Postgres errors filling logs, high cpu load

Question

Hi there, 
 I realized, that our Firewalls are very slow due to CPU load of 97%-99% caused by filling-up system-log with postgres errors: 
 
 ... 
 2016:07:26-15:14:40 firewall-1 postgres[30097]: [3-1] ERROR: function ins_websecurity(unknown, unknown, unknown, unknown, unknown, unknown, unknown, unknown, unknown, unknown, unknown, unknown, unknown, integer, integer, integer, bigint) does not exist at character 8 
 2016:07:26-15:14:40 firewall-1 postgres[30097]: [3-2] HINT: No function matches the given name and argument types. You might need to add explicit type casts. 
 2016:07:26-15:14:40 firewall-1 postgres[30097]: [3-3] STATEMENT: select ins_websecurity($1, $2, $3, $4, $5, $6, $7, $8, $9, $10, $11, $12, $13, $14::int4, $15::int4, $16::int4, $17::int8) 
 ... 
 
 ...and so on, almost every 3 seconds 
 
 Checking the logs shows, that this error came up after Installing 9.402-7 (on May 13th 2016). Right now the dual SG-230 HA system runs on actual 9.404-5 
 
 Any hints to correct this very quickly? I need more speed right now, because in 2 days, there will be the whole VOIP connect to the firewall... 
 
 tnx and meeow! 
 Clyde

BAlfson · Accepted Answer

Although this will delete the data in the data bases and graphs, Clyde, it's the only way to fix your issue. If you're uncomfortable doing this yourself, ask Sophos Support. As root at the command line: /etc/init.d/postgresql92 rebuild 
 Cheers - Bob

sachingurung · Answer

Hi, 
 This is a known issue, which is taken into consideration in NUTM- 3882. We are waiting for an update from the Dev Team on this instance. 
 Thanks

Clyde · Answer

Hi Bob, 
 
 Good things first: it works again! 
 I made a backup of the current config, took the "Standby-Node" Firewall out of the network and did a factory-reset on it. I planned to reconfigure the machine step-by-step, but then decided, to reinstall the backup i've just made. 
 And lucky me, it works right the way it should: CPU load went down to 6% (up to 40% peaks). Switched the nodes and done the same thing to the seccond appliance. 
 After sync, they both working togehter again as a stable HA-System. Don't know what happened before, but if it works this way: well, fine for me! :-) 
 
 Thanks for the help! meeeow! Clyde