IPP Software Navigation Tools IPP Links Communication Pan-STARRS Links
wiki:MinidvodbCopy

MinidvodbCopy

This describes how to use minidvodbcopy.pro (with pantasks) to automatically transfer the minidvodbs to a machine.

Using the module in pantasks

General description

This module is intended to be loaded into the addstar pantasks running on ipp004. This is because this is where the minidvodbs live. Like addstar.pro/minidvodb.pro, this does not need controller host add (host).

server module minidvodbcopy.pro

To queue up and start the copy process:

add.minidvodb.copy (minidvodb_group) (host) (minidvodbcopy_path)

This will queue up all the minidvodbs in 'merged' state, will copy them to /data/ipp022.0/minidvodbs/ThreePi.V1.Mini on ipp022(2). Note that we are using ipp0222 in order for the transfer to work. Also note that ipp004 must be able to ssh/rsync to ipp022 without a password. There are checks on this for each minidvodb queued (there must be 97% or less used, and there must be enough space to transfer the minidvodb onto the host). This is similar to the queueing mechanism for creating dvodbs - it will only queue minidvodbs that have not been queued yet for ipp022(2), there should only be 1 copy of a minidvodb on a machine.

Then, you need to tell pantasks which hosts you want to turn on the transfer - this is a bit like adding labels to determine what is processed:

add.minidvodb.host (host)

Example

An example of how to queue up and copy ThreePi.V1 onto ipp022(2):

add.minidvodb.copy ThreePi ipp0222 /data/ipp022.0/gpc1/mini-catdirs/ThreePi
add.minidvodb.host ipp0222

Other useful tasks

del.minidvodb.host (host) deletes the host from the list of hosts to transfer
show.minidvodb.host shows the current hosts used for transfer
del.minidvodb.copy (host) (minidvodb_group) removes (host) (minidvodb_group) from queuing/loading/copying
show.minidvodb.copy shows which (host).(minidvodbs) are in pantasks
minidvodb.copy.on turns on/off all the minidvodb.copy tasks
minidvodb.copy.off
minidvodb.copy.load.on turns off individual copy tasks
minidvodb.copy.load.off
minidvodb.copy.queue.on
minidvodb.copy.queue.off
minidvodb.copy.run.on
minidvodb.copy.run.off

MinidvodbCopy table

Information about the state of the minidvodb transfers are available on the minidvodbCopy table:

Field Type Null Key Default Extra Comment
minidvodbcopy_id bigint(20) NO PRI NULL auto_increment
minidvodb_id bigint(20) YES MUL NULL which minidvodb_id to be copied
minidvodb_rsync_path varchar(255) NO NULL where to copy it (path)
destination_host varchar(255) NO NULL where to copy it (host)
fault smallint(6) NO NULL were there faults copying it
state varchar(64) NO NULL should be either 'new' or 'full'
epoch timestamp NO CURRENT_TIMESTAMP Time of copy (or when the table was last updated
dtime float YES NULL Total time the transfer took

minidvodbtool

You can use minidvodbtool to find out information about the minidvodbCopy table, to revert/queue/update items. minidvodbtool -h gives the available options. Currently there is no revert task - I need to think about how to do this, which ones can and should be reverted. For now, if there are faults, please investigate carefully why they failed, then revert by hand using -updateminidvodbcopy

minidvodbcopy logs

Please, if you run into problems, or do unusual/irregular things, please update the log.

http://svn.pan-starrs.ifa.hawaii.edu/trac/ipp/wiki/MinidvodbCopyLog

Last modified 15 years ago Last modified on Jun 13, 2011, 12:10:00 PM
Note: See TracWiki for help on using the wiki.