FAQ
Q. If I enter Ctrl+S during interactive jobs, I cannot enter keys after that
This is because standard terminal emulators for macOS, Windows, and Linux have Ctrl+S/Ctrl+Q flow control enabled by default. To disable it, execute the following in the terminal emulator of the local PC:
$ stty -ixon
Executing while logged in to the interactive node has the same effect.
Q. The Group Area is consumed more than the actual size
Generally, any file systems have their own block size, and even the smallest file consumes the capacity of the block size.
ABCI sets the block size of the Group Area 1,2,3 to 128 KB and the block size of the home area and Group Area to 4 KB. For this reason, if a large number of small files are created in the Group Area 1,2,3 , usage efficiency will be reduced. For example, if you want to create a file that is less than 4KB in the Group Area, you need about 32 times the capacity of the home area.
Q. Singularity cannot use container registries that require authentication
SingularityPRO has a function equivalent to docker login that provides authentication information with environment variables.
[username@es1 ~]$ export SINGULARITY_DOCKER_USERNAME='username'
[username@es1 ~]$ export SINGULARITY_DOCKER_PASSWORD='password'
[username@es1 ~]$ singularity pull docker://myregistry.azurecr.io/namespace/repo_name:repo_tag
For more information on SingularityPRO authentication, see below.
Q. NGC CLI cannot be executed
When running NGC Catalog CLI on ABCI, the following error message appears and execution is not possible. This is because the NGC CLI is built for Ubuntu 14.04 and later.
ImportError: /lib64/libc.so.6: version `GLIBC_2.18' not found (required by /tmp/_MEIxvHq8h/libstdc++.so.6)
[89261] Failed to execute script ngc
By preparing the following shell script, it can be executed using Singularity. This technique can be used not only for NGC CLI but also for general use.
1 2 3 4 5 6 | |
Q. I want to assign multiple compute nodes and have each compute node perform different processing
If you give -l rt_F=N or -l rt_AF=N option to qrsh or qsub, you can assign N compute nodes. You can also use MPI if you want to perform different processing on each assigned compute node.
$ module load openmpi/2.1.6
$ mpirun -hostfile $SGE_JOB_HOSTLIST -np 1 command1 : -np 1 command2 : ... : -np1 commandN
Q. I want to avoid to close SSH session unexpectedly
The SSH session may be closed shortly after connecting to ABCI with SSH. In such a case, you may be able to avoid it by performing KeepAlive communication between the SSH client and the server.
To enable KeepAlive, set the option ServerAliveInterval to about 60 seconds in the system ssh configuration file (/etc/ssh/ssh_config) or per-user configuration file (~/.ssh/config) on the user's terminal.
[username@yourpc ~]$ vi ~/.ssh/config
[username@yourpc ~]$ cat ~/.ssh/config
(snip)
Host as.abci.ai
ServerAliveInterval 60
(snip)
[username@userpc ~]$
Note
The default value of ServerAliveInterval is 0 (no KeepAlive).
Q. I want to use a newer version of Open MPI
ABCI offers CUDA-aware and CUDA non-aware versions of Open MPI, and you can check the availability provided by Open MPI.
The Environment Modules provided by ABCI will attempt to configure CUDA-aware Open MPI environment when loading openmpi module only if cuda module has been loaded beforehand.
For the combination where CUDA-aware MPI is provided (cuda/10.0/10.0.130.1, openmpi/2.1.6), therefore, the environment settings will succeed:
$ module load cuda/10.0/10.0.130.1
$ module load openmpi/2.1.6
$ module list
Currently Loaded Modulefiles:
1) cuda/10.0/10.0.130.1 2) openmpi/2.1.6
For the combination where CUDA-aware MPI is not provided (cuda/9.1/9.1.85.3, openmpi/3.1.6), the environment setup will fail and openmpi module will not be loaded:
$ module load cuda/9.1/9.1.85.3
$ module load openmpi/3.1.6
ERROR: loaded cuda module is not supported.
WARNING: openmpi/3.1.6 cannot be loaded due to missing prereq.
HINT: at least one of the following modules must be loaded first: cuda/9.2/9.2.88.1 cuda/9.2/9.2.148.1 cuda/10.0/10.0.130.1 cuda/10.1/10.1.243 cuda/10.2/10.2.89 cuda/11.0/11.0.3 cuda/11.1/11.1.1 cuda/11.2/11.2.2
$ module list
Currently Loaded Modulefiles:
1) cuda/9.1/9.1.85.3
On the other hand, there are cases where CUDA-aware version of Open MPI is not necessary, such as when you want to use Open MPI just for parallelization by Horovod. In this case, you can use a newer version of Open MPI that does not support CUDA-aware functions by loading openmpi module first.
