Jump to content
Sign in to follow this  
kensonuken

Afanasy ( email acknowledgement )

Recommended Posts

Im tyring to use afanasy and I would like to know that is there any email commands which will be sent by the afanasy render client or server in order to identify the failed tasks or executed tasks or tasks on hold or etc.....

Share this post


Link to post
Share on other sites

No, Afanasy can't do it (yet). But I am thinking about it.

It not so hard to make it.

For example Afanasy can execute a command with environment variables contains job name, task name, user name, ... etc.

Command can be described in Afanasy XML config file.

Simple python program, sending emails can be provided as an example.

Who must execute command server or render host?

Launching such commands on remote host can save server resources.

Where email addresses will be stored, in Afanasy database? And whose attribute (address) it must be? Job or User?

May be each user can have an address, and any job can override it.

May be better not to store addresses, store commands to execute on task|job status change. This commands can contain anything and do anything.

Share this post


Link to post
Share on other sites

+ Afanasy has a very simple web based GUI. It allows only to watch statuses, not to change anything (for security reasons it may not be able to change anything). And it is very raw now. But "It works."

(Excuse my english if mistakes)

Share this post


Link to post
Share on other sites

sounds good and would love to see some sample py script which can send a mail... I can automate something but atleast if i know some triggering event so that i can use that event to trigger the mails...

some tip will be good..

Im trying to trigger the mail using the server though...

There are two things that I want to use Afanasy for...

1. To run couple of simple python scripts like backup scripts for migrating data over the network stuff after render is done.... ( i dont know how to do this but would like to see some of them... )

2. For rendering of frames over the farm...

pls suggest me some stuff for my above requirements...

Share this post


Link to post
Share on other sites

I thinking about some command which can be executed with some identifiers (user, job, task, status), for sending emails, for example. Some people wants me to realize it in Afanasy (and I'll do it).

Or you want each task to have a commands to execute on success or fail?

For example our developers wrote python scripts to launch our working soft through. We launch Houdini and Nuke which render frame(s) locally in some 'temp' directory, then copy result on file server and clean 'temp' (it saves some traffic).

So you can launch any custom things you wrote, not to launch Nuke or Houdini directly.

Share this post


Link to post
Share on other sites

yeah thank u I mean

we just need to know about the render statistics like... render machine name

start time

end time

avg time

start memory

end memory

avg memory

max memory

user name who triggered the renders

list of machines which rendered the complete sequence will be a big advantage

list of machines which failed to render and mark it as failed would be better to see which systems are giving the failed renders...

it will be better if we have

frame 101 machine name rendered "memory used" "time taken"

frame 102 machine name failed memory used time taken

at least we would know how many frames are rendered properly and

Grouping of tasks into one single task ( show as single task instead of multiple tasks ) ( we may render per light outputs in such cases as one group rather than thousands of light tasks on the farm )

while rendering the scripts its better to have dependencies of scripts sequences $F rather than just dependencies of tasks numbers any more ideas would be nice..

custom defined colors for specific state of renders will be good just like Frantic films Deadline has...

Edited by kensonuken

Share this post


Link to post
Share on other sites

Nice wish list )

I think i'll improve Afanasy Python API.

When task came to remote host, this render host instance some Python class with some initial parameters.

Before task start 'pre' method will be called (by this render), with some parameters.

After task finish 'post' method will be called. This post method can send emails, insert data in your statistics database, or even can check something and switch task state error.

It will be useful i hope.

I am thinking about how to specify where to store any custom parameters you want to pass to this class.

"list of machines which failed to render and mark it as failed would be better to see which systems are giving the failed renders..."

- Afanasy already stores 'blask lists' - error hosts for each block and each task. Host name last task session was, and a log what, where and when was happen with task. - But it all only in server RAM, it does not store it anywhere else. You can query it in 'Watch' GUI.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  

×