Segmentation fault

Emoncms
1

I’ve just deleted an unused feed, not realizing that there was a redundant input process still in existence, so the input list showed an ERROR against that process.
Logical step then was to delete the input process, but they still remained after clicking the trash can (the input is redundant - so not recreated because of an update).
So… then rebooted hoping that it may clear, but things then got worse!

The FEEDS page is blank apart from the spinning ‘busy’ wheel, (should be about 30 phpfina feeds), and checking my browser, I see a console error of GET http://192.168.1.10/emoncms/feed/list.json net::ERR_EMPTY_RESPONSE VM47:1.

The INPUT’s page shows all of my inputs, except they are not updating, and the process lists are showing wait...

My apache log is showing [Mon Sep 11 10:14:59.059545 2017] [core:notice] [pid 1260] AH00052: child pid 3435 exit signal Segmentation fault (11)

No errors in the emoncms log since reboot.

Any ideas?

Emoncms Version 9.8.10 | 2017.08.17 (self build)

2

The lines in the log just prior to the segfault may offer some insight.
Could you post about 10 lines from that part of the log?

3

Thanks Bill, here are the preceding entries.
The 09:41:12 is the first entry for today.
Feed 76 is the feed which I mentioned above.

[Mon Sep 11 09:41:12.343460 2017] [:error] [pid 12775] [client 192.168.1.9:53214] PHP Warning:  fopen(/var/lib/phpfina/76.dat): failed to open stream: Permission deni$
[Mon Sep 11 10:04:09.032362 2017] [mpm_prefork:notice] [pid 1239] AH00169: caught SIGTERM, shutting down
[Mon Sep 11 10:04:42.829508 2017] [mpm_prefork:notice] [pid 15755] AH00163: Apache/2.4.10 (Raspbian) OpenSSL/1.0.1t configured -- resuming normal operations
[Mon Sep 11 10:04:42.830000 2017] [core:notice] [pid 15755] AH00094: Command line: '/usr/sbin/apache2'
[Mon Sep 11 10:04:43.836211 2017] [core:notice] [pid 15755] AH00052: child pid 15760 exit signal Segmentation fault (11)
[Mon Sep 11 10:04:43.836582 2017] [core:notice] [pid 15755] AH00052: child pid 15761 exit signal Segmentation fault (11)
4

Looks like a permision error.

Is this a system that’s been running for some time, or a new system?

5

No, not new, it’s running constantly.
The reason it couldn’t open 76.dat was because I had deleted it, as I mentioned in my first post

6

Right, bu that shouldn’t give you a permission error.
I’m thinking more along the lines of a file not found type of error.

What came to mind, is failing flash memory. Especially if it’s a system that has been running for a long time.
It looks as if it’s trying to re-create the missing data, but can’t, which is what made me think “flash going south.”

7

It’s an external HDD, not a SD card.

8

Just thought of that. I recall a comment you made a few years ago about one of the side bennies of having a Pi on an HDD was the server providing strorage for other things…

9

At any rate, it might be worth a look at the directory the file was in, to verify the permissions are OK.

10

Yep, they are fine.
I can’t help but think that this is Redis related…
If you look at the para starting ‘March 2016’ in the link below that mentions ‘Segmentation errors in Apache log’ and I did have php5-redis v2.2.5 installed (which I have now updated to v2.2.7 but still the same problem).

Paul

11

the updated version of php5-redis

And the updated module, that’s been installed too?

12

Do you mean phpredis V3.1.1?
If so, no because the guide says;

‘This guide will be updated to reflect this change once tested in due course.’

Paul

13

Is 3.1.1 redis itself or the php module?

I was speaking of the module.

Looks like this is one for @glyn.hudson

14

I’ve not specifically updated php, and it’s showing
PHP v5.6.30-0+deb8u1 (Zend Version 2.6.0)
Redis version
v2.8.17

Paul

15

Doesn’t matter now, I’m rebuilding the system from scratch in Raspbian Stretch, using PHP7.0 and many other updated libraries.

I’m documenting as I’m progressing, and will put together an installation guide shortly.

Paul