$ module load openmpi/3.1.6
$ module load cuda/9.1/9.1.85.3
module list
Currently Loaded Modulefiles:
1) openmpi/3.1.6 2) cuda/9.1/9.1.85.3
Note
The functions of CUDA-aware versions of Open MPI can be found on the Open MPI site: FAQ: Running CUDA-aware Open MPI
Q. I want to know how ABCI job execution environment is congested
ABCI operates a web service that visualizes job congestion status as well as utilization of compute nodes, power consumption of the whole datacenter, PUE, cooling facility, etc.
The service runs on an internal server, named vws1, on 3000/tcp port.
You can access it by following the procedure below.
You need to set up SSH tunnel.
The following example, written in $HOME/.ssh/config on your PC, sets up the SSH tunnel connection to ABCI internal servers through as.abci.ai by using ProxyCommand.
Please also refer to the procedure in Login using an SSH Client::General method in ABCI System User Environment.
Host *.abci.local
User username
IdentityFile /path/identity_file
ProxyCommand ssh -W %h:%p -l username -i /path/identity_file as.abci.ai
You can create an SSH tunnel that transfers 3000/tcp on your PC to 3000/tcp on vws1.
[username@userpc ~]$ ssh -L 3000:vws1:3000 es.abci.local
You can access the service by opening http://localhost:3000/ on your favorite browser.
Q. Are there any pre-downloaded datasets?
Please see Datasets.
Q. Image file creation with Singularity pull fails in batch job
When you try to create an image file with Singularity pull in a batch job, the mksquashfs executable file may not be found and the creation may fail.
INFO: Converting OCI blobs to SIF format
FATAL: While making image from oci registry: while building SIF from layers: unable to create new build: while searching for mksquashfs: exec: "mksquashfs": executable file not found in $PATH
The problem can be avoided by adding /usr/sbin to PATH like this:
Example)
[username@g0001 ~]$ export PATH="$PATH:/usr/sbin"
[username@g0001 ~]$ module load singularitypro
[username@g0001 ~]$ singularity run --nv docker://caffe2ai/caffe2:latest
Q. I get an error due to insufficient disk space, when I ran the singularity build/pull on the compute node.
The singularity build and pull commands use /tmp as the location to create temporary files.
When you build a large container on the compute node, it may cause an error due to insufficient space in /tmp.
FATAL: While making image from oci registry: error fetching image to cache:
while building SIF from layers: conveyor failed to get: writing blob: write
/tmp/0123456789.1.gpu/bundle-temp-0123456789/oci-put-blob0123456789:
no space left on device
If you get an error due to insufficient space, set the SINGULARITY_TMPDIR environment variable to use the local storage as shown below:
[username@g0001 ~]$ SINGULARITY_TMPDIR=$SGE_LOCALDIR singularity pull docker://nvcr.io/nvidia/tensorflow:20.12-tf1-py3
Q. How can I find the job ID?
When you submit a batch job using the qsub command, the command outputs the job ID.
[username@es1 ~]$ qsub -g grpname test.sh
Your job 1000001 ("test.sh") has been submitted
If you are using qrsh, you can get the job ID by retrieving the value of the JOB_ID environment variable.This variable is available for qsub (batch job environment) as well.
[username@es1 ~]$ qrsh -g grpname -l rt_C.small=1 -l h_rt=1:00:00
[username@g0001 ~]$ echo $JOB_ID
1000002
[username@g0001 ~]$
To find the job ID of your already submitted job, use the qstat command.
[username@es1 ~]$ qstat
job-ID prior name user state submit/start at queue jclass slots ja-task-ID
------------------------------------------------------------------------------------------------------------------------------------------------
1000003 0.00000 test.sh username qw 08/01/2020 13:05:30
To find the job ID of your completed job, use qacct -j. The -b and -e options are useful for narrowing the search range. See qacct(1) man page (type man qacct on an interactive node). The following example lists the completed jobs that started on and after September 1st, 2020. jobnumber has the same meaning as job-ID.
[username@es1 ~]$ qacct -j -b 202009010000
==============================================================
qname gpu
hostname g0001
group grpname
owner username
:
jobname QRLOGIN
jobnumber 1000010
:
qsub_time 09/01/2020 16:41:37.736
start_time 09/01/2020 16:41:47.094
end_time 09/01/2020 16:45:46.296
:
==============================================================
qname gpu
hostname g0001
group grpname
owner username
:
jobname testjob
jobnumber 1000120
:
qsub_time 09/07/2020 15:35:04.088
start_time 09/07/2020 15:43:11.513
end_time 09/07/2020 15:50:11.534
:
Q. I want to run the Linux command on all allocated compute node
ABCI provides the ugedsh command to execute Linux commands in parallel on all allocated compute nodes.
The command specified in the argument of the ugedsh command is executed once on each node.
Example)
[username@es1 ~]$ qrsh -g grpname -l rt_F=2
[username@g0001 ~]$ ugedsh hostname
g0001: g0001.abci.local
g0002: g0002.abci.local
Q. What is the difference between Compute Node (A) and Compute Node (V)
ABCI was upgraded to ABCI 2.0 in May 2021. In addition to the previously provided Compute Nodes (V) with NVIDIA V100, the Compute Nodes (A) with NVIDIA A100 are currently available.
