Hi,
we have a question involving excessive DNS queries from a CryoSPARC installation on campus. This is a faculty-owned server and I do not know much about CryoSPARC as I am not a user of the software, but rather a facilitator working in Research Computing trying to help the faculty solve the networking issue.
The problem is the following: It seems to be that every time there is a new client computer accessing the CryoSPARC server, the CryoSPARC server swamps our campus DNS server with lookup requests trying to resolve the IP address for the client.
The etc/hosts file of the CryoSPARC lists its own hostname ( i.e. 127.0.0.1 localhost cryosparcserver) as suggested in this forum post: Excessive DNS lookups created by cryoSPARC? - #2 by stephan as this was a problem before.
We also added the client’s IP addresses to the etc/hosts file and this seems to stop the behavior with the excessive lookup requests to the DNS server.
However, this is rather inconvenient as static IP addresses are a limited pool and on our campus require renewal annually making this a cumbersome and error-prone process.
We are wondering if other people have encountered this issue and if there is a better way of doing this.
Not sure if this is an issue that can be caused by CryoSPARC or if this is an underlying OS issue. Perhaps changing DNS lookup behavior via etc/host.conf and/or etc/nsswitch.conf would solve the issue?
Any insight is appreciated.
Thanks,
Joe
Welcome to the forum @HPRC_1
With “client”, do you mean a computer running a web browser to access the CryoSPARC UI hosted on the CryoSPARC master computer?
Thank you for responding. That’s correct.
Regards,
Joe
@HPRC_1 Do you agree that the DNS requests in your case differ from those mentioned under the other topic Excessive DNS lookups created by cryoSPARC??
Are users accessing the the CryoSPARC web app via a reverse proxy?
I apologize for the late reply: the CRYOSPARC_MASTER_HOSTNAME has been added to the /etc/hosts file and as far as I know the requests were still happening after. Adding client’s IP addresses stops the behavior. This is what is portrayed in the case in the mentioned post I believe.
My question was going further: “However, this is rather inconvenient as static IP addresses are a limited pool and on our campus require renewal annually making this a cumbersome and error-prone process.”
We are also wondering why the CryoSPARC server does not cache the IP address after the first request, but keeps “pestering” the DNS server for the address over and over again.
According to our faculty that runs the CryoSPARC:
No, users do not access the CryoSPARC web app via a reverse proxy.
They use ssh via terminal like this: ssh -N -f -L localhost:39004:localhost:39000 cryosparc_user@cxxxxxxx.
Thank you!
Which process launches the DNS queries to which you refer?
With server, do you mean the host that runs CryoSPARC software or the (software) server components of CryoSPARC?
Thank you for this detail. To confirm: the DNS queries are for the address of the client where the ssh command was run, not for worker (GPU) hosts?
With server we refer to the physical system that hosts the CryoSPARC software, registers with a fully qualified domain name on our network and has a static IP address. This hostname was added in the etc/hosts file as specified by your previous post.
Correct, the DNS queries are for the client computers where the ssh command was run.
As to which software process of the CryoSPARC software (if any) might be involved in the requests that is part of what we are trying to find out.
Thank you!
We are not aware of how the DNS queries for the ssh client would originate within CryoSPARC and do not have any suggestion for the resolution of this issue. Please post here any additional information that your investigation may reveal.