Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add sphand ros #2707

Merged
merged 36 commits into from
May 14, 2020
Merged

Add sphand ros #2707

merged 36 commits into from
May 14, 2020

Conversation

pazeshun
Copy link
Collaborator

Add updated version of https://github.com/pazeshun/sphand_ros
Those packages are required for my demo.

@wkentaro
Copy link
Member

wkentaro commented Jan 23, 2020

Are there some specific reasons why you don’t add your sphand_ros to the rosinstall?

@pazeshun
Copy link
Collaborator Author

Are there some specific reasons why you don’t add your sphand_ros to the rosinstall?

I just want to place my code under a repository of our lab, not to let my code forgotten.
This placing also makes maintenance easy, because sphand_ros packages depend on jsk_apc packages.
Someone who changes jsk_apc drastically may adjust sphand_ros if it is included in jsk_apc.

@wkentaro
Copy link
Member

wkentaro commented Jan 24, 2020 via email

@pazeshun pazeshun changed the title Add sphand ros [WIP] Add sphand ros Feb 8, 2020
@knorth55 knorth55 added this to the 4.3.0 milestone Apr 1, 2020
@pazeshun pazeshun changed the title [WIP] Add sphand ros Add sphand ros Apr 28, 2020
@pazeshun pazeshun closed this May 8, 2020
@pazeshun pazeshun reopened this May 8, 2020
@pazeshun
Copy link
Collaborator Author

pazeshun commented May 9, 2020

Travis got timeout in cache setting:

fetching PR.2707/cache--linux-trusty-7471d3ccbbdeaec95cea57133066e84e8dedc6006c1eff07012e6f7e0064ef76--compiler-gpp.tgz

fetching PR.2707/cache-linux-trusty-7471d3ccbbdeaec95cea57133066e84e8dedc6006c1eff07012e6f7e0064ef76--compiler-gpp.tgz

fetching PR.2707/cache--compiler-gpp.tgz

fetching master/cache--linux-trusty-7471d3ccbbdeaec95cea57133066e84e8dedc6006c1eff07012e6f7e0064ef76--compiler-gpp.tgz

found cache

0.00sadding /home/travis/.ccache to cache

adding /home/travis/.cache/pip to cache

adding /home/travis/.ros/data to cache

creating directory /home/travis/.ros/data

No output has been received in the last 10m0s, this potentially indicates a stalled build or something wrong with the build itself.

Check the details on how to adjust your build configuration on: https://docs.travis-ci.com/user/common-build-problems/#build-times-out-because-no-output-was-received

The build has been terminated

I assume the master cache is too large and it cannot be expanded in the time frame.
https://travis-ci.community/t/builds-timeout-during-or-shortly-after-creating-directory-home-travis-cache-sccache/5001
Screenshot from 2020-05-09 14-04-37
I'll try deleting the master cache. If that trial succeeded, we have to consider changing the cache setting or commenting out allow_failures job to reduce cache size.

@pazeshun pazeshun changed the title Add sphand ros [WIP] Add sphand ros May 9, 2020
@pazeshun
Copy link
Collaborator Author

pazeshun commented May 9, 2020

I'll try deleting the master cache.

Travis becomes passed after this deletion, so my assumption (travis failed as cache is too large) may be correct.
And cache of this PR is 22372.95MB, which is more than the previous master cache and seems too heavy:
Screenshot from 2020-05-09 17-32-00
I'll try to reduce this.

pazeshun added a commit to pazeshun/jsk_apc that referenced this pull request May 9, 2020
@pazeshun
Copy link
Collaborator Author

pazeshun commented May 9, 2020

Reached to download limit. Now waiting.

$ gdown https://drive.google.com/uc?id=0B9P1L--7Wd2vS1VaWWVFNDVFQ1k
/usr/local/lib/python2.7/dist-packages/requests/__init__.py:83: RequestsDependencyWarning: Old version of cryptography ([1, 2, 3]) may cause slowdown.
  warnings.warn(warning, RequestsDependencyWarning)
Access denied with the following error:

 	Too many users have viewed or downloaded this file recently. Please
	try accessing the file again later. If the file you are trying to
	access is particularly large or is shared with many people, it may
	take up to 24 hours to be able to view or download the file. If you
	still can't access a file after 24 hours, contact your domain
	administrator. 

