Need to spin up a m5.24xlarge 16.04 Ubuntu instance: Number of vCPUs 96: Attach a 15 T Volume (used this as the location to import raw data and output saved results): Modify the volume so that it automounts each time you start and stop instance as follows:
/etc/fstab
with changes:
UUID=37ef5286-57c8-4ab2-83a7-401fcf9a8a0d /data-extended ext4 defaults,nofail 0 2
the UUID will be different for each drive
Can get the volume device name using:
lsblk
And then use volume device name to get the UUID for the drive using
blkid
Raw data and processed data can be upwards of 6TB for 6 data sets that were 180GB of raw data. SMRTlink itself is 130G. Each job is approximately 200 -300 G
download the smrtlink.run file through link from PacBio website and secure copy the .run file into your 16.04 Ubuntu instance
In a tmux or screen session run:
chmod +x smrtlink_6.0.0.47841.run
Next need to export the path to run commands from any directory:
./smrtlink_6.0.0.47841.run
export PATH=$PATH:/home/ubuntu/smrtlink/admin/bin
echo $PATH
Then run:
services-start
Use all the default ports and parameters, next run:
import-canneddata
run-sat-services
First run:
services-stop
Then run command in order to set the DNS name :
smrt_reconfigure
Keep all existing settings however - choose NO option for DNS
of the choices specify 5) and enter new DNS name We gave the instance an elastic IP and a DNS entry in Route 53
Then Run:
services-start
**If configured DNS as above then only need to do services-start each time you start instance **If done using instance do services-stop prior to stopping instance
You need to reconfigure the DNS using smrt_reconfigure and use the public DNS name for the instance each time you restart the instance.
In our security groups we added inbound rules for ports : Port 9090 - this is the port the SMRTlink uses as a default. Added inbound rules for port 8243 also.
Differs depending on where data coming from. Can use lftp or other file transfer protocol.
#The data directory: pushing it to the GUI specified the location of the volume in the pbservice import dataset command #and the results diretory are pushed into the 15T volume attached
pbservice import-dataset --user xxxx --ask-pass --host xxxx /data/data2/pb/r54290_20190202_025740/5_E01/m54290_190203_201037.subreadset.xml
Then go to Google Chrome type the DNS name into chrome and add --> :9090
Then use login specified in set-up (default admin, admin)
You can now interact via the GUI interface.