Stuck with 503 error on first time install

First time I have ever tried a Cyberpanel install was over the last few days.

I am using VMs on homelab before moving my business accounts to a VPS.

I have been testing other panels and have been running development servers for awhile. I have clean minimal up to date VMs that I can spinup.

First install was on Ubuntu 18.04 – got a 503 trying to access cyberpanel (8090) and could not get to LSCP (7080) . I did some reading and saw that centos was the first choice.

Tried a Centos 7 install - got a 503 trying to access cyberpanel (8090) and got to LSCP (7080) after opening the firewall for port.

I tried several things (sudo install, root install and forced upgrade). Always with OLP and have tried with other options marked yes or no.

OLP works fine and I can server pages…have tested PHP and that is fine as well.

I have been through the logs. The only trace I can find is in lscp/logs where a “ExtConn timed out while processing” shows up.

The big issue is I just started with 1.8.2 (so most of the help is off topic e.g. gunicorn) and I did not realize that at some point yesterday I was getting 1.8.3.

Is it a user:group thing (not sure which directories should be owned by which).

Or is it related to my first install being with 1.8.2 after your recent changes?

--------------->
2019-05-06 14:17:58.424013 [NOTICE] Loading LiteSpeed/0.1.8 Open …
2019-05-06 14:17:58.424274 [NOTICE] Using [BoringSSL]
2019-05-06 14:17:58.688809 [NOTICE] [LscpdCommSocket] server socket: uds://usr/local/lscpd/admin/comm.sock and fd 11 uid 1001 gid 1001
2019-05-06 14:17:58.726852 [NOTICE] [AutoRestarter] new child process with pid=2747 is forked!
2019-05-06 14:17:58.815920 [NOTICE] [child: 2747] Successfully change current user to lscpd
2019-05-06 14:17:58.817567 [NOTICE] Child: 2747] Core dump is disabled.
2019-05-06 14:17:58.863955 [NOTICE] [Child: 2747] Setup swapping space…
2019-05-06 14:17:58.865517 [NOTICE] [Child: 2747] LiteSpeed/0.1.8 Open starts successfully!
2019-05-06 14:17:59.075705 [INFO] checkStatusFile() called with uid 5001 randoffset -1.
2019-05-06 14:17:59.075967 [INFO] checkStatusFile() return 1.
2019-05-06 14:18:29.043501 [NOTICE] [192.168.1.26:1173] ExtConn timed out while processing.
2019-05-06 14:18:32.075624 [NOTICE] [192.168.1.26:1173] ExtConn timed out while processing.
2019-05-06 14:18:35.083415 [NOTICE] [192.168.1.26:1173] ExtConn timed out while processing.
2019-05-06 14:18:38.088476 [NOTICE] [192.168.1.26:1187] ExtConn timed out while processing.
2019-05-06 14:18:41.105783 [NOTICE] [192.168.1.26:1187] ExtConn timed out while processing.
2019-05-06 14:18:44.080732 [NOTICE] [192.168.1.26:1187] ExtConn timed out while processing.
2019-05-06 14:20:39.006023 [NOTICE] [192.168.1.26:13656] ExtConn timed out while processing.
2019-05-06 14:20:42.069339 [NOTICE] [192.168.1.26:13656] ExtConn timed out while processing.
2019-05-06 14:20:45.077560 [NOTICE] [192.168.1.26:13656] ExtConn timed out while processing.

after some more work on ubuntu 18.04 this got me past the “503” error with 1.8.3:

chmod -R 777 /tmp/lshttpd
systemctl restart lscpd

Can you provide these details → Bug Report Template - Bug Report - CyberPanel Community

and KVM or OpenVZ?

I did read that before posting…but I attempted 5 clean installs and 4 forced upgrades. Only the last one after my post.

Local machine with virtualbox (homelab). VMs had 3 cores / 1 gig mem / 30g drive. Always selected OpenLitespeed

1 install and one forced upgrade on a minimal install of Centos 7.6-1810 fully updated.