You may still be able to access the file from the browser:

	 https://drive.google.com/uc?id=0B9P1L--7Wd2vS1VaWWVFNDVFQ1k 

@knorth55
Copy link
Member

knorth55 commented May 9, 2020

can you try pip install gdown="3.11.0"?
i heard new gdown suppress the error.

@pazeshun
Copy link
Collaborator Author

pazeshun commented May 9, 2020

can you try pip install gdown="3.11.0"?

I tried, but faced the same error.

@pazeshun
Copy link
Collaborator Author

pazeshun commented May 9, 2020

Currently, there seems no solution: wkentaro/gdown#26 (comment)
Perhaps there is a tricky solution, but I don't want to spend my time on finding it.

@wkentaro
Copy link
Member

wkentaro commented May 9, 2020

It seems this file is downloaded extremely many times before, so I can't access the file even from the browser without logging in. (incognito mode on Chrome)
Screen Shot 2020-05-09 at 15 47 56

@wkentaro
Copy link
Member

wkentaro commented May 9, 2020

Currently most of the large files in arc2017_competition directory (in which the original file is located) are not downloadable. (so maybe the drive directory (arc2017_competition) may be in a tricky condition)

@pazeshun
Copy link
Collaborator Author

pazeshun commented May 11, 2020

commenting out allow_failures job to reduce cache size

This is tried at 98c558f and cache size is reduced to 18304.37MB:
Screenshot from 2020-05-11 10-02-18
Now I'll restart travis and check if current cache can be expanded.

@wkentaro
Copy link
Member

Now it's back to normal, so we can see the caching helps to keep the file downloadable for a longer period.

% gdown  https://drive.google.com/uc?id=0B9P1L--7Wd2vS1VaWWVFNDVFQ1k                                                                                                                                                                  
Downloading...                                                                                                                                                                                                                        
From: https://drive.google.com/uc?id=0B9P1L--7Wd2vS1VaWWVFNDVFQ1k                                                                                                                                                                     
To: /home/wkentaro/fcn32s_arc2017_datasetv2_cfg003_20170612.npz                                                                                                                                                                       
52.0MB [00:05, 9.18MB/s]

BTW, is file caching enabled for ~/.ros (or something similar) directory? If it's enabled, I don't think this problem won't happen that frequently.

@wkentaro
Copy link
Member

It was enabled

- $HOME/.ros/data

But I found that the file in the issue is not used in the system because datasetv3 is used instead of v2. So probably we can just comment it out.

symlink_src = os.path.join(
PKG_PATH, 'data/models/fcn32s_arc2017_datasetv3_cfg009_20170724.npz')
symlink_dst = os.path.join(PKG_PATH, 'data/models/fcn32s.npz')
if os.path.exists(symlink_dst):
print('[%s] File already exists, so skipping.' % symlink_dst)
else:
if os.path.islink(symlink_dst):
os.remove(symlink_dst)
print('[%s] Creating symlink to: %s' % (symlink_dst, symlink_src))
os.symlink(symlink_src, symlink_dst)

@pazeshun
Copy link
Collaborator Author

pazeshun commented May 12, 2020

Now I'll restart travis and check if current cache can be expanded.

