You are here: NRAO Public Wiki>VLBA Web>Software>Mark6MountingRaidModule (2020-01-31, MarkWainright) Edit wiki text Edit Attach Print version

Mark6 Mounting and Correlating a RAID Module

Put the Mark6 RAID module in a Mark6 playback unit by itself. Currently the RAID modules interfere with mk5daemon properly detecting normal Mark6 modules. Key on the RAID module.

Run the 'lsblk' command to show the RAID device.

The result should look like: 

difx@mark6fx06 VLBADIFX7-2.5.2 ~> lsblk
NAME        MAJ:MIN RM   SIZE RO TYPE  MOUNTPOINT
sda           8:0    0 698.7G  0 disk  
├─sda1        8:1    0   200M  0 part  /boot/efi
├─sda2        8:2    0  39.1G  0 part  /
├─sda3        8:3    0     8G  0 part  [SWAP]
└─sda4        8:4    0 651.4G  0 part  /export/home/mark6
sdb           8:16   0   7.3T  0 disk  
└─md127       9:127  0  14.6T  0 raid0 /export/home/datacopy
sdc           8:32   0   7.3T  0 disk  
└─md127       9:127  0  14.6T  0 raid0 /export/home/datacopy
sdd           8:48   0   7.3T  0 disk  
└─md126       9:126  0  36.4T  0 raid0 
  └─md126p1 259:0    0  36.4T  0 md
sde           8:64   0   7.3T  0 disk  
└─md126       9:126  0  36.4T  0 raid0 
  └─md126p1 259:0    0  36.4T  0 md
sdf           8:80   0   7.3T  0 disk  
└─md126       9:126  0  36.4T  0 raid0 
  └─md126p1 259:0    0  36.4T  0 md
sdg           8:96   0   7.3T  0 disk  
└─md126       9:126  0  36.4T  0 raid0 
  └─md126p1 259:0    0  36.4T  0 md
sdh           8:112  0   7.3T  0 disk  
└─md126       9:126  0  36.4T  0 raid0 
  └─md126p1 259:0    0  36.4T  0 md
sr0          11:0    1  1024M  0 rom

The result indicates five disk devices, sdd to sdh, linked together as RAID0 device md126. We will mount the device at the 'md' (multiple device) partition location which is md126p1 in this case.

User 'difx' has 'sudo mount' and 'sudo umount' permissions.

To mount the RAID module use the 'sudo mount' command with the appropriate 'md' partition location with '/dev/' prefix and mount point '/mnt/raid'. For this case: 

sudo mount -t xfs /dev/md126p1 /mnt/raid

Running 'lsblk' again should confirm the mount. 

difx@mark6fx06 VLBADIFX7-2.5.2 ~> lsblk
NAME        MAJ:MIN RM   SIZE RO TYPE  MOUNTPOINT
sda           8:0    0 698.7G  0 disk  
├─sda1        8:1    0   200M  0 part  /boot/efi
├─sda2        8:2    0  39.1G  0 part  /
├─sda3        8:3    0     8G  0 part  [SWAP]
└─sda4        8:4    0 651.4G  0 part  /export/home/mark6
sdb           8:16   0   7.3T  0 disk  
└─md127       9:127  0  14.6T  0 raid0 /export/home/datacopy
sdc           8:32   0   7.3T  0 disk  
└─md127       9:127  0  14.6T  0 raid0 /export/home/datacopy
sdd           8:48   0   7.3T  0 disk  
└─md126       9:126  0  36.4T  0 raid0 
  └─md126p1 259:0    0  36.4T  0 md    /mnt/raid
sde           8:64   0   7.3T  0 disk  
└─md126       9:126  0  36.4T  0 raid0 
  └─md126p1 259:0    0  36.4T  0 md    /mnt/raid
sdf           8:80   0   7.3T  0 disk  
└─md126       9:126  0  36.4T  0 raid0 
  └─md126p1 259:0    0  36.4T  0 md    /mnt/raid
sdg           8:96   0   7.3T  0 disk  
└─md126       9:126  0  36.4T  0 raid0 
  └─md126p1 259:0    0  36.4T  0 md    /mnt/raid
sdh           8:112  0   7.3T  0 disk  
└─md126       9:126  0  36.4T  0 raid0 
  └─md126p1 259:0    0  36.4T  0 md    /mnt/raid
sr0          11:0    1  1024M  0 rom

We can see '/mnt/raid' following all the 'md' partition locations confirming the mount.

Run 'll /mnt/raid' to see a long listing of the module contents. It might look something like: 

difx@mark6fx06 VLBADIFX7-2.5.2 ~> ll /mnt/raid
total 56
drwxrwxr-x 2 xguest xguest  8192 Jan 11 22:52 rv139_hh
drwxrwxr-x 2 xguest xguest  8192 Jan 13 11:02 rv139_ny
drwxrwxr-x 2 xguest xguest 12288 Jan 12 11:26 rv139_on
drwxrwxr-x 2 xguest xguest 12288 Jan 12 05:06 rv139_wz

For this particular observation Haystack provided the scans for four sites in separate directories.

The next step is to generate .filelist files for the sites. Use 'm5bsum -s' to generate file lists for mark5b data. Use 'vsum -s' to generate file lists for VDIF data. Run the commands in the destination directory for the file lists.

This particular observation is mark5b, so here are the mark5b command examples: 

m5bsum -s /mnt/raid/rv139_hh/* > rv139_hh.filelist
m5bsum -s /mnt/raid/rv139_ny/* > rv139_ny.filelist
m5bsum -s /mnt/raid/rv139_on/* > rv139_on.filelist
m5bsum -s /mnt/raid/rv139_wz/* > rv139_wz.filelist

Here are a few lines of the rv139_hh.filelist output file as an example: 

/mnt/raid/rv139_hh/rv139_hh_008-1737 58856.73457176 58856.73803241
/mnt/raid/rv139_hh/rv139_hh_008-1744b 58856.73944444 58856.74146991
/mnt/raid/rv139_hh/rv139_hh_008-1751a 58856.74434028 58856.74577546
/mnt/raid/rv139_hh/rv139_hh_008-1756 58856.74729167 58856.74805555
/mnt/raid/rv139_hh/rv139_hh_008-1759 58856.74953704 58856.75003472

The sites with RAID data, hh, ny, on and wz in this case, will need to be updated in the .v2d file with the 'filelist=rv139_xx.filelist' tag, where xx is the site.

After the RAID sites are updated in the .v2d, normal use of mk62v2d and vex2difx should generate the jobs.

After the jobs are queued, to correlate we will have to manually run genmachines on the jobs. The RAID site file datastreams will probably show up in the .machines file as swcNNN units. The .machines file will have to be edited to replace the swcNNN datastream machine with the mark6 playback machine containing the RAID module.

Finally, we will have to run the jobs manually with 'startdifx -v -f -n .input'.

When everything is finished use the 'sudo umount' command to unmount the RAID module. 

sudo umount /mnt/raid

Lastly, the module should be re-initialized as a normal scatter/gather module for distribution to VLBA sites.
Edit  | Attach | Print version | History: r1 | Backlinks | View wiki text | Edit wiki text | More topic actions
Topic revision: r1 - 2020-01-31, MarkWainright
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding NRAO Public Wiki? Send feedback