Print

Print


Hi,

our DPM which is using native xrootd 4.9.1 yesterday created two
different 128TB files on two different diskservers that were received as
a job output from two different worker nodes. The size of original file
was few megabytes and when I check beginning of stored files up to
original size & checksum everything seems to be OK - except these are
sparse files with size 128TB. Original transferred files from WN looked
like:

mc15_13TeV:log.18636890._001059.job.log.tgz.1 - original size 3101899, adler32 742e1336
mc16_5TeV:AOD.18595963._000786.pool.root.1 - original size 737474514, adler32 dc68cb21

but on our diskservers they were stored as

[root@dpmpool22 ~]# ls -l /mnt/fs1/atlas/2019-07-17/log.18636890._001059.job.log.tgz.1.rucio.upload.209928.1563376157
-rw-rw---- 1 dpmmgr dpmmgr 140560257079236 Jul 17 17:11 /mnt/fs1/atlas/2019-07-17/log.18636890._001059.job.log.tgz.1.rucio.upload.209928.1563376157
[root@dpmpool22 ~]# du --block-size=1 /mnt/fs1/atlas/2019-07-17/log.18636890._001059.job.log.tgz.1.rucio.upload.209928.1563376157
3108864	/mnt/fs1/atlas/2019-07-17/log.18636890._001059.job.log.tgz.1.rucio.upload.209928.1563376157

[root@dpmpool24 ~]# ls -l /mnt/fs2/atlas/2019-07-17/AOD.18595963._000786.pool.root.1.rucio.upload.210480.1563377083
-rw-rw---- 1 dpmmgr dpmmgr 139734345056408 Jul 17 17:31 /mnt/fs2/atlas/2019-07-17/AOD.18595963._000786.pool.root.1.rucio.upload.210480.1563377083
[root@dpmpool24 tmp]# du --block-size=1 /mnt/fs2/atlas/2019-07-17/AOD.18595963._000786.pool.root.1.rucio.upload.210480.1563377083
737480704	/mnt/fs2/atlas/2019-07-17/AOD.18595963._000786.pool.root.1.rucio.upload.210480.1563377083

I was told that DPM is not really involved in transfering / storing
files with xrootd protocol, so I guess something bad must happened
inside XRootD libraries. Do you have any idea why these two transfers
generated such extreme files? Unfortunately our xrootd logging is pretty
low and I was able to find only info about file creation

[root@dpmpool22 ~]# grep log.18636890._001059.job.log.tgz /var/log/xrootd/dpmdisk/xrootd.log
190717 17:10:39 332092 ofs_Posc: Creator changed from atlasprd.2757:90@[2001:718:401:6e03::1:21] to atlasprd.2757:39@[2001:718:401:6e03::1:21] for /dpm/farm.particle.cz/home/atlas/atlasdatadisk/rucio/mc15_13TeV/26/98/log.18636890._001059.job.log.tgz.1.rucio.upload

[root@dpmpool24 tmp]# grep AOD.18595963._000786.pool.root /var/log/xrootd/dpmdisk/xrootd.log
190717 17:29:26 14204 ofs_Posc: Creator changed from atlasprd.2757:33@[2001:718:401:6e03::1:32] to atlasprd.2757:95@[2001:718:401:6e03::1:32] for /dpm/farm.particle.cz/home/atlas/atlasdatadisk/rucio/mc16_5TeV/09/bc/AOD.18595963._000786.pool.root.1.rucio.upload


Petr

########################################################################
Use REPLY-ALL to reply to list

To unsubscribe from the XROOTD-L list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-L&A=1