Copied UKCA version not running on ARCHER2

Hi there, I copied a working UKCA version on Archer2 (u-dd742), but it did not run (with RosePopenError). It seems due to a small bug, but I do not know how to find it, could you help? Thanks.

Hi Xin,

There should be more information about the error in the pop up window. You may need to expand it.

Otherwise in the ~/roses/u-dd742 please run:

rose suite-run

and you should see an error message in the terminal window.

Regards,
Ros.

Hi Ros,

Thanks for the help.
I expanded the error message; it seems that the major error is “no hosts selected” (see attached screenshots). I have tried using different HPC account groups (no2-surfeit, n02-bas, n02) but they all failed.
Another thing is that “emxin” is my HPC account and used for HPC tasks (under suite conf), but I used “xinyang” as “owner” under “suite info”. Do they cause confliction?

Best,
Xin

This email and any attachments are intended solely for the use of the named recipients. If you are not the intended recipient you must not use, disclose, copy or distribute this email or any of its attachments and should notify the sender immediately and delete this email from your system. UK Research and Innovation (UKRI) has taken every reasonable precaution to minimise risk of this email or any attachments containing viruses or malware but the recipient should carry out its own virus and malware checks before opening the attachments. UKRI does not accept any liability for any losses or damages which the recipient may sustain due to presence of any viruses.

Hi Xin,

The owner variable has nothing to do with suite submission - that’s just suite discovery metatdata.

On the puma2 command line please try running:

rose host-select archer2

assuming that doesn’t work try

ssh ln01

And also run ssh-add -l

and post me the output of all those please.

From the message above it looks like you ssh connection from puma2 to archer2 is broken. Have you run a suite from puma2 before or is this the first time?

Regards,
Ros.

Hi Ros,

Thanks! Yes, this is my first time to run suite on puma2. See attached screenshot for the commands. When I run ssh ln01, it asked for PW input, but I forgot which PW I used before (sorry I forgot what I did). Please let me know if I need to re-set the PW.

Best,
Xin

This email and any attachments are intended solely for the use of the named recipients. If you are not the intended recipient you must not use, disclose, copy or distribute this email or any of its attachments and should notify the sender immediately and delete this email from your system. UK Research and Innovation (UKRI) has taken every reasonable precaution to minimise risk of this email or any attachments containing viruses or malware but the recipient should carry out its own virus and malware checks before opening the attachments. UKRI does not accept any liability for any losses or damages which the recipient may sustain due to presence of any viruses.

Hi Xin,

Please follow the puma2 setup instructions at: Transition to PUMA2 to setup you puma2 environment and connection to/from archer2.

The PW is your ARCHER2 password.

Regards,
Ros

Hi Ros,

Sorry, I think I passed a wrong message in my previous emial regarding 'ssh-add -l". the correct message is:

2048 SHA256:0akYBC5l9gF2Dw/5Daua6KwYgj0mxn0467+VsEphJLg /home/n02/n02/emxin/.ssh/id_rsa_puma (RSA)

Does it means something differently?

Below are a few other things:

  1. I am confused by the “ARCHER2 password”. When I ssh to my archer2 account, I was asked to enter “passphrase for key” and then the 6-digit number generated by my mobile authentication app, they worked well. Is the “passphrase for key” the PW needed for “ssh ln01” from puma? Since this passphrase did not work, thus I assume there must be a different PW, but I cannot recall where I entered this PW for ARCHER2 login, and please let know how to get it back.

  2. from the webpage instruction step 3: “Once you are on the ARCHER2 login nodes, type ssh -Y puma2. Enter your ARCHER2 password when prompted.”
    I could successfully log into puma2 from archer2 without entering any PW. This means that I have already set up passwordless access to puma2 when I moved puma files to puma2, but I forgot the PW I used.

  3. from instruction step 7: "You should be prompted for your MOSRS password, then username. (Note that it asks for your password first.) If the password caching works, you should see:


Subversion password cached
Rosie password cached"

However, I was not asked to input PW and username.

I believed that I had done something re environment setup, but do I need to do everything again?
Maybe I could just update some missing parts? Please help!

