to the end of the store of s1, and returns a new seq s3 where the new elements are visible.

Create a new rfun by applying the first argument only.

This "format" recognizes the format for input files by the file suffix (".var", "fixed<n>", or nothing).

Average number of runnable tasks (excl.

Average number of running tasks

Average number of running or runnable tasks

Starts the transaction: begin_transaction_e c clustername exp_rev.

The effective size of bigblocks.

bigblock_size fs path suggested: Returns the real bigblock size for the suggested value.

blit_memory src srcoff dst dstoff len copies len characters from buffer src, starting at character number srcoff, to string dst, starting at character number dstoff

Reads the blockalloc table for this datastore ID.

same, using an ro transaction

Updates the blockalloc table

Create a blockmap object, and initialize it from the blockalloc db table

Returns dn_blocksize

Returns the blocksize

block map row size - how many bits are in one row of the block map.

The name of the user with minimum rights: "pnobody" (Plasma nobody)

broadcast c restarts all processes waiting on the condition variable c.

Converts a normal blocklist into a set of block ranges

Returns the buffer.

Returns the block position of buffer

Returns the length of buffer in blocks

The number of covered blocks (incl.

check_up_to_date_e icm mc lt inode seqno: Checks whether seqno is the most recent committed sequence number of inode.

a clean buffer is continued to be used as clean buffer after a read access

Deletes all temp files (sync)

Deletes all files at this place the files function returns.

Create a client

Flushes unwritten data to disk (when writing to a store), and closes all OS resources

Closes the reader

Closes all file descriptors permanently

Closes the store

the cluster to which a transaction belongs

Returns the cluster name

Returns the namenodes passed to open_cluster

Commits changes, and closes the db.

Commits a transaction, and makes the changes of the transaction permanent.

Commits all prepared transactions at system startup

commit_e get_journal get_journal_mon post_commit shared_state debug_name:

Commits the transaction

Compare here the tasks in an execution-friendly order, i.e.

Declare that no more inputs will be added.

configure_auth c nn_user dn_user get_password: Configures that accesses to the namenode are authenticated on the RPC level as nn_user and accesses to datanodes are authenticated as dn_user.

Configures that the authentication daemon is contacted to get evidence about the current user.

Configures that this ticket proves the identity of the user.

configure_buffer c n: configures to use n buffers.

configure_default_user_group c user group: Normally, new files are created as the user and group corresponding to the current impersonation.

configure_plan jc conf

Configures that the data nodes with the given identities are preferred for the allocation of new blocks.

Similar to configure_pref_nodes, but this list overrides the global setting for this inode.

Configures the RPC server so the passed level of access protection is required.

Configures the RPC server so the passed level of access protection is required.

Configures a shared memory manager.

Performs a quite costly consistency check.

copy_in_e c inode pos fd len: Copies the data from the file descriptor fd to the file given by inode.

copy_in_from_buf c inode pos buf len: Copies the data from buf to the file denoted by inode.

copy_out_e c inode pos fd len Copies the data from the file referenced by inode to file descriptor fd.

copy_out_to_buf_e c inode pos buf len Copies the data from the file denoted by inode to the buffer buf.

Creates the line splitter thread

Creates a new empty place.

create fs name startblock endblock bufnum

Create a new authentication ticket (verifier) for this user ID.

Creates the shm object and returns its name

create_bl_cache mc lt inode: Create a cache for this transaction.

create_buffer_system n_bufs blocksize esys

Creates a new client.

create m: Creates a condition variable, and pushes it to the heap, using the mutator m.

The default values for the configurations are taken from mapred_env, and are effectively taken from the .conf files.

create_dir fs name: Creates a directory exclusively

create_file pl:

create_file fs name: Creates this file exclusively.

Creates a regular file (inode plus name) or a symlink.

Create a new inode.

create_manager esys: Creates a new manager.

The owner is an opaque ID which is guaranteed to be unique in the whole Netplex system

Creates a plan.

Creates a reserve_info record

Especially how many tasks are started per task server.

Like the above function, but gets many values from a mapred_config object

Creates a wait_set in the heap designated by the mutator.

let (user,group,supp_groups) = current_user c: Get the identity of the current client principal on the filesystem level.

