Difference between revisions of "Checksumming support in SRM implementations"

From PDP/Grid Wiki
Jump to navigationJump to search
m
Line 17: Line 17:
  
 
'''Note''': so-called <tt>gridftp checksums</tt> are checksums that can be calculated ''on-the-fly'' using a special GridFTP command:
 
'''Note''': so-called <tt>gridftp checksums</tt> are checksums that can be calculated ''on-the-fly'' using a special GridFTP command:
<tt>QUOTE CKSM ADLER32 0 -1 full-path</tt>
+
 
 +
<tt> QUOTE CKSM ADLER32 0 -1 full-path</tt>
  
 
= dCache =
 
= dCache =

Revision as of 16:37, 10 March 2010

Some applications have requested the availability of checksumming support on files stored on the storage elements in the Netherlands. The first VO to request this was Atlas, but other VOs are also very interested. Unfortunately, checksumming support seems to be non-homogenously implemented across the different storage backend systems.

This Wiki page is the result of a short investigation into the checksumming differences between dCache, DPM, StoRM and CASTOR. First a short overview is given of the checksumming features of each storage system. At the end the output of the (relevant) commands used in this investigation is shown.

Why only dCache, DPM, StoRM and CASTOR?

At the time of writing (August 25th 2009) the VO dteam had access to 319 SEs in total, of which

  • 195 DPM
  • 64 dCache
  • 35 StoRM
  • 16 CASTOR
  • 3 unknown (but at least one looks like DPM)
  • 6 down

So I did not have access to other SRM systems.

Note: so-called gridftp checksums are checksums that can be calculated on-the-fly using a special GridFTP command:

QUOTE CKSM ADLER32 0 -1 full-path

dCache

  • srmping SARA: v2.2 dCache production-1.9.3-3
  • seems to support only adler32 checksums (which is what WLCG seems to want) ; according to Ron md5 is also supported but this requires a server reconfiguration
  • checksums
    • are computed when the file is transferred to dCache or during a move between pool nodes.
    • are stored in the dCache namespace
    • can be retrieved using the srmls -l command

DPM

  • srmping Nikhef: v2.2 DPM 1.7.0-5
  • supports adler32, md5 and crc32 checksums
  • supports only gridftp checksums; adler32 and crc32 are supported using a special DPM-DSI GridFTP plugin
  • checksums
    • can be computed using the lcg-get-checksum command
    • are stored in the DPM namespace but can never be retrieved from there ( !! ) This is fixed in DPM 1.7.2+
    • are not displayed when using the srmls -l command. This is fixed for DPM 1.7.2+ servers
    • are recalculated every time. This is fixed for DPM 1.7.2+ servers

StoRM

  • srmping Groningen: v2.2 StoRM <FE:1.4.0-01.sl4><BE:1.4.0-00>
  • supports only md5 gridftp checksums
  • checksums
    • can be retrieved using the lcg-get-checksum command; however, the checksum is always the same
    • are not displayed when using the srmls -l command
    • are recalculated every time

CASTOR

  • srmping: v2.2 CASTOR v2_7_15 2.1.7
  • supports only adler32 checksums
  • checksums
    • are computed in the background
    • cannot be specified with the --checksum option
    • can be retrieved using the srmls -l command
    • are stored in the CASTOR name space
  • gridftp checksums are not supported

Command output

dCache

lcg-cr output:

$ lcg-cr -l /grid/pvier/janjust/my-dcache-file-checksum-adler32 -d $SRM file:/user/janjust/myfile --checksum 
guid:d156757d-ecca-472e-adb9-64ebefde23c4

srm-ls output:

$ srmls -l $SRM/pnfs/....
 8508 $SRM/pnfs/....
 space token(s) :15253459
 storage type:PERMANENT
 retention policy:CUSTODIAL
 access latency:NEARLINE
 locality:NEARLINE
 - Checksum value:  cd5d9820
 - Checksum type:  adler32
 UserPermission: uid=18010 PermissionsRW
 GroupPermission: gid=1276 PermissionsR
 WorldPermission: R
 created at:2009/08/22 00:02:18
 modified at:2009/08/22 00:02:18
 - Assigned lifetime (in seconds):  -1
 - Lifetime left (in seconds):  -1
 - Original SURL:  $SRM/pnfs/....
 - Status:  null
 - Type:  FILE

lcg-get-checksum output:

$ lcg-get-checksum $SRM/pnfs/....
cd5d9820	$SRM/pnfs/....

