Tony Meyer's questions

(It could be that this belongs on SO; I can't figure out how to work out if it's a code issue or something else).

I have a server running:

• Debian 7.6
• nginx 1.6.0
• uWSGI 1.9.16
• Django 1.6.5

I have code that does this:

1. accepts a form post containing a Word (application/vnd.openxmlformats-officedocument.wordprocessingml.document) file
2. processes the file creating a new Word file
3. returns the new file

This works fine in Safari, Firefox, Chrome, etc. It works fine in IE<10. It works fine in IE 11.0.9600.17126, update version 11.0.9 and earlier versions of IE 11.

In IE 11.0.9600.17207, update version 11.0.10, it does not work (tested on multiple computers, all Windows 7, some in a domain, some not). The page in IE is non-responsive; there's no timeout (for several minutes at least) or error (the spinner is still spinning and the close tab button works). In the network activity section of the developer tools it just shows as pending. Changing the User-Agent string doesn't make any difference.

The code executes at least as far as the return in the view handler. There's nothing in any of the code that behaves differently depending on the browser/user-agent.

How can I diagnose what is happening here? Is there a way to tell whether IE understands that it has finished sending the file? Or to know whether it is trying to get the response, or whether it has received any response? This is all over SSL, although I could temporarily disable that if necessary. Is there something special that needs to be done with nginx/uWSGI/Django to handle this latest version of IE?

I have a very simple Exim configuration that's just forwarding all mail to a set of destination addresses. When immediate delivery to an address fails, the message is added to the queue (and then processed by the retry rules). I want to change this so that if immediately delivery fails, the message is :blackhole:d. (It's ok if a bounce is generated instead, as I'll just redirect the bounce to the :blackhole:).

This needs to occur for temporary failures (i.e. 4xx) as well as permanent (i.e. 5xx) ones.

I understand that this means that if delivery can't be done immediately the message will be permanently and irretrievably lost. In this particular context, that isn't a problem.

Reading over this, it sounds suspiciously like "how can I improve my spamming Exim server". That really isn't what this is for, and if you can figure out a way I can prove that, I'm happy to do so!

I have a few scripts that automatically send out small email notifications when necessary. I get a small number of auto-response messages (e.g. out-of-office or vacation) back from these. Is there a header I can add to the email that will indicate to (well-behaved) auto-responders that I'm not interested in any response?

See also this related StackOverflow question.

Following on from this question. Debian, if that matters.

I understand that some memory is shared between forked processes. How then can I determine how much memory is used by a process/set of forked processes?

Using the smem tool recommended in this similar question, I get values like:

Command                         Swap      USS      PSS      RSS 
/usr/sbin/spamd --create-pr        0    16820    24974    41628
spamd chil                         0    19388    27581    44176
spamd chil                         0    32328    40038    55708

I understand the swap column, and the RSS column is what is normally reported (e.g. in ps). From reading the smem documentation, it sounds like USS is the memory specifically for that child, and PSS is a portion of the memory shared across the processes. However, adding USS to PSS gives higher values than RSS, and I thought that the sharing would mean less (so I'm presumably interpreting incorrectly).

I'm not tied to the smem tool. I would just like some way of getting a "memory use" number that somewhat accurately reflects the amount of real memory that the process is using.

I'm using SpamAssassin on Debian (the default configuration with Pyzor, AWL and Bayes disabled, and sa-compile enabled), and each of the spamd child processes consume around 100 to 150MB of memory (around 50MB of real memory) on the 32 bit servers, and about double this (logically enough) on the 64 bit servers. There are generally two child processes, but at busy times there can be five (the maximum) running.

ISTM that 200 to 600MB is a lot of memory for this task. I'd like to continue using SA as part of my filtering structure, but it's becoming difficult to justify so much memory.

Are there any ways to reduce the amount of memory that each child process uses? (Or alternatively, make a single child process so fast that I can set the maximum children to something like 2?). I'm willing to consider any options, including ones that will or may result in reduced accuracy.

I've already read the "Out of Memory Problems" page on the SA wiki; nothing there is of any use. Messages larger than 5 MB are not scanned with SA.

In particular, how do you choose between MyISAM and InnoDB, when neither is missing a required feature (e.g. you don't need foreign keys).

Does it always come down to trying both and measuring? Or are there good rules of thumb regarding the number and frequency of reads versus writes, and other measures like that? Does the size of the table have any effect on the typical choice?