Yesterday I could log into reForis with no problems. Today, I can’t.
I get:
“503 Service Unavailable”
I have no problem logging into LuCI, but I can’t get into reForis.
I have no idea what may have happened between
then and now to cause the change in access.
Would anyone have any clues as to
where I might begin to look to remedy this?
Thanks.
I am also seeing this, wwa working yesterday afternoon, now 503. Reforis is running. Luci is working fine.
So it’s running, but you can’t access it? (Since you get the 503 error?) How do you know it’s running?
/etc/init.d/foris-controller status
I’m afraid that’s not quite right way to check status of reforis.
This only tells you that foris-controller (reforis backend) is running.
To have reforis completely operational, you need foris-controller, lighttpd and reforis itself running.
To be precise, there is no separate service for reforis, because reforis is started by lighttpd as a FastCGI process. Therefore (re)starting reforis means (re)starting lighttpd and checking status of reforis itself is a slightly more complicated.
This looks like an issue involving lighttpd (essentially a proxy).
Could you please try restarting lighttpd and check again?
And if that doesn’t help, could you post content of /var/log/lighttpd/error.log?
Actually, the same thing happened to me yesterday on one router. So if there are any clues that would help to pinpoint the culprit of this behavior, we would be happy to investigate it.
I have had the error 503 issue many times but can usually get past it by refreshing the page or closing the tab and trying again.
I have seen the same issue on many other webpages and logins lately and believe this is a Chrome browser issue.
All the Chrome based browsers are wanting to use HTTPS everywhere and also their so called secure DNS.
Turning off these “security” features in the browser settings has so far stopped the webpage / login problem.
@AdrianH:
It’s happening to me in Firefox and Brave.
/etc/init.d/lighttpd start
did not enable me to get past the 503 reForis error.
(Line spacing below added for readability)
root@Turris_Omnia:~# cat /var/log/lighttpd/error.log
2023-03-12 19:58:30: (../src/server.c.1588) server started (lighttpd/1.4.67)
2023-03-12 20:07:52: (../src/gw_backend.c.528) connect /tmp/fastcgi.turris_auth.socket-0: Connection refused
2023-03-12 20:08:10: (../src/gw_backend.c.528) connect /tmp/fastcgi.reforis.socket-0: Connection refused
2023-03-12 20:22:20: (../src/gw_backend.c.283) establishing connection failed: socket: unix:/tmp/fastcgi.turris_auth.socket-0: Connection refused
2023-03-12 20:22:20: (../src/gw_backend.c.993) all handlers for /reforis/? on /reforis are down.
2023-03-12 20:22:22: (../src/gw_backend.c.360) gw-server re-enabled: unix:/tmp/fastcgi.turris_auth.socket-0 0 /tmp/fastcgi.turris_auth.socket
2023-03-13 11:24:31: (../src/gw_backend.c.283) establishing connection failed: socket: unix:/tmp/fastcgi.turris_auth.socket-0: Connection refused
2023-03-13 11:24:31: (../src/gw_backend.c.993) all handlers for /reforis/? on /reforis are down.
2023-03-13 11:24:33: (../src/gw_backend.c.360) gw-server re-enabled: unix:/tmp/fastcgi.turris_auth.socket-0 0 /tmp/fastcgi.turris_auth.socket
2023-03-13 11:25:01: (../src/gw_backend.c.283) establishing connection failed: socket: unix:/tmp/fastcgi.turris_auth.socket-0: Connection refused
2023-03-13 11:25:01: (../src/gw_backend.c.993) all handlers for /reforis/? on /reforis are down.
[ and so on]
[snip]
Same issue here on a freshly installed Omnia with Turris OS 6.2.4. Re-flashed with the latest medkit, performed initial (basic) setup, and installed Sentinel and NetMetr. Everything was working as expected… until the next day!
The applications dashboard and LuCI are accessible, but reForis gives 503 Service Unavailable. The lighttpd log shows similar messages as posted above by @watchpocket.
The problem disappeared after rebooting the router. It’s not a browser issue (tested with Firefox 111.0). Let’s hope it doesn’t reappear tomorrow!
Hello there,
are you all guys using current HBS TOS 6.2.4? Thanks in advance!
I had the same issue. I fixed it by ssh to the router and run pkgupdate
| Device | Turris Omnia |
|---|---|
| reForis version | 1.4.1 |
| Turris OS version | 6.2.4 |
| Turris OS branch | HBS |
| Kernel version | 5.15.94 |
# uptime
... up 2 days, 2:15...
More than 2 days since last reboot and no issue yet!
root@Turris_Omnia:~# pkgupdate
INFO:Target Turris OS: 6.2.4
WARN:Requested package reforis-snapshots-plugin-l10n-de that is missing, ignoring as requested.
WARN:Requested package reforis-netmetr-plugin-l10n-de that is missing, ignoring as requested.
WARN:Requested package reforis-diagnostics-plugin-l10n-de that is missing, ignoring as requested.
WARN:Requested package reforis-haas-plugin-l10n-de that is missing, ignoring as requested.
WARN:Requested package reforis-data-collection-plugin-l10n-de that is missing, ignoring as requested.
root@Turris_Omnia:~#
root@Turris_Omnia:~#
When I go to:
uci/admin/system/opkg
and try to manually install reforis-snapshots-plugin,
I see this:
Executing package manager
Package pakon version 1.2.1-1 has no valid architecture, ignoring.
Package ucollect-config version 117 has no valid architecture, ignoring.
Package openvpn-openssl version 2.4.6-2 has no valid architecture, ignoring.
Package https-cert version 3 has no valid architecture, ignoring.
Package reforis-snapshots-plugin (1.2.1-3.9-4) installed in root is up to date.
I don’t know what to do from here, but I still see:
503 Service Unavailable
when I try to bring up reForis.
It’s a dumb suggestion, I know, but you could simply try reboot! This seems to have fixed the issue somehow on my router…
Maybe @MadSpindel updated some package that also required a reboot?!? 
Not at dumb suggestion at all, and it worked. In fact I meant to reboot the router immidiately after posting the above but got side-tracked.
Even if it hadn’t worked, your suggestion isn’t a dumb one.
Thanks!