Releases · zhenrong-wang/hpc-now

25 Dec 05:46

0.3.1.beta9

379a251

IMPORTANT: This upgrade is not compatible with 0.2.x

Main Changes: Added MD5sum and BASE64 modules

As you may know, the HPC-NOW needs to calculate the md5 checksums and base64 decoded strings. Previously, it relies on system utilities such as certutil (for Microsoft Windows) md5sum base64 (for GNU/Linux and macOS distros), etc.

Now, we have introduced the internal modules for calculating MD5sum and base64.

We also added a CONTRIBUTING.md to this project, looking forward to your feedback and contributions!

HPC-NOW: Start your HPC journey in the cloud now.

Assets 5

0 Join discussion

21 Dec 10:12

zhenrong-wang

0.3.1.beta7

1fe35d9

[BETA] Version 0.3.1.0007 released!

Information security is the top issue.
IMPORTANT: This upgrade is not compatible with 0.2.x

Main Changes:

1. Your SSH private keys are encrypted and protected.

As you may know, the HPC-NOW generates and manages SSH keys for you to connect with the cloud clusters. SSH private keys are sensitive, especially the global operator private key in the SSH_KEY_FOLDER:

For Microsoft Windows: C:\Programdata\hpc-now\now-ssh
For GNU/Linux: /usr/.hpc-now/.now-ssh
For macOS (Darwin): /Applications/.hpc-now/.now-ssh

Previously, the private keys were hidden, but not encrypted. Now, with this release, the private keys will be encrypted locally.

2. New commands added.

For developers, we introduced 2 important commands: hpcopr encrypt and hpcopr decrypt. These 2 commands are convenient for users to decrypt sensitive files of one or several clusters. Therefore, users can check the TF files, cloud secrets, as well as cluster users' passwords easily.

However, the hpcopr decrypt command is VERY RISKY!!!, and WE DO NOT RECOMMEND USE IT UNLESS YOU KNOW ALL THE CONSEQUENCES VERY CLEARLY:

The cloud secrets will be decrypted to the vault directory
The user passwords will be decrypted to the vault directory
The Terraform/openTofu files containing sensitive information will be decrypted to the stack directory
The users SSH private keys will be decrypted to the cluster's sshkey folder under SSH_KEY_FOLDER (See the paths above)

Please do encrypt the decrypted files by using hpcopr encrypt as soon as possible.

We also added a CONTRIBUTING.md to this project, looking forward to your feedback and contributions!

HPC-NOW: Start your HPC journey in the cloud now.

Assets 5

0 Join discussion

14 Dec 08:14

zhenrong-wang

0.3.1.beta1

6ffda93

[BETA] First beta version of 0.3.1 released!

Information security is the top issue.
IMPORTANT: This upgrade is not compatible with previous versions, that is, 0.2.x

Main Changes: Strongly enhanced the information security by introducing AES-128 encryption

The HPC-NOW services handles sensitive information, typically are cloud access key (known as AK) and access secret key(known as SK), or key files (for Google Cloud Platform). Meanwhile, when managing clusters in the cloud, more sensitive information such as passwords are generated. All the information mentioned MUST be handled and protected carefully.

Since Day One of this project, we have been considering information security. That's why we have an independent crypto module named now-crypto for the running of HPC-NOW services ever since the first release of this project. However, the previous crypto module has been a weak point, because it is simple, and a not standard one.

Therefore, after days of work, we introduced a genuine, data block level, industrial standard, wide-recognized crypto module - AES-128 (ECB mode) with PKCS padding. With the new crypto module, the security level of HPC-NOW really meets the requirement of production. In our development plan, this AES-based module will be a long-term part of HPC-NOW.

We strongly recommend you start your HPC-NOW journey from this version - 0.3.x

And, VERY IMPORTANT: DO NOT directly update from 0.2.x to 0.3.x. If you'd like to update, please submit issues to us. We are also developing a utility to make the update safe and smooth.

We also added a CONTRIBUTING.md to this project, looking forward to your feedback and contributions!

HPC-NOW: Start your HPC journey in the cloud now.

Assets 5

2 Join discussion

10 Dec 12:50

zhenrong-wang

0.2.1.beta.25

a7238a2

[BETA] Release version v0.2.1 evolved to 0.2.1.0025!

This is an update to the Version 0.2.1.0019.

Main changes:

1. Introduced interactive mode when initializing a cluster

This is a major update. Previously, when initializing a cluster by the command hpcopr init, you need to specify parameters in the command line, which means that only batch mode is supported.

Now you can enter interactive mode to select options one-by-one, from region selection, az selection, image, to shared-volume capacity.

Please try it out.

We also added a CONTRIBUTING.md to this project, looking forward to your feedback and contributions!

HPC-NOW: Start your HPC journey in the cloud now.

Assets 5

1 Join discussion

05 Dec 10:46

zhenrong-wang

0.2.1.beta.19

d19cac6

[BETA] Release version v0.2.1 evolved to 0.2.1.0019!

This is an update to the Version 0.2.1.0015.

Main changes:

1. Enhanced the switch between terraform and tofu

Once a cluster initialized, the tf execution should not be switched between terraform and tofu. Therefore, we enhanced this feature by putting the tf execution configuration into the cluster level.

Now, each cluster has its own specific tf execution configuration during its lifecycle, and this configuration has the highest priority.

To be more clear, here is the priority of tf execution configuration options:

Highest: the cluster's own configuration after being initialized
Medium: the options specified in command line by using --tf-run --dbg-level and/or --max-time
Lowest: the static configuration saved in current hpcopr environment (you can update it by hpcopr set-tf command)

Therefore, please take care of these 3 types of options.

