admin console not loading

[nullx8]

every time lsws runns longer than 10 minutes, i can not access the management site (7080) at all ...
its simply loads forever while the (only one) production host
serves 120-350pages/sec with out any issues at all.

this morning the server did restart basically every 2 minutes to fix 503 errors.
i upgraded then to "Enterprise v4.2.1 (2 CPU)"

now the 503 restarts are gone and the production site works flawless without me having any code changed.

however if i go now to administations-urls like config/confMgr.php?m=serv
i get error 500

and have to restart lsws via commandline to get arround this (soft restart via web interface does not do anything)

any idea what that could be ?

the machine is a dedicated webserver,
has only 2 working tasks running lsws and lsphp5
always about 2gb memory free, no swap used, load never higher as 0.05

 

[NiteWave]

strange issue, 1st time to hear about.

how about you change port 7080 to other one like 7100 ?
what's the output of:
#netstat -ntlp |grep 7080

 

[nullx8]

how about you change port 7080 to other one like 7100 ?

sure i can do that, but what shoud that help ?

what's the output of:
#netstat -ntlp |grep 7080

now the error happens the netstat output is (as it is always since the port is still open):

[root@profiles ~]# netstat -ntlp |grep 7080
tcp        0      0 0.0.0.0:7080                0.0.0.0:*                   LISTEN      12720/litespeed (ls

 

[nullx8]

stderr shows:

==> stderr.log <==
lsapi: failed to open secret file: /usr/local/lsws/admin/cgid/secret/cgid.sock!
2012-11-30 03:52:05.255 [STDERR] lscgid: secret is not set properly!
2012-11-30 03:52:05.257 [STDERR] lsapi: failed to open secret file: /usr/local/lsws/admin/cgid/secret/cgid.sock!
2012-11-30 03:52:05.257 [STDERR] lscgid: secret is not set properly!

as told before, this happens only if the service do run for a while (10 min or more)

and since the upgrade the administration does produce the error500
before the admin sites just never responded to any http calls

 

[mistwang]

Can you also send us the error log around that time? Need to know what happened by that time.
Or, let our staff login to take a look. send email to bug@litespeed... or PM temp root access.

 

[nullx8]

The only entries in the servers error log are this:

==> /usr/local/lsws/logs/error.log <==
2012-11-30 04:00:33.215 [WARN] [config:server:rails] Cannot find ruby interpreter, Rails easy configuration is turned off
2012-11-30 04:00:33.215 [ERROR] [config:vhost:profiles] missing <certFile> in <vhssl>

Where both make no sense to me (ruby is not used and ssl working just fine)


I noticed that the service still do crash and produce the 503 error, and right after that the admin console stops working.

The crash is caused by glibc

==> /usr/local/lsws/logs/stderr.log <==
*** glibc detected *** lsphp5: double free or corruption (out): 0x00007f102c10a040 ***