This section describes the differences between the Compute Node (A) and the Compute Node (V), and points to note when using the Compute Node (A).
Resource type name
The Resource type name is different between the Compute Node (A) and the Compute Node (V). Compute Node (A) can be used by specifying the following Resource type name when submitting a job.
| Resource type | Resource type name | Assigned physical CPU core | Number of assigned GPU | Memory (GiB) |
|---|---|---|---|---|
| Full | rt_AF | 72 | 8 | 480 |
| AG.small | rt_AG.small | 9 | 1 | 60 |
For more detailed Resource types, see Available Resource Types.
Accounting
Compute Node (A) and Compute Node (V) have different Resource type charge coefficient, as described at Available Resource types. Therefore, the number of ABCI points used, which are calculated based on Accounting, is also different.
The number of ABCI points used when using the Compute Node (A) is as follows:
| Resource type name Execution Priority |
On-demand or Spot Service Execution Priority: -500 (default) (point/hour) |
On-demand or Spot Service Execution Priority: -400 (point/hour) |
Reserved Service (point/day) |
|---|---|---|---|
| rt_AF | 3.0 | 4.5 | 108 |
| rt_AG.small | 0.5 | 0.75 | N/A |
Operating System
The Compute Node (A) and the Compute Node (V) use different Operating Systems.
| Node | Operating System |
|---|---|
| Compute Node (A) | Red Hat Enterprise Linux 8.2 |
| Compute Node (V) | CentOS Linux 7.5 |
Since the versions of kernels and libraries such as glibc are different, the operation cannot be guaranteed when the program built for the Compute Node (V) is run on the Compute Node (A) as it is.
Please rebuild the program for the Compute Node (A) using the Compute Node (A) or the Interactive Node (A) described later.
CUDA Version
The NVIDIA A100 installed on the compute node (A) is Compute Capability 8.0 compliant.
CUDA 10 and earlier does not support Compute Capability 8.0. Therefore, Compute Node (A) should use CUDA 11 or later that supports Compute Capability 8.0.
Note
Environment Modules makes CUDA 10 available for testing, but its operation is not guaranteed.
Interactive Node (A)
ABCI provides the Interactive Nodes (A) with the same software configuration as the Compute Node (A) for the convenience of program development for the Compute Node (A). The program built on the Interactive Node (A) does not guarantee the operation on the Compute Node (V).
Please refer to the following for the proper use of Interactive Nodes:
| Interactive Node (V) | Interactive Node (A) | |
|---|---|---|
| Can users log in? | Yes | Yes |
| Can users develop programs for Compute Nodes (V)? | Yes | No |
| Can users develop programs for Compute Nodes (A)? | No | Yes |
| Can users submit jobs for Compute Nodes (A)? | Yes | Yes |
| Can users submit jobs for Compute Nodes (A)? | Yes | Yes |
For more information on Interactive Node (A), see Interactive Node.
Q. How to use previous ABCI Environment Modules
ABCI provides previous ABCI Environment Modules.
ABCI Environment Modules for each year are installed below, so set the path of the year you want to use in the MODULE_HOME environment variable and load the configuration file.
Please note that previous ABCI Environment Modules is not eligible for the ABCI System support.
| ABCI Environment Modules | Installed Path |
|---|---|
| 2020 version | /apps/modules-abci-1.0 |
| 2021 version | /apps/modules-abci-2.0-2021 |
The following is an example of using the 2021 version of ABCI Environment Modules.
sh, bash:
export MODULE_HOME=/apps/modules-abci-2.0-2021
. ${MODULE_HOME}/etc/profile.d/modules.sh
ch, tcsh:
setenv MODULE_HOME /apps/modules-abci-2.0-2021
source ${MODULE_HOME}/etc/profile.d/modules.csh
Q. What are the new Group Area and data migration?
In FY2021, we expanded the storage system. Refer to Storage Systems for details. As the storage system is expanded, the configuration of the Group Area has been changed, and data have been migrated from the Group Area used until FY2020 (hereinafter referred to as Old Area) to the new Group Area (hereinafter referred to as New Area).
The all data in the Old Area have been transferred to the New Area, then the source paths, were allocated to the Old Areas have been replaced with symlinks to the destination directories in the New Areas, so that it can be accessed at the same path as the Old Area.
The sources and the destinations of data migration are as follows.
| Source | Destination |
|---|---|
d002 users'/groups[1-2]/gAA50NNN/ |
/projects/datarepository/gAA50NNN/migrated_from_SFA_GPFS/1 |
others'/groups[1-2]/gAA50NNN/ |
/groups/gAA50NNN/migrated_from_SFA_GPFS/ |
/fs3/d001/gAA50NNN/ |
/projects/d001/gAA50NNN/migrated_from_SFA_GPFS/ |
/fs3/d002/gAA50NNN/ |
/projects/datarepository/gAA50NNN/migrated_from_SFA_GPFS3/1 |