Skip to content

Update checksums for R 3.6.0 packages and add patch for ppc#11088

Merged
boegel merged 2 commits intoeasybuilders:developfrom
Flamefire:20200811121519_new_pr_R360
Sep 1, 2020
Merged

Update checksums for R 3.6.0 packages and add patch for ppc#11088
boegel merged 2 commits intoeasybuilders:developfrom
Flamefire:20200811121519_new_pr_R360

Conversation

@Flamefire
Copy link
Copy Markdown
Contributor

(created using eb --new-pr)

Comment thread easybuild/easyconfigs/r/R/R-3.6.0-fosscuda-2019a.eb Outdated
@boegel boegel added the bug fix label Aug 12, 2020
@boegel boegel added this to the next release (4.2.3?) milestone Aug 12, 2020
@Flamefire Flamefire force-pushed the 20200811121519_new_pr_R360 branch from 3740f78 to 37b8044 Compare August 12, 2020 11:08
@Flamefire Flamefire changed the title Update checksums for R 3.6.0 packages Update checksums for R 3.6.0 packages and add patch for ppc Aug 12, 2020
@Flamefire
Copy link
Copy Markdown
Contributor Author

I also added the patch for liquidSVM which makes it build on PPC (hope it still applies)

@boegel
Copy link
Copy Markdown
Member

boegel commented Aug 13, 2020

@Flamefire Is there a test report coming for ppc?

@Flamefire
Copy link
Copy Markdown
Contributor Author

Makes sense, yes. Just submitted a job to test that. Gonna take a whole though, so maybe in 10 hrs :/

@boegel
Copy link
Copy Markdown
Member

boegel commented Aug 16, 2020

Test report by @boegel
SUCCESS
Build succeeded for 1 out of 1 (1 easyconfigs in this PR)
node3300.joltik.os - Linux centos linux 7.8.2003, x86_64, Intel(R) Xeon(R) Gold 6242 CPU @ 2.80GHz (cascadelake), Python 3.6.8
See https://gist.github.com/362457742f53521870b101eb6518ce42 for a full test report.

@boegel
Copy link
Copy Markdown
Member

boegel commented Aug 16, 2020

Test report by @boegel
SUCCESS
Build succeeded for 2 out of 2 (2 easyconfigs in this PR)
node2612.swalot.os - Linux centos linux 7.8.2003, x86_64, Intel(R) Xeon(R) CPU E5-2660 v3 @ 2.60GHz (haswell), Python 2.7.5
See https://gist.github.com/9d1242b6122202e25831157ada95b141 for a full test report.

@boegel
Copy link
Copy Markdown
Member

boegel commented Aug 16, 2020

Test report by @boegel
SUCCESS
Build succeeded for 2 out of 2 (2 easyconfigs in this PR)
node3411.kirlia.os - Linux centos linux 7.8.2003, x86_64, Intel(R) Xeon(R) Gold 6240 CPU @ 2.60GHz (cascadelake), Python 2.7.5
See https://gist.github.com/7b1ab67eee8348d048b4227c03e58dc4 for a full test report.

@boegel
Copy link
Copy Markdown
Member

boegel commented Aug 17, 2020

Looking good on x86_64... Any word on a test report from PPC @Flamefire?

If it's too much trouble I'll just go ahead and merge this, I trust you did the right thing there (and it's the same patch for liquidSVM we already have in other R easyconfigs)

@Flamefire
Copy link
Copy Markdown
Contributor Author

I have no idea what went wrong with the job, I assumed it would have already submitted here but now I can't even see the job on the cluster anymore. Restarting...

@Flamefire
Copy link
Copy Markdown
Contributor Author

Test report by @Flamefire
FAILED
Build succeeded for 6 out of 8 (2 easyconfigs in this PR)
taurusml3 - Linux RHEL 7.6, POWER, 8335-GTX, Python 2.7.5
See https://gist.github.com/b9714f54eebeb7d207b2446663219067 for a full test report.

@Flamefire
Copy link
Copy Markdown
Contributor Author

I'm now running into the same issue I mentioned earlier: Installing viridis fails (likely) due to viridisLite being installed first. No idea how to solve as I blessfully know nothing about R and its installation ;) Maybe there is something else somewhere in the environment which R picks up? No idea how to check that

Message I get is:

'viridis.map' is not an exported object from 'namespace:viridisLite'

The "owner" of viridis says sjmgarnier/viridis#57 (comment)

install viridis and viridis only. Do not install viridisLite first.

But it seems viridis would otherwise download and install viridisLite first, so we need to...