DPM

lcg-cr output:

$ lcg-cr -l /grid/pvier/janjust/my-dpm-file-checksum-adler32 -d $SRM file:/user/janjust/myfile --checksum 
guid:2352aac8-37b9-4cfd-ba91-5eaee71fd5f1

srm-ls output:

$ srmls -l $SRM
 8508 /dpm/science.uu.nl/....
space token(s) :none found
 storage type:PERMANENT
retentionpolicyinfo : null
 locality:ONLINE
  UserPermission: uid=/O=dutchgrid/O=users/O=nikhef/CN=Jan Just Keijser PermissionsRW
  GroupPermission: gid=pvier PermissionsRW
 WorldPermission: R
created at:2009/08/21 23:56:27
modified at:2009/08/21 23:56:27
 - Lifetime left (in seconds):  -1
 - Original SURL:  /dpm/science.uu.nl/....
- Status:  null
- Type:  FILE

lcg-get-checksum output:

$ lcg-get-checksum $SRM/dpm/science.uu.nl/....
cd5d9820	$SRM/dpm/science.uu.nl/....

StoRM

lcg-cr output:

$ lcg-cr -l /grid/pvier/janjust/my-storm-file-checksum-adler32 -d $SRM file:/user/janjust/myfile --checksum 
[LCG-UTIL][lcg_cr4][] Destination may be corrupted: 
Source checksum (cd5d9820) != Destination checksum (d41d8cd98f00b204e9800998ecf8427e )
[LCG-UTIL][lcg_cp4][] $SRM/pvier/.... has been DELETED, please try again (temporary network problem? )
lcg_cr: Resource temporarily unavailable

a different file:

$ lcg-cr -l /grid/pvier/janjust/my-storm-file-checksum-adler32 -d $SRM file:/user/janjust/wms_soap_msgs.txt --checksum 
[LCG-UTIL][lcg_cr4][] Destination may be corrupted:
Source checksum (8d9e8485) != Destination checksum (d41d8cd98f00b204e9800998ecf8427e)
[LCG-UTIL][lcg_cp4][] $SRM/pvier/.... has been DELETED, please try again (temporary network problem?)
lcg_cr: Resource temporarily unavailable

Interesting: the destination checksum is the same, even though it's a completely different file ?!?!?!?!

srm-ls output: not shown as the file was not successfully copied.

For another file that was copied to StoRM without --checksum:

$ lcg-get-checksum $SRM/pvier/....
d41d8cd98f00b204e9800998ecf8427e	$SRM/pvier/....

Wait! That's the exact same checksum as shown above !!

Let's try a different StoRM SRM:

$ lcg-get-checksum srm://storm-fe-cms.cr.cnaf.infn.it/dteam/testfile-cp-20090727-171238-6e877e21562d1.txt
d41d8cd98f00b204e9800998ecf8427e	srm://storm-fe-cms.cr.cnaf.infn.it/dteam/testfile-cp-20090727-171238-6e877e21562d1.txt

$ lcg-get-checksum srm://storm-fe-cms.cr.cnaf.infn.it/dteam/helloworld.txt1248346365.88
d41d8cd98f00b204e9800998ecf8427e	srm://storm-fe-cms.cr.cnaf.infn.it/dteam/helloworld.txt1248346365.88

Again, the same checksum. Note that these files were already present on the SRM system.

So at least we know it's not a site-specific issue ;-)

CASTOR

lcg-cr output:

$ lcg-cr -d $SRM/myfile2 -l lfn:/grid/pvier/janjust/my-castor-file2 file://$PWD/myfile --checksum
[LCG-UTIL][lcg_get_checksum_surls][] $SRM/castor/....:
Communication error on send 
guid:64344a07-f3f3-41e5-9079-c86bf61a874e
lcg_cr: Communication error on send 

Afterwards the file is stored in SRM but not in the LFC.

srmls output (after 10 minutes; the checksum calculation happens in the background):

$ srmls -l $SRM/myfile
 8508 /castor/....
 space token(s) :none found
 type: null
 retentionpolicyinfo : null
  locality:ONLINE_AND_NEARLINE
  - Checksum value:  0xcd5d9820
  - Checksum type:  adler32 
   UserPermission: uid=dteam001 PermissionsRWX
   GroupPermission: gid=dteam PermissionsRWX
  WorldPermission: RW
 created at:2009/08/20 17:11:06

lcg-get-checksum output:

$ lcg-get-checksum s$SRM/castor/....
0xcd5d9820	$SRM/castor/....

Notice that the checksum is preceded here by 0x !!