Deletes the datastore with this ID and all rows referencing it

Find a datastore by identity

same, using an ro transaction

Gets a list of all datastores

same, using an ro transaction

Updates the datastore table.

Encode/decode a binary string using the "ALT-64" method.

delete db key: Deletes this key, if it exists

delete identity: Remove this store

Deletes this entry from the cache

Deletes this inode from the cache

Delete this file or empty dir

Delete the inode

Deletes this file or directory recursively

Delete all the objects on disk, and clear the registry

Destroys these objects

regular_ii c mode: Creates an inodeinfo record for a new directory, where the mode field is set to mode modulo the current mask

Looks the filename up relative to a directory (given as inode) and returns the inode number.

A clean or invalid buffer is set to dirty.

disable identity: Sets enabled to false

Disables the cache.

disable_ticket_verifier tv f_post: Schedules that tv will be disabled soon.

Disconnects from this datanode

Alternative main program for dispatching between several types of jobs.

Divides buffer_size and buffer_size_tight by this number

Returns the list of ranges covered by the blocklist (i.e.

Drops all buffers of this inode, including dirty ones.

Gets the contents of the cache

This module provides an internal cache of the datastore ID to datastore identity mapping.

A dummy condition is non-functional, but can be used to put something into condition-typed variables

A dummy wait_set is non-functional, but can be used to put something into wait_set-typed variables

debug report

Writes the buffer table to log (debug level)

Do the election.

an empty blocklist

A dummy ticket (used during blocklist construction)

enable identity: Sets enabled to true

Enables the cache.

Pushes a safetrans secret to all enabled datanodes.

Turn the time (as seconds since epoch) into the RPC version

Returns the event system

Performs the passed modification

Command-line interface for running this job.

Returns all tasks that are executable.

Returns an MD5 sum of the running executable

Executes this task

Expands the blockinfo structs, and returns only structs with length=1

Returns the sequence view on the end of the store.

extend bgreader endblock: Sets a new endblock (higher value than the old endblock)

Extracts a config section from a Netplex config file like:

Extracts a config section from a Netplex config file like:

Extracts a config section from a Netplex config file like:

Config section must look like:

Extracts the db_config from this Netplex config file:

Config section must look like (see Dn_config.dn_config for explanations):

let (jc, mjc) = extract_job_config cf args custom_params:

Config section must look like (see Nn_config.nn_node_config for explanations):

Config section must look like (see Nn_config.nn_node_config for explanations):

Factory for the processor "mapred_task_server"

Returns this as factory for inclusion into Netplex_main.startup

Returns this as factory for inclusion into Netplex_main.startup

Returns this as factory for inclusion into Netplex_main.startup

Get the length of the file in blocks

file_exists fs name: whether this file, directory or symbolic link exists

Returns the absolute file name

Returns the files of the place

Fills the blockmap from db (synchronous)

fill_memory mem pos len x: Fills the buffer mem from pos to pos+len-1 with character x.

Looks for the first LF char from mem[pos] to mem[endpos-1] and returns the position.

Disables further registrations (at the end of the initialization).

A record has exactly n bytes.

Flushes the underlying store

Flushes the internal buffers (for stores writing to files)

flush_e inode pos len: Flushes all buffered data of inode from pos to pos+len-1, or to the end of the file if len=0.

Flushes all buffers.

fold f a0 s: if the sequence contains the elements [x0; x1; ...; x(n-1)] this function computes

	  f (f (f a0 x0) x1 ... ) x(n-1)

Fold operator

Forces that the new version is fetched

forget_blocks blc block n: Forgets these blocks

Frees a wait_entry

Picks up an existing location, and makes the files there accessible as place

Return statistics

Return the graph as SVG (once plan complete)

The sorter orders the records first by the hash of the key.

