[
https://issues.apache.org/jira/browse/DLAB-1974?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vira Vitanska updated DLAB-1974:
--------------------------------
Description:
1. If lib installation fails due to wrong version convey invalid version
How it works now (for example if version contains space):
- R package: Docker runs with '0', shows 'installation error' and it is retry
- Pip/apt/yum packages: Docker runs with '1' and status is still in
'installing'
(i) removed versions from tmp log file name, for R it shows invalid version in
version contains invalid characters.
(+) Invalid characters are blocked from UI side
----
----
2. Augment error message for library installation (<bossa> from r package).
Available version value is empty
(i) "available_versions": [""] was not empty, now it is []. error message is
installation_error
(+) Now status is 'invalid name'
----
----
3. For EMR pandas/matplotlib are installed but it is shown 'error installation'
due to some dependencies are not installed. This case is reproduced for pip3
group.
(i) now "installation_error" status is shown only if there was any kind of
error during lib installation and lib has not been installed. error message is
returned to backend, but I am note sure whether it will be shown on UI or not.
(should be shown)
(+)
----
----
4. For EMR ggplot2 (r package) is not installed via DLab UI, but it can be
installed via terminal
(i) fixed
(!) Blocked by ticket https://issues.apache.org/jira/browse/DLAB-1995
----
----
5. If user does not type library version and installation fails - convey
'installation error' instead of 'invalid version'. It was found for EMR (others
group) 'pivottable'
(i) previously returned invalid version with available versions = none. Now
returns installation_error with error message
The bug is still reproduced for compute/Notebook. Docker runs with '1' and
status is stuck in 'installing'.
(i) it is shown as 'invalid name' due to absence of any version for this lib
----
----
6. For EMR (it was found on RStudio, perhaps it is reproduced and for the other
template): If install library from apt/yum package and indicate wrong version
docker runs with '1' and library status remains 'installing' in DLab UI
(i) fixed
(+)
----
----
7. Rstudio (others group): Pivottablejs installation with correct version
<0.2.0> was stuck in 'installing' status and docker run with '1'
(i) there was a whitespace in version < 0.2.0>
(+) Now whitespace is blocked from UI side
----
----
8. TensorFlow with RStudio (r package -> NAM ): docker runs with '1' and
installation status is still 'installing' in Web UI (i) fixed
(+)
was:
1. If lib installation fails due to wrong version convey invalid version
How it works now (for example if version contains space):
- R package: Docker runs with '0', shows 'installation error' and it is retry
- Pip/apt/yum packages: Docker runs with '1' and status is still in
'installing'
(i) removed versions from tmp log file name, for R it shows invalid version in
version contains invalid characters.
(+) Invalid characters are blocked from UI side
----
----
2. Augment error message for library installation (<bossa> from r package).
Available version value is empty
(i) "available_versions": [""] was not empty, now it is []. error message is
installation_error
(+) Now status is 'invalid name'
----
----
3. For EMR pandas/matplotlib are installed but it is shown 'error installation'
due to some dependencies are not installed. This case is reproduced for pip3
group.
(i) now "installation_error" status is shown only if there was any kind of
error during lib installation and lib has not been installed. error message is
returned to backend, but I am note sure whether it will be shown on UI or not.
(should be shown)
(+)
----
----
4. For EMR ggplot2 (r package) is not installed via DLab UI, but it can be
installed via terminal
(i) fixed
(!) Blocked by ticket https://issues.apache.org/jira/browse/DLAB-1995
----
----
5. If user does not type library version and installation fails - convey
'installation error' instead of 'invalid version'. It was found for EMR (others
group) 'pivottable'
(i) previously returned invalid version with available versions = none. Now
returns installation_error with error message
The bug is still reproduced for compute/Notebook. Docker runs with '1' and
status is stuck in 'installing'.