Best,
Xin

Best,
Xin

This email and any attachments are intended solely for the use of the named recipients. If you are not the intended recipient you must not use, disclose, copy or distribute this email or any of its attachments and should notify the sender immediately and delete this email from your system. UK Research and Innovation (UKRI) has taken every reasonable precaution to minimise risk of this email or any attachments containing viruses or malware but the recipient should carry out its own virus and malware checks before opening the attachments. UKRI does not accept any liability for any losses or damages which the recipient may sustain due to presence of any viruses.

Hi Xin,

The ssh-add -l output indicates that you have added your puma key to the ssh-agent but not the archer2 one which is what is needed to get from puma2 → archer2 without being prompted for your ARCHER2 password.

I don’t know what previous setup you had already done so it’s very hard for me to advise which bits you do or don’t need to do in the instructions. Possibly only section 7 (i),(iii) & (iv) which setups up to get from puma2 → archer2.

In answer to your questions:

  1. ARCHER2 password is the password you used to use to login to ARCHER2 before they moved to the 6-digit number. If you have forgotten that password and if you need it again in the future then you will need to go into ARCHER2 SAFE and reset the password. Go through section 7 first - you shouldn’t need that password for setting up the ssh-agent from puma2 → archer2.

  2. If you can get to puma2 from archer2 without password that is all set up fine. Nothing to do there.

  3. Again it looks like you already had your MOSRS access set up so nothing to do there either.

Regards,
Ros

Hi Ros,

Thanks for the help. The suite is now working!

Best,
Xin

This email and any attachments are intended solely for the use of the named recipients. If you are not the intended recipient you must not use, disclose, copy or distribute this email or any of its attachments and should notify the sender immediately and delete this email from your system. UK Research and Innovation (UKRI) has taken every reasonable precaution to minimise risk of this email or any attachments containing viruses or malware but the recipient should carry out its own virus and malware checks before opening the attachments. UKRI does not accept any liability for any losses or damages which the recipient may sustain due to presence of any viruses.

Hi Ros,

I re-generated the ssh Key for archer2 and followed the instruction to re-do the setups, but the suite job crashed, see the error message shown in the attached screenshot.

FYI: when run “rose host-select archer2”, it gave “ln02” without waning indicating it works well.

“ssh ln01” seems also give the right answer (see attached screenshot.

However, “ssh-add -l” gave an error message:

Could not open a connection to your authentication agent.

Please help!

Best,
Xin

This email and any attachments are intended solely for the use of the named recipients. If you are not the intended recipient you must not use, disclose, copy or distribute this email or any of its attachments and should notify the sender immediately and delete this email from your system. UK Research and Innovation (UKRI) has taken every reasonable precaution to minimise risk of this email or any attachments containing viruses or malware but the recipient should carry out its own virus and malware checks before opening the attachments. UKRI does not accept any liability for any losses or damages which the recipient may sustain due to presence of any viruses.

Hi Xin,

Your setup from puma2 to archer2 is fine. You don’t need to re-do anything.
You ran ssh-add -l on archer2 which isn’t running ssh-agent, if you run it on puma2 it will list your keys fine.

In your previous message you said the suite is working fine.
Once a suite is running you can’t re-run it without stopping it first. The rose suite-run message is just saying u-dd742 is already running.

On puma2 in the ~/roses/u-dd742 directory run rose sgc to bring back up the cylc GUI to see the status of your running suite.

Regards,
Ros.

Hi Ros,

Thank you. Yes, now ssh-add -l gave the right answer on puma2.

In addition, job u-dd742 was submitted and hopefully it will run smoothly.

One thing that looks strange is: after I re-set up ssh keys, I have to input PW and passphrase every time when I try to “ssh puma2” from archer2. I remember I typed the “enter” key as PW when I re-generated the ssh key for puma2. Anyhtig that I can do to re-check whether the key was generated OK?

Best,
Xin

This email and any attachments are intended solely for the use of the named recipients. If you are not the intended recipient you must not use, disclose, copy or distribute this email or any of its attachments and should notify the sender immediately and delete this email from your system. UK Research and Innovation (UKRI) has taken every reasonable precaution to minimise risk of this email or any attachments containing viruses or malware but the recipient should carry out its own virus and malware checks before opening the attachments. UKRI does not accept any liability for any losses or damages which the recipient may sustain due to presence of any viruses.

Hi Xin,

Make sure the ~/.ssh/config on ARCHER2 is pointing to the correct ssh-key. And that the public part has been correctly copied to ~/.ssh/authorized_keys on puma2.