2. Enhanced the robustness by adding width control to the scanf() and fscanf() functions

3. Validated and updated the openTofu version to 1.6.0-beta2

Updated the version and MD5s of openTofu, please use hpcopr repair command to use the latest component.

4. Command deprecation

None.

We also added a CONTRIBUTING.md to this project, looking forward to your feedback and contributions!

HPC-NOW: Start your HPC journey in the cloud now.

Assets 5

0 Join discussion

04 Dec 09:54

zhenrong-wang

0.2.1.beta.15

e7b85a2

[BETA] Release version v0.2.1 evolved to 0.2.1.0015!

This is a major update to the Version 0.2.1.0011.

Main changes:

1. INTRODUCED the switch between terraform and tofu

Now, HPC-NOW generates and manages a configuration file for tf running (either terraform or opentofu). Your can use the new command below to check and update the configuration file.

hpcopr set-tf: without extra command options, the current config will be displayed.

hpcopr set-tf --tf-run TF_NAME --dbg-level DEBUG_LOG_LEVEL --max-time MAX_WAITING_SECONDS, where

TF_NAME can be either terraform or tofu, currently, terraform is the default option.
DEBUG_LOG_LEVEL can be one of trace debug info warn error off.
MAX_WAITING_SECONDS can be a value between 600 and 1200

2. INTRODUCED 3 command keywords to switch between terraform and opentofu in running

Now, you can use the command keywords --tf-run TF_NAME --dbg-level DEBUG_LOG_LEVEL --max-time MAX_WAITING_SECONDS to specify the tf running config for the current run. This will override the configs set by the set-tf command above.

Example: hpcopr init -b --tf-run terraform --dbg-level error --max-time 800

This example command will start initializing a cluster with terraform, only output error logs, and with maximum execution time 800 seconds. Check the effect below:

3. Better wrapped tf_execution

The tf_execution() function was updated with a new struct type tf_exec_config. With this method, the whole codes are much more readable and extendable.

4. Extra information was included in the repo

Uploaded the MD5s and versions of the third-party components. See the new directory vers_md5s folder.

5. Command deprecation

None.

6. Deleted the dev-0.2.2-opentofu branch

Because the switch between terraform and opentofu was introduced, the previously-branched dev-0.2.2-opentofu is now outdated. We deleted it from the repository.

We also added a CONTRIBUTING.md to this project, looking forward to your feedback and contributions!

HPC-NOW: Start your HPC journey in the cloud now.

Assets 5

0 Join discussion

01 Dec 23:14

zhenrong-wang

0.2.1.beta.11

34924b4

[BETA] Release version v0.2.1 evolved to 0.2.1.0011!

This is a fix to the Version 0.2.1.0009.

Main changes:

1. FIXED a major bug introduced by the batch mode.

hpcopr addc.

This command was corrupted after introducing the batch mode. We fixed it in this minor version. Please update immediately by running installer update.

2. Make AWS EC2 hyperthreading ON by default.

HT on is the default option for all cloud services. Previously, the without explicitly specifying --ht ON, the default option was --ht OFF.

However, turning off the HT may cause potential problems expecially when reconfiguring the compute nodes - it usually needs to rebuild the nodes, not replace them.

In this version, we changed this to HT on.

3. Optimized the command logic

Now, for delc, shutdownc, turnonc, we optimized the logic and finally eliminated the command flag --htoff. Now the command to control HT is unified to --ht HT_FLAG.

4. Command update/deprecation

The flag --htoff is now deprecated. Previously: hpcopr reconfc --conf a4c8g --htoff . Now: hpcopr reconfc --conf a4c8g --ht OFF

The flag --all for the command delc shutdownc turnonc is deprecated. Previously: hpcopr delc --all . Now: hpcopr delc --nn all

Please try the new version out.

We also added a CONTRIBUTING.md to this project, looking forward to your feedback and contributions!

HPC-NOW: Start your HPC journey in the cloud now.

Assets 5

1 Join discussion

30 Nov 07:26

zhenrong-wang

0.2.1.beta9

70fe623

[BETA] Release version v0.2.1 evolved to 0.2.1.0009!

This is a major evolution of Version 0.2.1.0007.

Main changes:

1. Added 2 important command keywords to control the running of tf.

--dbg-devel DEBUG_INFO_LEVEL.

With this option, you can specify the verbosity of tf debug log.

Options (from most verbose to least verbose): trace, debug, info, warn, err, off

--max-time MAXIMUM_WAIT_TIME.

With this option, you can define the maximum wait time for tf execution. You can specify a number in the range 600 ~ 1200

2. Now you can view the tf debug log with the hpcopr viewlog command.

Command reference: hpcopr viewlog -b --log dbg

3. Updated the now-server module with the recently-introduced batch mode.

Please run the server by ./my-server.ext --client-io PORT_NUMBER. Then the server will be standby to accept requests on the specified port number. If no port number specified, the server will use default port #19885.

4. Fixed known bugs.

Please check the source codes and the packages attached.

5. Command update/deprecation

The command hpcopr viewlog got updated. Previously, there are only 2 streams of log: std (standard output), err (standard err output). Because we introduced the debug log (dbg), now you need to specify a stream by the --log STREAM_TYPE command keyword. Example:

Previous command (Deprecated): hpcopr viewlog -b --err
New command: hpcopr viewlog -b --log err

We also added a CONTRIBUTING.md to this project, looking forward to your feedback and contributions!

HPC-NOW: Start your HPC journey in the cloud now.

Assets 5