This trial failed, and I realized that the problem is not about the whole cache size.
And further debugging on https://github.com/pazeshun/jsk_apc/tree/add-sphand_ros-cache told me:

  • On both kinetic test and indigo test, the following files exist under $HOME/.ros/data and cached:
/home/travis/.ros/data/jsk_arc2017_common/fcn32s_arc2017_datasetv2_cfg003_20170612.npz
/home/travis/.ros/data/jsk_arc2017_common/fcn32s_arc2017_datasetv2_cfg003_20170721.npz
/home/travis/.ros/data/jsk_arc2017_common/fcn32s_arc2017_datasetv3_cfg009_20170724.npz
/home/travis/.ros/data/jsk_2016_01_baxter_apc/sib_right_softkinetic.bag.tar.gz
/home/travis/.ros/data/jsk_2016_01_baxter_apc/sib_kinect2.bag.tar.gz
/home/travis/.ros/data/jsk_2016_01_baxter_apc/sib_right_softkinetic.bag
/home/travis/.ros/data/jsk_2016_01_baxter_apc/sib_kinect2.bag
/home/travis/.ros/data/jsk_2016_01_baxter_apc/2016-04-30-16-33-54_apc2016-bin-boxes.bag
/home/travis/.ros/data/jsk_apc2016_common/vgg16_rotation_translation_brightness_372000.chainermodel
/home/travis/.ros/data/jsk_apc2016_common/tabletop_objects.bag
/home/travis/.ros/data/jsk_apc2016_common/vgg16_96000.chainermodel
/home/travis/.ros/data/jsk_apc2016_common/tabletop_objects.bag.tgz
/home/travis/.ros/data/jsk_apc2016_common/fcn32s_v2_148000.chainermodel
/home/travis/.ros/data/jsk_apc2016_common/fcn32s_6000.chainermodel
  • Only on kinetic test, extracting /home/travis/.ros/data/jsk_arc2017_common doesn't finish (even though this directory is the same as indigo test and is lighter than /home/travis/.ros/data/jsk_apc2016_common (I confirmed file size on local) ):
fetching add-sphand_ros-cache/cache--linux-trusty-8defbc9ec54de23ff9c8cc77fb06464a913b1a901ef3ad077d9f613bfc5f48ae--compiler-gpp.tgz
found cache
adding /home/travis/.ccache to cache
adding /home/travis/.cache/pip to cache
adding /home/travis/.ros/data/jsk_2016_01_baxter_apc to cache
creating directory /home/travis/.ros/data/jsk_2016_01_baxter_apc
adding /home/travis/.ros/data/jsk_apc2016_common to cache
creating directory /home/travis/.ros/data/jsk_apc2016_common
adding /home/travis/.ros/data/jsk_arc2017_common to cache
creating directory /home/travis/.ros/data/jsk_arc2017_common

No output has been received in the last 10m0s, this potentially indicates a stalled build or something wrong with the build itself.
Check the details on how to adjust your build configuration on: https://docs.travis-ci.com/user/common-build-problems/#build-times-out-because-no-output-was-received
The build has been terminated

@wkentaro Do you mean we can comment out here and here?
Then this will be the best solution I know.

@wkentaro
Copy link
Member

Sent a PR #2712

@pazeshun pazeshun force-pushed the add-sphand_ros branch 3 times, most recently from c899bf5 to dcc34a4 Compare May 14, 2020 01:55
…ure on travis due to googledrive access limit

I tried travis caching of those data, but cache couldn't be expanded before travis timeout:
https://travis-ci.org/github/start-jsk/jsk_apc/jobs/686632934
https://travis-ci.org/github/start-jsk/jsk_apc/jobs/686826882
@pazeshun pazeshun changed the title [WIP] Add sphand ros Add sphand ros May 14, 2020
@pazeshun
Copy link
Collaborator Author

Changes to existing code by this PR are 6f7e1d9 and 40c5cea.
6f7e1d9 simply adds new packages added by this PR to indigo test.
40c5cea changes grasp_fusion package not to download heavy data from GoogleDrive on travis for avoiding access limit and build failure.
(As described in the commit message, travis cache doesn't work in this case.)
@knorth55 Could you review this?

@knorth55
Copy link
Member

@pazeshun Good work! Thank you!

@knorth55 knorth55 self-requested a review May 14, 2020 15:49
@knorth55 knorth55 merged commit 129f6cc into start-jsk:master May 14, 2020
@pazeshun
Copy link
Collaborator Author

Thank you!

@pazeshun pazeshun deleted the add-sphand_ros branch May 14, 2020 23:19
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants