How to Move data from pool to pool (how to drain a pool)

usecase:

we want to retire the node c-2-33 from the pool nodes..so we need to move the data from pool c-2-33_1(physically the data are stored under /dcache/pool/data directory ,we can see the total size is about 135G) to other available pools. One way we could use is the CopyManager Cell provided by the latest dcache..

but there is a limitation about the CopyManager,you should make sure the destination pool has enough space to store the files moved from the source pool..As CopyManager can only move files from a pool to another pool,but can not distribut files to a set of pools.

1 we can check on the web monitor http://head01.aglt2.org:2288/usageInfo ,to see which pool could be used as the destination pool. in this case, we choose pool c-2-32_1 which has enough free space (about 1TB) to hold the source files.

2 configure the CopyManager cell on our headnode head01.aglt2.org or head02.aglt2.org. following the instructionsL

a) make up a copy.batch file in the dir /opt/d-cache/config..its content can be as follows:

root@head01 ~# more /opt/d-cache/config/copy.batch

#
set printout default 3
set printout CellGlue none
onerror shutdown
#
check -strong setupFile
#
copy file:${setupFile} context:setupContext
#
#  import the variables into our $context.
#  don't overwrite already existing variables.
#
import context -c setupContext
#
#   Make sure we got what we need.
#
check -strong serviceLocatorHost serviceLocatorPort
#
create dmg.cells.services.RoutingManager  RoutingMgr
#
#   The LocationManager Part
#
create dmg.cells.services.LocationManager lm \
       "${serviceLocatorHost} ${serviceLocatorPort} "
#
#
#
create diskCacheV111.replicaManager.CopyManager copy0 \
       "default -export"
#
create diskCacheV111.replicaManager.CopyManager copy1 \
       "default -export"
#
create diskCacheV111.replicaManager.CopyManager copy2 \
       "default -export"
#

b) run /opt/d-cache/jobs/initPackage.sh to generate the copy scripts in jobs dir.

root@head01 /opt/d-cache/jobs# ./initPackage.sh

 Checking MasterSetup  ./config/dCacheSetup O.k.

   Scanning dCache batch files

    Processing adminDoor
    Processing copy
    Processing dCache
    Processing dir
    Processing door
    Processing gPlazma
    Processing gridftpdoor
    Processing gsidcapdoor
    Processing httpd
    Processing infoProvider
    Processing lm
    Processing pnfs
    Processing pool
    Processing replica
    Processing srm
    Processing statistics
    Processing utility
    Processing xrootdDoor

 Checking Users database .... Ok
 Checking Security       .... Ok
 Checking JVM ........ Ok
 Checking Cells ...... Ok
 dCacheVersion ....... Version production-1-7-0-38

3 start the CopyManager

root@head01 ~# /opt/d-cache/jobs/copy start

4 check the logfile of copymanager to see if there are any errors

root@head01 ~# tail /var/log/copyDomain.log

5 login the admin interface to move the data:

head01 ~# ssh -l admin -c blowfish -p 22223 localhost

admin@localhost's password: 
    dCache Admin (VII) (user=admin)

(local) admin > 
(local) admin > cd c-2-32_1   # change to the destination pool cell
(c-2-32_1) admin > pp set pnfs timeout 300  #set the pnfs timeout
(c-2-32_1) admin > save
(c-2-32_1) admin > ..
(local) admin > cd c-2-33_1  #change to the source pool cell
(c-2-33_1) admin > p2p set max active 10  #set max active number
(c-2-33_1) admin > mover ls  #check if there are any mover running
(c-2-33_1) admin > pp ls     #check if there are any transfers 
(c-2-33_1) admin > ..
(local) admin > cd copy0     #change to the copymanager cell,the name is correspondding to the one you specified in your copy.batch file
(copy0) admin > copy c-2-33_1 c-2-32_1 -max=5  #copy data from source pool to destination pool
(copy0) admin > info  #check the process of data moving.
     Name : copy0
    Class : diskCacheV111.replicaManager.CopyManager
  Version : $Id: CopyManager.java,v 1.4 2005/10/03 21:45:00 patrick Exp $
     Mode : ACTIVE 
   Status : Processing 2134 files
 Transfer : c-2-33_1 -> c-2-32_1 
  Started : Thu Sep 20 16:55:15 EDT 2007
    Param : Progress(f,b)=1,7346639;Failed(f,b)=0,0;Active(c,m)=5,5
 Progress
 +----------------------------------------+
 |                                        | 0 %
 +----------------------------------------+
 Average Speed : 3.0515246 MBytes/second

(copy0) admin > info

  
   Name : copy0
    Class : diskCacheV111.replicaManager.CopyManager
  Version : $Id: CopyManager.java,v 1.4 2005/10/03 21:45:00 patrick Exp $
     Mode : IDLE 
   Status : Done with 2134 files
 Transfer : c-2-33_1 -> c-2-32_1 
  Started : Thu Sep 20 16:55:15 EDT 2007
 Finished : Thu Sep 20 18:54:11 EDT 2007
    Param : Progress(f,b)=2134,143715233259;Failed(f,b)=0,0;Active(c,m)=0,5
 Progress
 +----------------------------------------+
 |****************************************| 100 %
 +----------------------------------------+
 Average Speed : 19.206345 MBytes/second

now it is finished.

6 .check the destination pool

root@c-2-32 /dcache/pool/data# ll |wc -l

2142

which ,before the copy process,I checked the file in this dir,it is

root@c-2-32 /dcache/pool/data# ll |wc -l

8

it means all the files have been copied from pool c-2-33_1 to pool c-2-32_1..it has nothing to do with the replicas count.It does a complete pool to pool copy..

7 .remove the file on pool c-2-33_1 as you wish..

references:

https://twiki.grid.iu.edu/twiki/bin/view/Storage/DcacheWorkshop

http://www.dcache.org/manuals/experts_docs/Resilient_dCache_v1_0.html

http://t2.unl.edu/documentation/dcache_replicas

-- WenjingWu - 21 Sep 2007