Get the store record by identity string, or raise Not_found

Get the store record by identity string

Get the allocation size for the blocksize.

get_auth_ticket user: Get an authentication ticket for this user

get_blocklist_e t inode block n keep_flag Returns the list of blocks for blocks block to blocks+n-1.

get_blocks_e blc index number req_seqno: Gets the blocklist for inode for the index range index to index+number-1.

get_blocks_from_cache_e blc index: Tries to get the blocks for index from the cache for inode.

get_blocks_via_cache_e blc index: First tries to get the blocklist for index and inode from the cache (as in get_blocks_from_cache_e), and if this fails, requests the blocklist from the server (as in get_blocks_e).

Returns the block size

Get info about inode.

Requests an authentication ticket

get_config (): This is the standard way for clients to obtain a config object.

Return the coordinator in "host:port" syntax.

Return information about all enabled and live datanodes (includes identity, size, and host)

Return the values from the flush record (or Not_found)

Returns the ID

Returns the identity string

get_ii_e icm mc tid inode: Gets the inodeinfo for inode using transaction tid and the managed client mc.

get_ii_from_cache_e icm inode: Looks the inodeinfo up for inode.

get_ii_via_cache_e icm mc lt inode: First tries to get the inodeinfo for inode from the cache (as in get_ii_from_cache_e), and if this fails, requests the inodeinfo from the server (as in get_ii_e).

Get info about inode.

Get the values of these config settings

Returns the directory in the local filesystem where the files configured in task_files can be found.

Returns the directory in the local filesystem where log files can be placed.

Returns the record config.

get_rc me bigblock_size: Returns a record config for the given environment and the suggested size of the bigblock

Get the computed value (or an exception)

Return the role.

Returns the number of blocks

Get the current version of the user/group tables.

Extract uc from a queue of DB updates

Returns the designated new EOF value of pending writes.

Returns the designated new mtime value of pending writes.

Returns points for the task, and the points indicate how valuable the task is for a "greedy" job execution strategy.

returns the head element of the sequence

returns the first up to n elements as list

returns all configured task server hosts

impersonate c user group supp_groups ticket: Become the user on the filesystem level.

Initializations to be run before the server is started.

The esys is used for ro transactions

This can be called from component context to set the low-water and high-water marks.

init_dn_store config identity size: Initializes the store.

Deletes a row from inode

Whether there is a write error for this inode

Returns n inodes with set anonymous field

Loads the inode data if existing

same, using an ro transaction

Inserts a new row into inode

list all inode numbers

Returns the max ID

same, using an ro transaction

Sets the anonymous flag of the inode

Updates a row in inode.

Sets the mtime and/or ctime.

Delete the blocks for inode in the range from index blkidx to blkidx+len-1.

Gets the block information for inode in the range from index blkidx to blkidx+len-1.

same, using an ro transaction

Returns list of (blkidx,block,length) for this inode and this datastore dsid (for fsck)

Insert these blocks.

insert db key value: Inserts a new entry, or replaces an existing entry with the same key.

insert_channel db key size: Returns a channel into which the inserted value must be written.

insert_large db key size ch: Inserts a new entry, or replaces an existing entry with the same key.

int_serializer heap esys: Creates a serializer that works across processes.

int_serializer_heap pool n: initializes a shared structure for use with int_serializer.

invalidate_ii icm inode: Removes information about this inode from the cache

investigate_machines config auth_ticket job_id:

Prepare the job, but do not run yet any task.

is_alive identity: Returns whether the datanode having this store is good as seen by the monitor.

is_alive identity: Returns whether the datanode having this store is good as seen by the monitor.

file_exists fs name: whether this directory exists

tests whether the sequence is empty

Whether the place can store files

iter f s: iterates over the elements of s and calls f x for every element x

iter f blist: Calls f identity (min,max) bi for all elementary ranges that can be compressed to one blockinfo bi.

iterate db f: Calls the function f with all keys, in index order.

The streaming job.

Return the job ID

Kills the job (also async)

Returns dn_shm_queue_length

These two functions return the number of records.

Conversions to/from string

The line structured file format represents every record as a line in a text file.

Links a name with an inode

link_at trans dir_inode name inode: Adds the entry name into the directory dir_inode and connects the entry with inode.

Returns the number of links (also no locks)

Returns the list of all datastores

Returns the list of all datastores

Lists the contents of the directory, given by filename

Lists the contents of the directory, given by inode

local_filesystem root: Access the local filesystem where the directory root is the assumed root.

