Hi all,
I'm encountering a problem when I try to add a client to HQueue. The client gets added and says “NEEDS-UPGRADE”. I do what it tells me and I click on “upgrade”. An upgrade job is added to the client.
This is where the fun stops, the job is showing status “waiting for client…” and never continues with the upgrade. How does one get past this? Is there a different way to upgrade client? I havn't found one.
Both client and server run Ubuntu 12.04
Houdini 12.5, build 371
All the best,
Elias Ericsson Rydberg
HQueue: problem adding client
6788 7 4- Elias Ericsson Rydberg
- Member
- 41 posts
- Joined: 4月 2013
- Offline
- Elias Ericsson Rydberg
- Member
- 41 posts
- Joined: 4月 2013
- Offline
- tex1820
- Member
- 22 posts
- Joined: 3月 2011
- Offline
- Elias Ericsson Rydberg
- Member
- 41 posts
- Joined: 4月 2013
- Offline
- tex1820
- Member
- 22 posts
- Joined: 3月 2011
- Offline
- Luke Letellier
- Member
- 241 posts
- Joined: 4月 2014
- Offline
I'm having the same issue. The client is installed & technically recognized by the server, but there's no response from the client. No heartbeat. When I open up the client panel, under status I see “Waiting for Machine”.
When I try to restart the client, I get:
"Failed to restart '': coercing to Unicode: need string or buffer, NoneType found"
Thoughts?
Thanks,
Luke
When I try to restart the client, I get:
"Failed to restart '': coercing to Unicode: need string or buffer, NoneType found"
Thoughts?
Thanks,
Luke
- goldleaf
- スタッフ
- 4195 posts
- Joined: 9月 2007
- Offline
What OS are you installing under? The only time I've had luck adding a client via the web interface is when I changed my router. Originally (and this is how I recommend it) I just used the Houdini installer to setup clients, pointed them to the server, and I was good to go.
However, when I changed my router, the clients weren't responding since their IP addresses changed, so I removed those clients, and re-added the same machines via the web interface; this time it worked. Also, instead of IP addresses, I put $HOSTNAME.local (i.e. computer1.local, if a machine were called computer1). This was all on linux, btw.
But overall, I'd recommend just using the Houdini installer and only installing the client, when adding a machine to HQ!
However, when I changed my router, the clients weren't responding since their IP addresses changed, so I removed those clients, and re-added the same machines via the web interface; this time it worked. Also, instead of IP addresses, I put $HOSTNAME.local (i.e. computer1.local, if a machine were called computer1). This was all on linux, btw.
But overall, I'd recommend just using the Houdini installer and only installing the client, when adding a machine to HQ!
I'm o.d.d.
- eddgarpv
- Member
- 64 posts
- Joined: 7月 2005
- Offline
So I was having the same problem: “Failed to restart ‘': coercing to Unicode: need string or buffer, NoneType found”
Recently switched from Debian Crunchbang to Centos 7 and in debian everything was working fine, until I had to set up the HQueue server on the fresh centos workstation. It wasn’t working because the firewall was enabled and running. After disabling the firewall, I could talk to clients (enable/disable).
I still can't restart'em in the web interface because I get: “Failed to restart ‘thisnode’: Username, password or installation directory is not set. You can change the remote login in the Client Details section.” but I can ssh to the node and restart the client there (a bit of kerfuffle but at least the server/clients see each other).
I still have to check other security options in centos, maybe that will solve the restart node problem, but this hopefully will help others debugging errors.
-ed
Recently switched from Debian Crunchbang to Centos 7 and in debian everything was working fine, until I had to set up the HQueue server on the fresh centos workstation. It wasn’t working because the firewall was enabled and running. After disabling the firewall, I could talk to clients (enable/disable).
I still can't restart'em in the web interface because I get: “Failed to restart ‘thisnode’: Username, password or installation directory is not set. You can change the remote login in the Client Details section.” but I can ssh to the node and restart the client there (a bit of kerfuffle but at least the server/clients see each other).
I still have to check other security options in centos, maybe that will solve the restart node problem, but this hopefully will help others debugging errors.
-ed
-
- Quick Links