(i) it is shown as 'invalid name' due to absence of any version for this lib
----
----
6. For EMR (it was found on RStudio, perhaps it is reproduced and for the other
template): If install library from apt/yum package and indicate wrong version
docker runs with '1' and library status remains 'installing' in DLab UI
(i) fixed
(+)
----
----
7. Rstudio (others group): Pivottablejs installation with correct version
<0.2.0> was stuck in 'installing' status and docker run with '1'
(i) there was a whitespace in version < 0.2.0>
(+) Now whitespace is blocked from UI side
----
----
8. TensorFlow with RStudio (r package -> NAM ): docker runs with '1' and
installation status is still 'installing' in Web UI (i) fixed
(!) Could not check. It is blocked by ticket
https://issues.apache.org/jira/browse/DLAB-1995
> Show 'invalid version' if lib installation fails due to wrong version
> ---------------------------------------------------------------------
>
> Key: DLAB-1974
> URL: https://issues.apache.org/jira/browse/DLAB-1974
> Project: Apache DLab
> Issue Type: Task
> Reporter: Vira Vitanska
> Assignee: Leonid Frolov
> Priority: Minor
> Labels: AWS, AZURE, Debian, DevOps, GCP, RedHat,
> pull-request-available
> Attachments: Augment error message.png, Augment2.png, Available
> version.png, P5 - pivottable 2.png, P6 -docker.png, P6 provisioning.png, P6
> response.png, P7 - docker 2.png, P7 - provisioning.png, P7-docker1.png,
> P7-rstudio UI.png, Pandas1.png, Pandas2.png, Pandas3.png, Pip.png, Point 6
> UI.png, Point8 - Tensorflow-rstudio.png, Point8 TensorFlow-Rstudio3.png,
> Point8- Tensorflow-Rstuio2.png, Reopened - p5-pivottable.png, apt-yum.png,
> ggplot2.png, p5 - Pivottable.png
>
>
> 1. If lib installation fails due to wrong version convey invalid version
> How it works now (for example if version contains space):
> - R package: Docker runs with '0', shows 'installation error' and it is retry
> - Pip/apt/yum packages: Docker runs with '1' and status is still in
> 'installing'
> (i) removed versions from tmp log file name, for R it shows invalid version
> in version contains invalid characters.
> (+) Invalid characters are blocked from UI side
> ----
> ----
> 2. Augment error message for library installation (<bossa> from r package).
> Available version value is empty
> (i) "available_versions": [""] was not empty, now it is []. error message is
> installation_error
> (+) Now status is 'invalid name'
> ----
> ----
> 3. For EMR pandas/matplotlib are installed but it is shown 'error
> installation' due to some dependencies are not installed. This case is
> reproduced for pip3 group.
> (i) now "installation_error" status is shown only if there was any kind of
> error during lib installation and lib has not been installed. error message
> is returned to backend, but I am note sure whether it will be shown on UI or
> not. (should be shown)
> (+)
> ----
> ----
> 4. For EMR ggplot2 (r package) is not installed via DLab UI, but it can be
> installed via terminal
> (i) fixed
> (!) Blocked by ticket https://issues.apache.org/jira/browse/DLAB-1995
> ----
> ----
> 5. If user does not type library version and installation fails - convey
> 'installation error' instead of 'invalid version'. It was found for EMR
> (others group) 'pivottable'
> (i) previously returned invalid version with available versions = none. Now
> returns installation_error with error message
> The bug is still reproduced for compute/Notebook. Docker runs with '1' and
> status is stuck in 'installing'.
> (i) it is shown as 'invalid name' due to absence of any version for this lib
> ----
> ----
> 6. For EMR (it was found on RStudio, perhaps it is reproduced and for the
> other template): If install library from apt/yum package and indicate wrong
> version docker runs with '1' and library status remains 'installing' in DLab
> UI
> (i) fixed
> (+)
> ----
> ----
> 7. Rstudio (others group): Pivottablejs installation with correct version
> <0.2.0> was stuck in 'installing' status and docker run with '1'
> (i) there was a whitespace in version < 0.2.0>
> (+) Now whitespace is blocked from UI side
> ----
> ----
> 8. TensorFlow with RStudio (r package -> NAM ): docker runs with '1' and
> installation status is still 'installing' in Web UI (i) fixed
> (+)
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@dlab.apache.org
For additional commands, e-mail: dev-h...@dlab.apache.org
