4 title: "Storing and Retrieving data using Keep"
7 This tutorial introduces you to the Arvados file storage system.
10 *This tutorial assumes that you are "logged into an Arvados VM instance":{{site.baseurl}}/user/getting_started/ssh-access.html#login, and have a "working environment.":{{site.baseurl}}/user/getting_started/check-environment.html*
12 The Arvados distributed file system is called *Keep*. Keep is a content-addressable file system. This means that files are managed using special unique identifiers derived from the _contents_ of the file, rather than human-assigned file names (specifically, the md5 hash). This has a number of advantages:
13 * Files can be stored and replicated across a cluster of servers without requiring a central name server.
14 * Systematic validation of data integrity by both server and client because the checksum is built into the identifier.
15 * Minimizes data duplication (two files with the same contents will result in the same identifier, and will not be stored twice.)
16 * Avoids data race conditions (an identifier always points to the same data.)
18 h1. Putting Data into Keep
20 We will start with downloading a freely available VCF file from the "Personal Genome Project (PGP)":http://www.personalgenomes.org subject "hu599905":https://my.personalgenomes.org/profile/hu599905 to a staging directory on the VM, and then add it to Keep.
22 In the following tutorials, replace <b><code>you</code></b> with your user id.
24 First, log into the Arvados VM instance and set up the staging area:
26 notextile. <pre><code>~$ <span class="userinput">mkdir /scratch/<b>you</b></span></code></pre>
28 Next, download the file:
31 <pre><code>~$ <span class="userinput">cd /scratch/<b>you</b></span>
32 /scratch/<b>you</b>$ <span class="userinput">curl -o var-GS000016015-ASM.tsv.bz2 'https://warehouse.personalgenomes.org/warehouse/f815ec01d5d2f11cb12874ab2ed50daa+234+K@ant/var-GS000016015-ASM.tsv.bz2'</span>
33 % Total % Received % Xferd Average Speed Time Time Time Current
34 Dload Upload Total Spent Left Speed
35 100 216M 100 216M 0 0 10.0M 0 0:00:21 0:00:21 --:--:-- 9361k
39 {% include 'notebox_begin' %}
41 If you have your own data, for example @MyData.vcf@, you can use @scp@ or @rsync@ to copy from your local workstation to the shell VM (run this on your local workstation):
43 notextile. <pre><code>~$ <span class="userinput">scp MyData.vcf <b>you@shell</b>.arvados:/scratch/<b>you</b>/MyData.vcf</span></code></pre>
45 {% include 'notebox_end' %}
47 Now use @arv keep put@ to add your VCF data to Keep, then delete the local copy of the file:
50 <pre><code>/scratch/<b>you</b>$ <span class="userinput">arv keep put var-GS000016015-ASM.tsv.bz2</span>
51 c1bad4b39ca5a924e481008009d94e32+210
52 /scratch/<b>you</b>$ <span class="userinput">rm var-GS000016015-ASM.tsv.bz2</span>
56 The output value @c1bad4b39ca5a924e481008009d94e32+210@ from @arv keep put@ is the Keep locator. This enables you to access the file you just uploaded, and is explained in the next section.
58 h2(#dir). Putting a directory
60 You can also use @arv keep put@ to add an entire directory:
63 <pre><code>/scratch/<b>you</b>$ <span class="userinput">mkdir tmp</span>
64 /scratch/<b>you</b>$ <span class="userinput">echo "hello alice" > tmp/alice.txt</span>
65 /scratch/<b>you</b>$ <span class="userinput">echo "hello bob" > tmp/bob.txt</span>
66 /scratch/<b>you</b>$ <span class="userinput">echo "hello carol" > tmp/carol.txt</span>