Return the identities of the data nodes running on this machine (for configure_pref_nodes)

The name used for forming the name of the lock files

lookup db key: Looks this key up, and returns it.

Returns a registered function

lookup dir_inode name: Looks up the name in the cache, or raises Not_found

Looks up the inode in the cache, or raises Not_found

Looks the absolute filename up and returns the inode number

lookup_buffer sys inode index: Looks the buffer up for inode and index, or raises Not_found if there is none yet, or the existing descriptor is in `Dropped state.

lookup_large db key buf f: If the key is found, the function f is called with the size of the value.

Same as asynchronous function

Main program for the task server and job controller, for this single job only

let s3 = map f s1 s2: Maps the elements of s1 and adds them to s2

map f blist: Returns a blocklist where each element is mapped via f

mapl m pl_in pl_out conf: Runs a map-only job.

let s3 = map f s1 s2: Maps the elements of s1 to lists and adds them to s2

mapl_sort_fold <args> pl_in pl_out conf int_codec: This is map/reduce.

mapl_sort_fold ... p_in p_out:

Returns the config as object

Marks this task as finished

Marks this task as started, together with the node where it is running, and the runtime task ID and the misallocation flag

Return the maximum key size

The max of two time structs

size of buffers in bytes

merge blist1 blist2: Merges both blocklists.

Logs a message

Creates a directory

monitor_e cm uc f: Tries to get a slot for doing a commit described by uc.

multi_read_e c inode stream: This version of read allows it to read multiple times from the same file.

multi_tree trees default_tree:

Stats of tasks

The number of buffer requests that can be currently not satisfied

namelock trans dir name: Acquires an existence lock on the member name of directory dir.

Count names of an inode

same, using an ro transaction

Delete this name

Look up inode by dir_inode and name

Look up the parent inode of an inode.

same, using an ro transaction

same, using an ro transaction

Insert this name to inode mapping

Lists the contents of the directory identified by dir_inode.

same, using an ro transaction

Look up names by inode.

same, using an ro transaction

Access this cluster with the stream_fs interface:

In transactional mode, this function checks whether a concurrent writer has committed a new version in the meantime.

Same as async function

Looks for the next LF char.

creates an in-memory sequence (not linked to a file)

Creates a new in-memory store.

A new in-memory place for notebooks

We got an EIO and have to update the blocklist.

This "sorter" just copies the input unmodified to the output

Returns the number of cores (or virtual cores) on the system.

Convert datastore to/from the wire representation

open_cluster name namenodes: Opens the cluster with these namenodes (given as (hostname,port) pairs).

Same, but takes a Plasma_client_config.client_config object which can in turn be obtained via Plasma_client_config.get_config.

Opens a cluster with a configuration taken from an existing cluster.

Opens the store

opendb c basename flags: Opens the database using the client c.

Almost the same as async function.

Maps the shm object

override_blocks blc block n blocklist: Puts these blocks into the cache, overriding whatever is there.

Used in mapred.mlp to announce a new source file

Used in mapred.mlp to register a function with an automatic name

Returns a list of client parameters

Parses the config file and returns all included configs

Parses the admin tables passed as argument (as pairs (table_name,table_contents)).

Tries to get the amount of physical RAM, in bytes.

The name of the named pipe

returns the place where this store resides

Respond whether everything is planned

Whether everything is done

The total capacity of all machines (for planning)

plasma_filesystem cc configure:

Needs to be called from the post_add_hook

Needs to be called from the post_add_hook

This function must be called from the post_add_hook of the component

Reads the datastore table from the database, and starts discovering datanodes.

post_start_hook cont conf cauth pool master_index transaction_revision

post_start_hook pool cauth

pool

pool

pool

pool

Prepares the commit.

Describe the principal behind the ticket as pair (pw_entry, main_group, groups).

Returns (user,group,supp_groups)

Debug printing to stdout

The name of the privileged users: "proot" (Plasma root)

Pushes the db modification to the slave which the client connects to

Returns a string with n random bytes

Returns a random int 0 ..

Returns a random int64 number

Returns the sequence view such that the first element of the sequence is the first element of the store, i.e.