• I tried some gunicorn things (no longer relevant)
• I tried the manual python virtual environment activation listed somewhere.
• I discovered that opening 7080 in firewalld gave me the Litespeed panel
• tested database / webserver / php all worked fine.
• I did not revisit Centos install to test the “fix” for the 503

3 installs and 3 forced upgrades on a minimal install of ubuntu server 18.04 fully updated

• I tried without redis and memcache on some
• I tried some gunicorn things (no longer relevant)
• I tried the manual python virtual environment activation listed somewhere.
• After centos I opening 7080 in firewalld which gave me the Litespeed panel
• tested database / webserver / php all worked fine.
• I tried some change password things / deleted tmp files / restarted services
• I tried some owner / group changes listed somewhere – mainly because during install there are some onscreen errors about nogroup and nobody
• I found the fix mentioned above to get into Cyberpanel (chmod in tmp directory)

1 Current clean install on a minimal install of ubuntu server 18.04 fully updated

• I had made many changes in previous installs wanted a clean slate.
• the chmod in the tmp folder fixed the 503
• I have begun testing (adding websites / wordpress installs) so far very nice
• PhpMyadmin was missing - I used the install script as a guide (download_install_phpmyadmin section) and did it manually – is now working

Phpmyadmin is an error listed in the installLogs.txt – I have not tested modsec yet.
cyberpanel is a user and seems to be in proper groups I have not changed anything.

from installlogs.txt ------------------------->

[08-11-05-Mon-May-2019] [ERROR] We are not able to add user cyberpanel return code: 9 [setup_account_cyberpanel]
[09-10-02-Mon-May-2019] [ERROR] We are not able to add lscpd to important groups return code: 6 [fixCyberPanelPermissions]
[09-17-42-Mon-May-2019] [ERROR] We are not able to fix permissions /usr/local/CyberCP/public/phpmyadmin/tmp return code: 1 [fixCyberPanelPermissions]

[08-39-51-Mon-May-2019] [Errno 13] Permission denied [setupPHPAndComposer]
[08-55-50-Mon-May-2019] [Errno 2] No such file or directory [download_install_phpmyadmin]
[08-57-43-Mon-May-2019] [Errno 2] No such file or directory: ‘/usr/local/lsws/modules/mod_security.so’ [modSecPreReqs]

There are more errors shown on screen during install than are in the logs. e.g. I saw some error about Ebtables and rc.d (rc.d is no longer used by Ubuntu). But ebtables is not auto starting and is empty – so it did not effect the install.

BTW installs took hours…mostly during the make process of the multiple PHPs with litespeed api.

I am glad that I stuck with it…I currently use Cpanel and lightspeed on my public server. Cyberpanel / OLS is a good alternative.

I have been using virtualmin/webmin and have tested others – none of which I wanted to use in a public VPS.

Unfortunately, virtualbox (homelab) is not supported at all, we’ve never tested nor planning to test on such a setup. We support KVM (recommended), OpenVZ and Dedicated servers.

I get that,but this is really not about my hypervisor. I am just testing on an internal server, because I can’t deploy my public websites to something I don’t know.

On ubuntu 18.04 server I have some sort of owner:group issue going on with the tmp directory. I can force open the file permissions and get into the cyberpanel, but that does not survive a reboot.

The missing phpmyadmin was probably a non connect or download. Looking at the installer script there is no checks in the phpmyadmin install section for anything.

As opposed to the the rainloop install section (same general file location) this section checks for every step along the way and would retry a failed download.

Hopefully someone will come across a solution to the 503 error that does not involve Gunicorn since it is no longer used.

red

I get that,,,,but this is really not about my hypervisor. I am just testing on an internal server, because I can't deploy my public websites to something I don't know.

On ubuntu 18.04 server I have some sort of owner:group issue going on with the tmp directory. I can force open the file permissions and get into the cyberpanel, but that does not survive a reboot.

The missing phpmyadmin was probably a non connect or download. Looking at the installer script there is no checks in the phpmyadmin install section for anything.

As opposed to the the rainloop install section (same general file location) this section checks for every step along the way and would retry a failed download.

Hopefully someone will come across a solution to the 503 error that does not involve Gunicorn since it is no longer used.

red

Since I’ve not tested on VirtualBox, I can not say anything. But if it is not working then better go the official route rather then finding hacks to fix things as it will eventually cost you more time.

I did not want to leave this sitting here with misleading information.

In this case the 503 error was due to slow disk access…which is why it effected both new centos and ubuntu installs.

The VM host system was rebuilding the raid stack (hardware) and I did not know it.

The litespeed server was still fast and the occasional times I got into the panel was due to a lull in disk activity not anything I was trying to reset / chmod. Once the raid was no longer degraded the panel was accessible.

Not sure what the time out is set for on LSCPD…but I was able to get other wsgi things working while the raid was degraded by changing the initial timeout in Litespeed.

I am having issued with ubuntu and new website creation, but will open another topic

This worked for me, basically re-installs some of the dependencies and should not cause any damage to your install.

source /usr/local/CyberCP/bin/activate

pip install --ignore-installed -r /usr/local/CyberCP/requirments.txt

deactivate

virtualenv --system-site-packages /usr/local/CyberCP

systemctl restart gunicorn.socket

Regards,

don