If you can’t see what’s wrong, the easiest thing to do is to redo Section 5 “setup passwordless access to puma2”. If you do redo it then I would suggest moving the ~/.ssh/authorized_keys file on puma2 out of the way first as I can see you have so many versions of the ssh-key listed in there that could cause issues.

Regards,
Ros.

Hi Ros,

Yes, gave a new name of the authorized_keys file on puma and re-did the ssh key for puma2 connection following instruction in section 5. However, the problem remains. I attached wo screenshts for you to check, e.g., one message is

Load key “/home/n02/n02/emxin/.ssh/id_rsa_puma2”: bad permissions.

does it means something went wrong?

Another issue I am not sure is in the config file on puma2, I simply put below line at the end of the file, which may be a problem ?

ssh-add ~/.ssh/id_rsa

Best,

Xin

This email and any attachments are intended solely for the use of the named recipients. If you are not the intended recipient you must not use, disclose, copy or distribute this email or any of its attachments and should notify the sender immediately and delete this email from your system. UK Research and Innovation (UKRI) has taken every reasonable precaution to minimise risk of this email or any attachments containing viruses or malware but the recipient should carry out its own virus and malware checks before opening the attachments. UKRI does not accept any liability for any losses or damages which the recipient may sustain due to presence of any viruses.

Hi Xin,

The error message says the permissions on your ARCHER2 ~/.ssh directory are too open. The same is true on PUMA2. Your ssh keys must not be readable by others.

On ARCHER2 and PUMA2 run:

chmod 600 ~/.ssh

And then try logging into puma2 again.

Regards,
Ros.

Hi Ros,

Thanks! after I chmod to 600 for ~/.ssh on puma and archer, the old error message gone, showing a new error message:

-bash: /home/n02/n02/emxin/.ssh/ssh-setup: Permission denied

it seems this time it is related to the ssh-setup file. Could you please take a look and let me know what I should do. BTW, I still need to enter both PW and passphrase to log into puma2 from archer2.

Best,

Xin

This email and any attachments are intended solely for the use of the named recipients. If you are not the intended recipient you must not use, disclose, copy or distribute this email or any of its attachments and should notify the sender immediately and delete this email from your system. UK Research and Innovation (UKRI) has taken every reasonable precaution to minimise risk of this email or any attachments containing viruses or malware but the recipient should carry out its own virus and malware checks before opening the attachments. UKRI does not accept any liability for any losses or damages which the recipient may sustain due to presence of any viruses.

Sorry I miss-typed. Should have said:

chmod 700 ~/.ssh

Cheers,
Ros.

Hi Ros,

No worries. After I changed ~/.ssh and the files under /.ssh (on puma and archer ) to 700, things seem getting better, as I was only asked to input passphrase (no PW intput):

Last login: Thu Feb 22 17:12:55 2024 from 172.24.75.51

Enter passphrase for /home/n02/n02/emxin/.ssh/id_rsa:

Identity added: /home/n02/n02/emxin/.ssh/id_rsa (/home/n02/n02/emxin/.ssh/id_rsa)

It seems something is still not working well.

Best,

Xin

This email and any attachments are intended solely for the use of the named recipients. If you are not the intended recipient you must not use, disclose, copy or distribute this email or any of its attachments and should notify the sender immediately and delete this email from your system. UK Research and Innovation (UKRI) has taken every reasonable precaution to minimise risk of this email or any attachments containing viruses or malware but the recipient should carry out its own virus and malware checks before opening the attachments. UKRI does not accept any liability for any losses or damages which the recipient may sustain due to presence of any viruses.

Hi Xin,

That’s logging into which machine from which machine?
Can you please always include terminal prompt & command you ran so we can see.

Regards,
Ros.

Hi Ros,

Sorry about it. I now attach a screenshot for easy read. It was from archer2 to puma2. I was asked to enter passphrase to log in.

Best,
Xin

This email and any attachments are intended solely for the use of the named recipients. If you are not the intended recipient you must not use, disclose, copy or distribute this email or any of its attachments and should notify the sender immediately and delete this email from your system. UK Research and Innovation (UKRI) has taken every reasonable precaution to minimise risk of this email or any attachments containing viruses or malware but the recipient should carry out its own virus and malware checks before opening the attachments. UKRI does not accept any liability for any losses or damages which the recipient may sustain due to presence of any viruses.