Returns another store value reading the data from the same physical place.

read bgreader: reads the next blocks into the buffer.

read_e c inode pos s spos len: Reads data from inode, and returns (n,eof,ii) where n is the number of read bytes, and eof the indicator that EOF was reached.

read_admin_table_e key Returns the admin table key as text.

read_block st block blockpos mem mpos len: Reads the bytes blockpos to blockpos+len-1 from block into the range mpos to mpos+len-1 of mem.

Constructs a record reader that reads from the input readers one after the other.

Returns the stores at one place.

Read the admin tables for managing users and groups, and make the contents available as Plasma_ug.ug_admin object.

Needs to be called from the receive_message hook in order to arrange notifications about db updates

Needs to be called from the receive_message hook in order to subscribe to this news channel

register name f: Registers the function f under name.

Register a temporary file

Register a POSIX shm object

regular_ii c mode: Creates an inodeinfo record for a new empty regular file, where the mode field is set to mode modulo the current mask

Checks whether the buffers are in the right state after finishing or aborting a flush request.

remove (min,max) blist: Returns the blocklist where the interval (min,max) is deleted

Revokes mark_as_started or mark_as_finished

rename trans old_path new_path: Renames/moves the file or directory identified by old_path to the location identified by new_path.

rename_at trans old_dir_inode old_name new_dir_inode new_name: Moves the file old_name in old_dir_inode to the new location which is given by new_name in new_dir_inode.

Renew the ticket.

Re-opens the db after commit or abort

Same as async function

Which admin tables are required (normally "passwd" and "group", but in the future more such tables could be defined)

request_buffer_e sys inode index: This function is to be used when lookup_buffer raises Not_found to get a new buffer.

reads the elements from the store obtained by calling Store.read

Executes a ticket reset in the background

Resets the error

retry c name f arg: Executes f arg and returns the result.

Looks for the last LF char from mem[pos] to mem[endpos-1] and returns the position.

Returns the filenames linked with this inode number.

Returns the abs filename linked with this inode number which must be a directory.

Deletes all revisions

Get the checksum of all revision strings

Get the maximum revision string

same, using an ro transaction

This initial transaction clears all revisions.

Inserts or updates the revision string of a committer

Returns the esys for ro transactions

Same for read-only transactions.

Returns points for the task, such that tasks in later rounds get more points than tasks in earlier rounds.

Returns (rpc_user,rpc_password)

Returns a value suitable for the mclient_auth_methods field.

Returns a value suitable for the mclient_auth_methods field.

Run the tasks, and wait until the job is done

scan_dir fs dir deeply: Scans the directory dir for files, and returns the files (w/o directory prefix) as pairs (name,info).

Returns information about this file

If the buffer can be immediately invalidated this is done.

Same for a whole inode.

A taskfile manager that uses scp to copy file to remote nodes.

scp_to_cmd localfiles host dest: Generates a shell command that copies localfiles to host into the destination dest

select_for_flush sys inode: Returns a list of blocks that need to be written.

Returns the inodes that can be flushed, in the order of precedence.

serializer heap esys: Creates a serializer that works across processes.

serializer_heap pool: initializes a shared structure for use with serializer.

Make this process a blockmap server.

set info about inode.

Set the current version of the user/group tables.

Set up the RPC server so that Elect procedures are bound

Returns a limit on the amount of POSIX shm.

Returns the current manager

To be called at shutdown time

signal c restarts one of the processes waiting on the condition variable c.

signal_db_change id node_opt: Signal to the monitor that the db row with id has changed and must be reread.

let (mem,pos) = slot shm i: Accesses the slot i of the shm object shm.

snapshot trans inode: Takes a snapshot of the file, affecting buffered reads and writes, and a few other functions.

Loads the complete blocklist - which effectively means to take a snapshot of the whole file.

sort ~hash ~cmp s1 s2: Sorts the elements of s1 and appends the sorted result to s2, and returns this sequence.

ssh_cmd host cmd: Generates a shell command that runs cmd on host

standard_fs conf: Returns the standard filesystem that is used for map/red jobs.

