"Clients over time" graph is utterly broken (FTL beta)

Output of pihole -v:

$ pihole -v
  Pi-hole version is v3.3.1-136-ga7e7680 (Latest: v3.3.1)
  AdminLTE version is v3.3-130-g4355bde (Latest: v3.3)
  FTL version is vDev-5ecab0a (Latest: v3.0)

The clients over time graph looks like this:

Things I have done to try to fix the problem:

  1. Flush browser cache, repeatedly (I am aware that this is the go-to solution for graph problems—I have already done it and continue doing it in conjunction with other fixes)
  2. Delete /etc/pihole/pihole-FTL.db and force pihole to create a new database. This fixes the problem for ~12 hours, and then the graph breaks again. The amount of time it takes to break is variable. I can't tell if it breaks based on query count or time.
  3. Changed between boxed mode and wide display mode. No effect on problem behavior.

Other than "rip out and reinstall from scratch," I'm not sure what else to do. Any suggestions would be welcome.

Debug token is 9wf15eeyul.

Could you check the file /var/log/lighttpd/error.log for warnings and errors?

There's only a single entry in there:

2018-06-25 06:25:01: (server.c.1242) logfiles cycled UID = 0 PID = 28041 

The older lighttpd error logs contain some TLS errors for blocked sites, looks like, but I don't know if that's relevant. Lots of lines similar to this:

2018-06-24 09:45:08: (network.c.117) SSL: no certificate/private key for TLS server name id.google.com 
2018-06-24 09:45:08: (connections.c.305) SSL: 1 error:1412E0E2:SSL routines:ssl_parse_clienthello_tlsext:clienthello tlsext 
2018-06-24 09:45:08: (connections.c.305) SSL: 1 error:1408A0E3:SSL routines:ssl3_get_client_hello:parse tlsext 
2018-06-24 09:45:08: (connections.c.1702) SSL: 1 -1 error:140E0197:SSL routines:SSL_shutdown:shutdown while in init 
2018-06-24 09:46:22: (network.c.117) SSL: no certificate/private key for TLS server name id.google.com 
2018-06-24 09:46:22: (connections.c.305) SSL: 1 error:1412E0E2:SSL routines:ssl_parse_clienthello_tlsext:clienthello tlsext 
2018-06-24 09:46:22: (connections.c.305) SSL: 1 error:1408A0E3:SSL routines:ssl3_get_client_hello:parse tlsext 
2018-06-24 09:46:22: (connections.c.1702) SSL: 1 -1 error:140E0197:SSL routines:SSL_shutdown:shutdown while in init 
2018-06-24 09:51:49: (network.c.117) SSL: no certificate/private key for TLS server name id.google.com 
2018-06-24 09:51:49: (connections.c.305) SSL: 1 error:1412E0E2:SSL routines:ssl_parse_clienthello_tlsext:clienthello tlsext 
2018-06-24 09:51:49: (connections.c.305) SSL: 1 error:1408A0E3:SSL routines:ssl3_get_client_hello:parse tlsext 
2018-06-24 09:51:49: (connections.c.1702) SSL: 1 -1 error:140E0197:SSL routines:SSL_shutdown:shutdown while in init

But regardless, no current entries in the lighttpd error log—the problem behavior is ongoing, but the last error message was from yesterday AM.

Just a try:

change line 40

$out = fgets($socket, 128);

in /var/www/html/admin/scripts/pi-hole/php/FTL.php to

$out = fgets($socket, 256);

Clear the browser cache and reload the website.

If it doesn't work revert the above file change.

2 Likes

Ahh, excellent! That appears to have done the trick:

The graph is now readable. Thank you!

1 Like

There is already a fix

2 Likes

Yeah, it's doing it again—once I get a few days of data, the graph eats itself:

I'll just go back in and remove the max line length as suggested.

edit - yep, that fixed the issue:

1 Like

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.