Leginon

If redux server shows no record of receiving the request, then the port it opened didn't receive any instruction. Check to see if the port configured in myamiweb/config.php and in redux.cfg are matched. imcached does not use this mechanism, so it is not affected.

Anchi,

thank you for pointing me to the right direction. I was sure that the port number in the two files was the same, but actually there was a typo that I overlooked many times. Now the images are shown.
At this point I am trying to understand how to make the image caching work. I followed the instructions of the related page, and I am running imcached on the old sessions to generate the jpgs, but from the redux log there is not indication that images opened in the image viewer are from the cache: the Read operation is always on the mrc. Do I need to enable the cache also in the redux.cfg file? When I tried to do it, instead of the pics I get 'TypeError: can't write str to text stream', which is fired when trying to write to the disk cache (put_result in redux/pipe.py). Given the message I tried to open the disk cache in binary mode and, while the behavior is unstable (not all images are shown), I can find messages in the log specifying if the file is in memory or not, and only in two occasions it found the image in file (I assume it means in the directory designated as cache). However. still not sign of reading of jpgs.

Thanks,
Giovanni

Sorry for delay.

With imcache, you need to define CACHE_PATH in myamiweb/config.php to where imcache stores the jpeg files.

redux cache and imcache are two separate cache. You can use either.

Let's say the image you request from Redux go through a pipeline of read->binning->scale->format. Redux cache saves the result of each pipe in the pipeline in memory and on disk so that if you request from redux the same image with the same binning but at different intensity scale, it does not reload the image from the network, nor redo the binning. When the pipe result is not in memory, it looks in the disk cache. This saves a lot of results and is good if you need to do different operations on the same image often. Because the amount of information it saves, a limit is set to clear them automatically within the configuration.

IMCACHE makes jpeg files with redux at a fixed size as the default loaded to myamiweb. It is more permanent. Once you've configured myamiweb to look for these first, it saves time as long as the jpeg files are made first, as often the case when you look at an overnight session next morning. In some labs, they would archive the actual mrc files off network and keep only these jpeg files online for occasional comparison with recent sessions. The size limit of this "cache" is only the size of the disk available.

Anchi

Anchi,

thank you for the explanation. Still it puzzles me why the redux cache mechanism doesn't work right out of the box on our fresh new CentOS6 installation, but I'll try to investigate it more when I have some time.

Thanks again,
Giovanni