Another version of standard_fs: If client_config is passed, PlasmaFS will be enabled, otherwise disabled.

Starts the job which is running asynchronously with the event system

Starts a transaction

Updates on the datastores are published and can be retrieved from any other process using Nn_datastore_news

Get statistics

Stops the thread.

Stops any ongoing election (in the current process only)

The name of a parameterless message for stop_election.

The underlying store of the sequence

str_cmp s1 p1 l1 s2 p2 l2: The same as String.compare (String.sub s1 p1 l1) (String.sub s2 p2 l2), only faster

See str_cmp

debug string

For debugging

sub (min,max) blist: Returns the blocklist with the subset of blist that intersects with the interval (min,max)

switch_to_reading sys b e eof_opt:

The buffer is switched to `Writing state (from `Dirty).

regular_ii c target: Creates an inodeinfo record for a symlink pointing to target

Syncs the store to disk

Waits until the event system is done and returns the result of the engine

Whether we are running on Darwin (OS X)

Returns the partition of a string "<key> TAB <partition>"

Splits a line into (key, partition, value).

Same as let (key,_,_) = tab2_split s in key

Same as let (key,p,_) = tab2_split s in (key,p)

Returns the range of the key and the partition

Returns only the range of the key

Splits a line into key and value.

Same as fst (tab_split_key s)

Returns only the range of the key

Returns the list of tasks a given task depends on

Returns: number of input files, number of non-local input files Raise Not_found if the task has never been started

Returns a test object

Like text_field_range, but returns the contents and not the range

let (p,l) = text_field_range s: Extracts a field from s, and returns in p the byte position where the field begins, and in l the byte length.

Like text_field_range, but returns the contents of each field

Like text_field_range, but returns the ranges (p,l) for each field separately

returns the tail of the sequence

Returns the blocks for a blocklist

Converts any list of blocks to the official blocklist representation

String repr for debugging

This is a generic job definition that must be used together with the distributed algorithms in Mapred_toolkit.DSeq.

let e' = trans_e sh vl e debug_name: Runs the engine e until it is finished, and if e was successful, the modifications of the views vl are taken as the modifications to be done in the transaction, and the transaction is committed.

Finds an unused connection, and start a transaction.

Triggers the shutdown of this client

Sets EOF value, and all blocks beyond EOF are deallocated.

Encode bytes that cannot be directly put into text fields with escape sequences

Unlinks the name.

Unlinks the name.

unlink_at trans dir_inode name: Removes the entry name from the directory dir_inode.

Marshal and unmarshal

Updates the cache with this info

Updates the inode with this inodeinfo

update_allocinfo inode ailist bilist: Returns a hash table ht which maps each datanode identity to the selected allocinfo block.

Read the db version of this record, and trigger the update of the module copy of it.

Updates the values that are actually passed as arguments

Garbage-collects the db.

Returns a string with the encoded fields

Like var_field_range, but returns the contents and not the range

let (p,l) = var_field_range s: Extracts a field from s, and returns in p the byte position where the field begins, and in l the byte length.

Like var_field_range, but returns the contents of each field

Like var_field_range, but returns the ranges (p,l) for each field separately

A record consists of a variable number of bytes (at most a bigblock).

Returns (verifier)

wait_e we c m esys cont atomically unlocks the mutex m and suspends the calling engine on the condition variable c.

Busy-waits until the nameslave program is available

Wait until the slaves respond

will_read_block st block: announces that this block will be read in the near future

with_ro_trans f: Runs e = f conn esys for a shared read-only transaction and returns e

with_trans c f: Starts a new transaction t and runs f t.

write_e c inode pos s spos len: Writes data to inode and returns the number of written bytes.

write_admin_table_e key file: Sets the admin table key to file.

write_block st block blockpos mem mpos: Writes a complete block out.

The write is cancelled.

Records that the write is committed: If the buffer is in `Written false state it is reset to `Clean (or reassigned for a different purpose).

Records a write error.

write_multiple fs shm rc prefix size_limit lines_limit create_sync create_async rwf: Writes into a sequence of files whose names are composed of prefix followed by an integer k.

Returns a new store with an automatically generated name.

Write the admin tables back (after modification)