Forums

access.log logs stopped logging

I've recently updated my Django virtualenv installation due to the security release they issued and I'm sure the logs were working then. At the moment even if I visit any page I can't see the GET requests on my access.log file.

Is this a Django related problem, a webserver software related problem, or should I look for the access logs elsewhere? Thanks!

Hm, no, that's likely to be a bug on our side. We'll take a look...

Just an update to say we're still investigating this one. We did change the logging code recently, so it's likely that that broke it. But we're still trying to work out how.

Looks like it was a file permissions problem in our migration from the old system to the new one -- it affected you and a few other users. We've fixed it manually, it should be working now.

Thanks for reporting the problem!

Any chance this problem has recurred? My access and error logs are completely empty, and I know accesses are happening (server log appears to be fine). Any thoughts?

Yes, we've had some trouble with the logs after the Trusty upgrade. I have a fix in that should fix it. I'll keep an eye on it to make sure it's working.

The same problem has started occurring in my account. The access log file has stopped logging all the requests

Ok. I've fixed that for you.

Stopped working for me too. The timestamp of the last log I have is 22/Nov/2015:19:34:13

We're doing a system update tomorrow morning, so we'll see if that fixes it.

The logs do not seem to be working still. Can you please confirm when this issue will be resolved.

Ok. We have a solution and it looks like logs are updating within a few seconds again. We've also implemented some additional monitoring so we can keep an eye on them to ensure that they don't deteriorate again.

Let us know if you're still having trouble.

The logs are still not accessible. The count of hits is increased regularly but no change in access log file is there. It is still the same.

No logs here since dec 9 2015

Thanks for reporting that. We're still debugging -- for anyone who's interested, it looks like a problem with open file limits for a specific rsyslog process. Will report back when it's been updated.

Hi,

is this concerning the "logging" module in python3.4?

When using this module in console (on python3.4), it works fine (I copy/pasted between console and script).

When creating a script, I receive the following error:

    logging.basicConfig(level=logging.DEBUG, format=' %(asctime)s -%(levelname)s - %(message)s')
AttributeError: 'module' object has no attribute 'basicConfig'

The first 3 lines of the script are:

import logging
logging.basicConfig(level=logging.DEBUG, format=' %(asctime)s - %(levelname)s- %(message)s')
logging.debug('Start of program')

or do I need to create a seperate thread for this one.

No, this is a different thing -- it's about the access logs that you can see on the "Web" tab.

Best to move this to a different thread. I'll do that now.

[edit] Moved here.

Please fix my logs. Not updating, also "Web->Traffic" stats stuck at "tbc".

Sorry about that, it looks like one of our servers is taking a while to come back after this morning's system update. Access logging will come back when it's recovered. It's on its way and the hits are queued so your logs and stats should all be back-filled.

Log lines started to appear, but no queued entries. Still need to wait or they gone?

They are going to be gone in this case as the logging process locked up post deploy. Usually it should be queued/backfilled though.

Understand, thanks.

I just had the same problem as mentioned in this blog. All my logs are completely empty and Web -> Traffic is at tbc. Anyone know if there is a way to fix this currently?

fixed again, dont know why or how I had this problem and how it got fixed.

Sorry about that! We did some system maintenance last night -- it was finished at 00:20, about eight hours before your post, but it looks like there were some logging problems afterwards that we didn't pick up on. Everything should be OK as of 09:30 UTC (about three hours ago).

My error logs are not updating for the past 6 hours at least. Is there any maintenance activity going on?

We don't have any maintenance underway, no. Is it possible that your website just hasn't had any errors over this period?