Nginx nginx upgrade

Sounds like you ran out of memory during upgrade.

Do you have any persistent config file options set in /etc/centminmod/custom_config.inc ? You can check via this command:

Code (Text):

cat /etc/centminmod/custom_config.inc

Post output in CODE tags.

To troubleshoot, you need to check the the nginx upgrade log at /root/centminlogs and instructions under Sharing logs and errors heading for using Pastebin.com or Gists to share a sanitised version of the contents of the nginx_upgrade.log log. You can see full details at How to troubleshoot Centmin Mod initial install issues

also what's output of

Code (Text):

nginx -t

when you run centmin.sh menu option 4 there's a nginx upgrade log timestamped at /root/centminlogs

if you type this command it lists all logs in date ascending order so latest log at bottom

Code (Text):

ls -lArt /root/centminlogs

so copy the entire contents of latest nginx_upgrade log to gist.github.com or pastebin.com

you can use grep to filter the logs, i.e. look for nginx_upgrade in log name

Code (Text):

ls -lahrt /root/centminlogs/ | grep nginx_upgrade
-rw-r--r--  1 root root 3.2M Oct 11 15:55 centminmod_1.2.3-eva2000.09.001_111016-155345_nginx_upgrade.log
-rw-r--r--  1 root root 672K Oct 11 22:06 centminmod_1.2.3-eva2000.09.001_111016-220515_nginx_upgrade.log

So the last nginx upgrade log was named centminmod_1.2.3-eva2000.09.001_111016-220515_nginx_upgrade.log and located at /root/centminlogs/centminmod_1.2.3-eva2000.09.001_111016-220515_nginx_upgrade.log

then use cat command to output the contents of that log

clear your ssh window buffer/screen and type

Code (Text):

cat /root/centminlogs/centminmod_1.2.3-eva2000.09.001_111016-220515_nginx_upgrade.log

then select and copy and paste output to pastebin.com or gist.github.com file to share. If your SSH client's scroll buffer isn't large enough using cat might not output the entire log file contents, so you may need to download the log and use local text editor to open and copy and paste.

so only need content of one specific log, in this case most recent nginx_upgrade.log log

 

looks like you ran out of memory when it came to ngx_pagespeed part of nginx install/compile

how much memory do you have ?

what's output for this command in CODE tags

Code (Text):

free -m

what's output for sar memory usage profile for yesterday

Code (Text):

sar -r -f /var/log/sa/sa$(date +%d -d yesterday) 

and today

Code (Text):

sar -r

check around 14/02/17 20:51 time which is the timestamp of your nginx log

 

yup sar output for yesterday confirms it at ~9:20pm you had >100% committed memory so ran out of memory during nginx compile

Code (Text):

07:10:01 PM kbmemfree kbmemused  %memused kbbuffers  kbcached  kbcommit   %commit
07:20:01 PM     81752    420248     83.71      2860    124080   1290664     83.24
07:30:01 PM     83920    418080     83.28      3156    115940   1286676     82.98
07:40:01 PM     93540    408460     81.37      8788    156872   1220704     78.73
07:50:01 PM     70232    431768     86.01     11044    118792   1292944     83.38
08:00:01 PM     86172    415828     82.83      3072    107216   1321340     85.22
08:10:01 PM     99248    402752     80.23       380     77848   1359792     87.70
08:20:01 PM     80668    421332     83.93      9372    131244   1282708     82.72
08:30:02 PM     91940    410060     81.69      4788    134300   1236840     79.77
08:40:01 PM     73244    428756     85.41      5956    125392   1317188     84.95
08:50:01 PM     93524    408476     81.37      2724    121188   1273428     82.13
09:00:01 PM     95696    406304     80.94     10284    145084   1212692     78.21
09:10:01 PM     97152    404848     80.65      3372    140120   1246436     80.39
09:20:03 PM     74020    427980     85.25      1488     41560   1887792    121.75
09:30:03 PM     69568    432432     86.14      1680     28088   2177200    140.41
09:40:02 PM     91056    410944     81.86       464     30820   1970648    127.09
09:50:02 PM     88108    413892     82.45       740     24444   2031120    130.99

maybe try centmin.sh menu option 4 recompile/upgrade of nginx again - could be something else eating into memory at same time - though sar output shows constant 77-85% mem used so you're running on borderline.

Good time to switch to 1GB Linode at $5/month though https://community.centminmod.com/th...-month-digitalocean-vs-vultr-vs-linode.10437/

 

check benchmarks https://community.centminmod.com/th...-month-digitalocean-vs-vultr-vs-linode.10437/ including post install disk free space comparison

 

Yeah true, 2GB Linode fits that bill nicely

 

are you actually using ngx_pagespeed, if not you can disable it from nginx compiles to save memory usage too

so set /etc/centminmod/custom_config.inc to

Code (Text):

NGINX_PAGESPEED='n'
NGXDYNAMIC_NGXPAGESPEED='n'

 

Not really as 1.11.10 isn't a security release so should be fine. But generally better to keep updated.

have you tried a 2nd attempt at nginx upgrade as is ? try it

 

yeah you can do that i do that on 128MB VPS running CentOS 6 32bit with centmin mod. Stop php-fpm, mysql and memcached servers, update Nginx and restart

 

to prevent errors for visitors do it like this

Code (Text):

npstop
mysqlstop
memcachedstop

update nginx

Code (Text):

mysqlstart
memcachedstart
nprestart

 

Great to hear