h2o-release s3 bucket files gone

[sshenoy-mdsol]

H2O version, Operating System and Environment
3.44.0
Ubuntu 22.04.5 LTS-R version 4.1.3

Actual behavior
h2o R package fails to install due to missing s3 file: https://s3.amazonaws.com/h2o-release/h2o/rel-3.44.0/3/Rjar/h2o.jar

Expected behavior
Ability to install h2o R package

Error: Error: Error installing package 'h2o':

• installing source package ‘h2o’ ...
  ** package ‘h2o’ successfully unpacked and MD5 sums checked
  ** using staged installation
  ** R
  ** demo
  ** inst
  ** byte-compile and prepare package for lazy loading
  Performing one-time download of h2o.jar from
  https://s3.amazonaws.com/h2o-release/h2o/rel-3.44.0/3/Rjar/h2o.jar
  (This could take a few minutes, please be patient...)
  Warning in download.file(url = h2o_url, destfile = temp_file, mode = "wb", :
  cannot open URL 'https://s3.amazonaws.com/h2o-release/h2o/rel-3.44.0/3/Rjar/h2o.jar': HTTP status was '404 Not Found'
  Error in download.file(url = h2o_url, destfile = temp_file, mode = "wb", :
  cannot open URL 'https://s3.amazonaws.com/h2o-release/h2o/rel-3.44.0/3/Rjar/h2o.jar'
  Error: Error: unable to load R code in package ‘h2o’
  Execution halted
  ERROR: lazy loading failed for package ‘h2o’

[sshenoy-mdsol]

I assume this goes beyond this specific version of h2o since all links to the download page on https://docs.h2o.ai/h2o/latest-stable/h2o-docs/downloading.html are broken

https://h2o-release.s3.amazonaws.com/h2o/latest_stable.html

<Error>
<Code>NoSuchKey</Code>
<Message>The specified key does not exist.</Message>
<Key>h2o/latest_stable.html</Key>
<RequestId>RFQMW0HRAAHTMANC</RequestId>
<HostId>nykE99sOuBdi/vDh7oyBMIl/O6pXeq4pGJkNy7w9VdL7yJIOHjiN2U35QeBy9H6D6L6+cJ6vlDw=</HostId>
</Error>

[RoelVerbelen]

Facing same issue.

[moravveji]

I hit the same problem while installing the R-bundle-CRAN module on our cluster, using EasyBuild. My remedy was to update the EasyConfig file for pulling the h2o source to the following block, to make sure the source bundle is directly pulled from CRAN:

    ('h2o', '3.44.0.3', {
        'source_urls': ['https://cran.r-project.org/src/contrib/'],
        'sources': ['%(name)s_%(version)s.tar.gz'],
        'checksums': ['61a85f6c2f15e8e96839f8a4fd3a45eaa6bca90517bb20a4dd36e951d6fd0c82'],
    }),

Rebuilding now ... will check in few hours how far I get ;-)

[tomasfryda]

Sorry for the inconvenience. There is some internal migration in h2o. It should work again within couple days.

Until that you should be able to use the following:

> install.packages("h2o", type="source", repos="https://s3.amazonaws.com/h2o-release-migratedv2/h2o/rel-3.44.0/3/R")

For other languages than R, see https://s3.amazonaws.com/h2o-release-migratedv2/h2o/rel-3.44.0/3/index.html

[itsmunishbhardwaj]

TL;DR internal aws migration is causing this issue. Its resolved now. Can you please try again? @sshenoy-mdsol

https://h2oai.slack.com/archives/CT1BQCRV5/p1736742720346759?thread_ts=1733834158.624249&cid=CT1BQCRV5

[tomasfryda]

@itsmunishbhardwaj I tried it just now and it's still an issue.

You can reproduce it by running install.packages("h2o", type="source") in R.

> install.packages("h2o", type="source")
Installing package into ‘/Users/tomasfryda/Library/R/x86_64/4.4/library’
(as ‘lib’ is unspecified)
trying URL 'https://cloud.r-project.org/src/contrib/h2o_3.44.0.3.tar.gz'
Content type 'application/x-gzip' length 665867 bytes (650 KB)
==================================================
downloaded 650 KB

* installing *source* package ‘h2o’ ...
** package ‘h2o’ successfully unpacked and MD5 sums checked
** using staged installation
** R
** demo
** inst
** byte-compile and prepare package for lazy loading
Performing one-time download of h2o.jar from
     https://s3.amazonaws.com/h2o-release/h2o/rel-3.44.0/3/Rjar/h2o.jar
(This could take a few minutes, please be patient...)
Warning in download.file(url = h2o_url, destfile = temp_file, mode = "wb",  :
  cannot open URL 'https://s3.amazonaws.com/h2o-release/h2o/rel-3.44.0/3/Rjar/h2o.jar': HTTP status was '404 Not Found'
Error in download.file(url = h2o_url, destfile = temp_file, mode = "wb",  :
  cannot open URL 'https://s3.amazonaws.com/h2o-release/h2o/rel-3.44.0/3/Rjar/h2o.jar'
Error: unable to load R code in package ‘h2o’
Execution halted
ERROR: lazy loading failed for package ‘h2o’
* removing ‘/Users/tomasfryda/Library/R/x86_64/4.4/library/h2o’
* restoring previous ‘/Users/tomasfryda/Library/R/x86_64/4.4/library/h2o’
Warning in install.packages("h2o", type = "source") :
  installation of package ‘h2o’ had non-zero exit status

The downloaded source packages are in
	‘/private/var/folders/49/kh67vvnj633ftt08t8zfwsvh0000gn/T/RtmpegDIaa/downloaded_packages’

[stamenminkov]

Same issue. Hope it's solved quickly!

[sshenoy-mdsol]

Since the s3 bucket files are still unavailable I used the following workaround:

1. Find the h2o.jar file from an existing installation, e.g. in the R console

> h2o:::.h2o.downloadJar()
[1] "/home/XXX/.cache/R/renv/library/XXX-9d0219ea/R-4.1/x86_64-pc-linux-gnu/h2o/java/h2o.jar"

Alternatively, just download h2o.jar from the bucket mentioned above:
https://s3.amazonaws.com/h2o-release-migratedv2/
e.g.

https://s3.amazonaws.com/h2o-release-migratedv2/h2o/rel-3.44.0/3/Rjar/h2o.jar

2. Wherever you are installing R/h2o, set the env variable H2O_JAR_PATH to point to this file. If you are building a docker image you will need to COPY the file, e.g. in the Dockerfile

COPY h2o.jar /opt/ml/h2o.jar
ENV H2O_JAR_PATH=/opt/ml/h2o.jar

3. Proceed with R package installation steps as normal

Note: it is probably important to make sure you use h2o.jar from the same version of h2o that you are installing.
Ref:

h2o-3/h2o-r/h2o-package/R/connection.R

Line 787 in facf8f4

.h2o.downloadJar <- function(overwrite = FALSE) {