67 /scratch/<b>you</b>$ <span class="userinput">arv keep put tmp</span>
69 887cd41e9c613463eab2f0d885c6dd96+83
73 The locator @887cd41e9c613463eab2f0d885c6dd96+83@ represents a collection with multiple files.
75 h1. Getting Data from Keep
79 You may access collections through the "Collections section of Arvados Workbench":https://{{ site.arvados_workbench_host }}/collections located at "https://{{ site.arvados_workbench_host }}/collections":https://{{ site.arvados_workbench_host }}/collections . You can also access individual collections and individual files within a collection. Some examples:
81 * "https://{{ site.arvados_workbench_host }}/collections/c1bad4b39ca5a924e481008009d94e32+210":https://{{ site.arvados_workbench_host }}/collections/c1bad4b39ca5a924e481008009d94e32+210
82 * "https://{{ site.arvados_workbench_host }}/collections/887cd41e9c613463eab2f0d885c6dd96+83/alice.txt":https://{{ site.arvados_workbench_host }}/collections/887cd41e9c613463eab2f0d885c6dd96+83/alice.txt
84 h2(#arv-get). Using arv-get
86 You can view the contents of a collection using @arv keep ls@:
89 <pre><code>/scratch/<b>you</b>$ <span class="userinput">arv keep ls c1bad4b39ca5a924e481008009d94e32+210</span>
90 var-GS000016015-ASM.tsv.bz2
93 <pre><code>/scratch/<b>you</b>$ <span class="userinput">arv keep ls 887cd41e9c613463eab2f0d885c6dd96+83</span>
100 Use @-s@ to print file sizes rounded up to the nearest kilobyte:
103 <pre><code>/scratch/<b>you</b>$ <span class="userinput">arv keep ls -s c1bad4b39ca5a924e481008009d94e32+210</span>
104 221887 var-GS000016015-ASM.tsv.bz2
108 Use @arv keep get@ to download the contents of a collection and place it in the directory specified in the second argument (in this example, @.@ for the current directory):
111 <pre><code>/scratch/<b>you</b>$ <span class="userinput">arv keep get c1bad4b39ca5a924e481008009d94e32+210/ .</span>
115 You can also download indvidual files:
118 <pre><code>/scratch/<b>you</b>$ <span class="userinput">arv keep get 887cd41e9c613463eab2f0d885c6dd96+83/alice.txt .</span>
122 With a local copy of the file, we can do some computation, for example computing the md5 hash of the complete file:
125 <pre><code>/scratch/<b>you</b>$ <span class="userinput">md5sum var-GS000016015-ASM.tsv.bz2</span>
126 44b8ae3fde7a8a88d2f7ebd237625b4f var-GS000016015-ASM.tsv.bz2
132 Use @arv-mount@ to take advantage of the "File System in User Space / FUSE":http://fuse.sourceforge.net/ feature of the Linux kernel to mount a Keep collection as if it were a regular directory tree.
135 <pre><code>/scratch/<b>you</b>$ <span class="userinput">mkdir mnt</span>
136 /scratch/<b>you</b>$ <span class="userinput">arv-mount --collection c1bad4b39ca5a924e481008009d94e32+210 mnt &</span>
137 /scratch/<b>you</b>$ <span class="userinput">cd mnt</span>
138 /scratch/<b>you</b>/mnt$ <span class="userinput">ls</span>
139 var-GS000016015-ASM.tsv.bz2
140 /scratch/<b>you</b>/mnt$ <span class="userinput">md5sum var-GS000016015-ASM.tsv.bz2</span>
141 44b8ae3fde7a8a88d2f7ebd237625b4f var-GS000016015-ASM.tsv.bz2
142 /scratch/<b>you</b>/mnt$ <span class="userinput">cd ..</span>
143 /scratch/<b>you</b>$ <span class="userinput">fusermount -u mnt</span>
147 You can also mount the entire Keep namespace in "magic directory" mode:
150 <pre><code>/scratch/<b>you</b>$ <span class="userinput">mkdir mnt</span>
151 /scratch/<b>you</b>$ <span class="userinput">arv-mount mnt &</span>
152 /scratch/<b>you</b>$ <span class="userinput">cd mnt/c1bad4b39ca5a924e481008009d94e32+210</span>
153 /scratch/<b>you</b>/mnt/c1bad4b39ca5a924e481008009d94e32+210$ <span class="userinput">ls</span>
154 var-GS000016015-ASM.tsv.bz2
155 /scratch/<b>you</b>/mnt/c1bad4b39ca5a924e481008009d94e32+210$ <span class="userinput">md5sum var-GS000016015-ASM.tsv.bz2</span>
156 44b8ae3fde7a8a88d2f7ebd237625b4f var-GS000016015-ASM.tsv.bz2
157 /scratch/<b>you</b>/mnt/c1bad4b39ca5a924e481008009d94e32+210$ <span class="userinput">cd ../..</span>
158 /scratch/<b>you</b>$ <span class="userinput">fusermount -u mnt</span>
162 Using @arv-mount@ has several significant benefits:
164 * You can browse, open and read Keep entries as if they are regular files.
165 * It is easy for existing tools to access files in Keep.
166 * Data is downloaded on demand, it is not necessary to download an entire file or collection to start processing