​

​

Hello Reader, here is this month’s iRODS news and developments!

If you’re facing an issue with iRODS you’re not sure how to solve, please do drop me a line; if I’ve come across a solution or seen something relevant elsewhere I’ll do my best to let you know. Or just drop me a mail to say ‘Hi’. Always nice to hear from people, particularly in these pandemic times!

I’d love your thoughts and feedback on how this newsletter could be better for you.

News

​iRODS Development Update: December 2020​

​iRODS S3 Resource Plugin v4.2.8.1 is released​

I dont report on the commits to this repo at the moment (thinking of starting on the python one), do say if its something of interest to you. The big thing here, if I’m reading it right, apart from bug fixes and test improvements, is the cachless S3 functionality.

Main Repository Activity

Open Issues

​unixodbc version too old for postgres12​

I’m not up to date on postgreSQL world, but if you or your DBA’s are planning on moving to V12, you might want to add your support to this ticket?

​add atomic api endpoint for catalog operations​

To quote the issue; “this is in service of anticipated use cases for HPC, parallel filesystems, storage tiering, and AWS lambda irods/irods_client_aws_lambda_s3#12” - very intriguing. If I understand correctly, this is setting the stage for bulk atomic operations of operation. Could be huge for batch workloads; roll back a load of files and metadata if some of them fail to upload. At the moment, developers have to detect the failure and retry, and, well, that’s hard and not everyone does.

​imeta command returns 0 exit code on invalid input​

I think the argument goes that as the program didn’t crash, it returns 0. My gut feel is that isn’t right, and there is some kind of GNU or linux standard for abnormal (i.e. non zero) return codes, but, and I may be displaying my Comp Sci ignorance here, I can’t find it. Anyone know better?

​New data object does not follow O_RDWR semantics​

This is vs 4-2-stable so dont panic. Its a really nicely written report though with sample Python iRODSClient code and everything.

​iput using too much memory?​

I keep meaning to try this, see if I get a similar result with Lustre, but no Round Tuit’s so far. The current theory is that it isn’t iRODS though…

​ichksum -aK ignores status: down​

Possibly a subset of all the ways iRODS doesn’t quite interact with marking a resource down that you might not be aware of…

​msi_update_unixfilesystem_resource_free_space being run on incorrect server?​

Distributed systems are hard m’kay?

​pep_api_auth_response_pre can be triggered without _post or _except​

Found whilst attempting to notify users they were using the wrong authentication type for their login, it makes me wonder how many other of the pre, post, except, and finally Policy Enforcement Points dont get triggered, and if there was a test harness that looped through them all in some way to dteermine regression. If not, wnayone want to write one and submit it? ;-)

​ils does not pass auth_scheme from irods environment to agent​

Related to 5201, above. Might be all the icommands?

​Add PUT KW during iput.​

Allows the S3 plugin to avoid cache, if I understand correctly. If you use the S3 plugin, worth reading this, especially if you’re on the fence about upgrading to 4.2.9.

​pam_password_max_time does not allow pass_expiry_ts above 1209600​

Been open for a bit, but other users chiming in. Remember, the developers can’t work on everything, so leting them know which issues are causing the most issues helps them prioritise. That and consortium membership obvs.

​iquest cannot match names containing apostrophes via = operator​

Quoting the gift that keeps on giving…

​Support for Ubuntu 20.04​

Anyone running this in production? Systemd support is as tricky as expected!

​ichksum gives incorrect error when checking replica on downed resource​

Another in the list of all the ways iRODS doesn’t quite interact with marking a resource down. Using passthru/weighted resources is probably a better approach, but I’ve not experimented myself.

​DATA_OBJ_OPEN_AN API request replies with 0 on non-existing file​

This is an interesting issue not because of the issue itself (possible regression between 4.1.x and 4.2.3), but also the use of the PHP API for iRODS and whether it is an archived/defunct project (its gitgub page says so, so certinaely not being actively maintained as its not, to my knowledge, been taken on by the Consortium, unlike the R API). I think the python API is getting more love, assuming you’re not using the C API?

​Non-package install missing some dependencies​

Nice to see the non-package build getting some love.

Closed Issues

Actually, lots of closed pull requests but I dont comment on them, in case you’re wondering!

Sometimes it goes like that!

… Actually, there were a few as I wrote this, but they didn’t make the cut. Next Issue!

If you think someone else would appreciate this newsletter, they can sign up at https://theresource.metadata.school/​

7 Yaks were shaved in the making of this newsletter.

​