Installation file download error

Install, Configure and Update
1

Hello,

I’m trying to re-install Cryosparc on our master node. However, when running the command

curl -L https://get.cryosparc.com/download/master-latest/$LICENSE_ID -o cryosparc_master.tar.gz

I receive the error:

/programs/x86_64-linux/system/curl/current/bin/curl: error while loading shared libraries: liblber-2.4.so.2: cannot open shared object file: No such file or directory

Our admin says, that our servers are using a newer version of this library:

rpm -ql openldap-2.6.6-3.el9.x86_64 | grep liblber
/usr/lib64/liblber.so.2
/usr/lib64/liblber.so.2.0.200

  1. Is there an easy way to fix this (like downloading the files in a different way and proceeding with installation as usual)? Our admin thinks we might have to run CS from a container. If so, is running CS through a container having lower performance?

  2. Is there a recommendation where the master installation should be located? Is it better to install CS on a local drive on the master node?

Thanks for any help!
Stefan

2

What happens if you run the commands

LICENSE_ID="your-license-id" # replace with actual license id
/usr/bin/curl -L https://get.cryosparc.com/download/master-latest/$LICENSE_ID -o cryosparc_master.tar.gz
3

It downloads a file, but with a size of 4Kb

/usr/bin/curl -L https://get.cryosparc.com/download/master-latest/$LICENSE_ID -o cryosparc_master.tar.gz
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
100 42 100 42 0 0 119 0 --:–:-- --:–:-- --:–:-- 119

So the connection looks OK, the error is gone, but the filesize doesn’t match.

I was able to download the files using a different server, having an older OS. Installation with that file worked well.

Should the master package be installed locally on the master node for better performance in the GUI and handling the database (which is also locally installed) or does it not matter?

4

It is possible to install the cryosparc_master/ directory on an nfs share that is mounted on the computer where CryoSPARC master processes will run.
You may want to install and run CryoSPARC under a dedicated Linux account whose environment includes few, if any, customizations on top of the “standard” system environment.
For example, the curl command should be provided by the OS, likely under /usr/bin:

$ which curl
/usr/bin/curl
$

for the selected Linux account.
Some customizations, such as environment changes associated with certain software, may interfere with CryoSPARC operation.
You may also want to ensure that the OS of the selected server is still supported by security updates and not beyond EOL.

5

Thank you very much! So far both master and worker packages were installed on a lustre-fs, accessible from master and worker nodes.

We are experiencing performance issues of the GUI (e.g. when directories with many filenames are accessed) and stopping of the cryosparc process every other week (i.e. “cryosparcm status” returns “Cryosparc is not running.” while the GUI becomes very laggy. This requires a kill of all cryosparc-mongo processes and a restart of cryosparcm). Hence I am now testing, whether having the master package available locally on the master node may help here.

6

The precise cause of this problem should be investigated. This message can be triggered, for example, if a CryoSPARC-related socket file, typically at
/tmp/cryosparc-supervisord-*.sock, is improperly removed. Improper removal may occur if a regularly scheduled system “cleanup” task, such as systemd-tmpfiles, is improperly configured.