@boegel
Copy link
Copy Markdown
Member

boegel commented Aug 21, 2020

@Flamefire I'm clueless on that too... Why doesn't that issue occur on x86_64?!

@Flamefire
Copy link
Copy Markdown
Contributor Author

As I have no idea why it appears I don't know why it doesn't appear either. I tried to install it on Power till it fails and wanted to manually continue or retry with --skip but that doesn't seem to work: It reinstalls everything. If I could get a shell in the environment where it tries to install viridis at that point as EB would, I could maybe find out more. But as it seems our admins could install it successfully maybe we shouldn't bother...

@boegel
Copy link
Copy Markdown
Member

boegel commented Sep 1, 2020

Since the test installations are working out fine n x86_64, I'm calling this: time to merge.

Thanks for the effort @Flamefire!

@boegel
Copy link
Copy Markdown
Member

boegel commented Sep 1, 2020

Going in, thanks @Flamefire!

@boegel boegel merged commit 3edad73 into easybuilders:develop Sep 1, 2020
@Flamefire Flamefire deleted the 20200811121519_new_pr_R360 branch September 2, 2020 07:41
@Flamefire
Copy link
Copy Markdown
Contributor Author

Ok problem found: On Power we have a Lustre mounted over NFS and apparently the install creates and writes to a file but a subsequent getdents on the folder the file was created (and chmoded) in does not return it.
Workaround: Install in /tmp or similar for testing

Relevant part of strace:

[pid 74646] open("/software/R/3.6.0-fosscuda-2019a/lib64/R/library/00LOCK-viridisLite/00new/viridisLite/data/viridis.map.rda", O_WRONLY|O_CREAT|O_TRUNC, 0666) = 4
[pid 74646] fstat(3, {st_mode=S_IFREG|0644, st_size=27832, ...}) = 0
[pid 74646] mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2002cb230000
[pid 74646] read(3, "BZh91AY&SY>\232Q\6\0,\205\177\377\377\377\377\377\377\377\377\377\377\377\377\377\377"..., 8192) = 8192
[pid 74646] fstat(4, {st_mode=S_IFREG|0644, st_size=0, ...}) = 0
[pid 74646] mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x2002cb240000
[pid 74646] write(4, "BZh91AY&SY>\232Q\6\0,\205\177\377\377\377\377\377\377\377\377\377\377\377\377\377\377"..., 8192) = 8192
[pid 74646] read(3, "L\352O\317\317\354\321\33\343mq\224\330\213<\317$1U\33+4z\377\301<\307ZSN\245\306"..., 8192) = 8192
[pid 74646] read(3, "?\300v\360\374\243(\205\3068\375\376\274J\335\3729\217T\250a)\226\340\325M_\275/\340\337^"..., 8192) = 8192
[pid 74646] write(4, "L\352O\317\317\354\321\33\343mq\224\330\213<\317$1U\33+4z\377\301<\307ZSN\245\306"..., 8192) = 8192
[pid 74646] write(4, "?\300v\360\374\243(\205\3068\375\376\274J\335\3729\217T\250a)\226\340\325M_\275/\340\337^"..., 8192) = 8192
[pid 74646] read(3, "\224W\2023\303\310\365\341\2412\365\370\31\251\327\200\201\377-\273`M-\235\331\231,\n\341\310L\21"..., 8192) = 3256
[pid 74646] read(3, "", 8192)           = 0
[pid 74646] write(4, "\224W\2023\303\310\365\341\2412\365\370\31\251\327\200\201\377-\273`M-\235\331\231,\n\341\310L\21"..., 3256) = 3256
[pid 74646] close(4)                    = 0
[pid 74646] munmap(0x2002cb240000, 8192) = 0
[pid 74646] chmod("/software/R/3.6.0-fosscuda-2019a/lib64/R/library/00LOCK-viridisLite/00new/viridisLite/data/viridis.map.rda", 0644) = 0
[pid 74646] close(3)                    = 0
[pid 74646] munmap(0x2002cb230000, 8192) = 0
[pid 74646] openat(AT_FDCWD, "/software/R/3.6.0-fosscuda-2019a/lib64/R/library/00LOCK-viridisLite/00new/viridisLite/data", O_RDONLY|O_NONBLOCK|O_DIRECTORY|O_CLOEXEC) = 3
[pid 74646] getdents(3, /* 2 entries */, 32768) = 48
[pid 74646] getdents(3, /* 0 entries */, 32768) = 0

The 2nd last line should show "3 entries" as it does on x86 or /tmp

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants