Crash Reporter
-
Error from pfSense:
[23-Jan-2024 10:14:58 America/Chicago] PHP Fatal error: str_ireplace(): Cannot use output buffering in output buffering display handlers in /usr/local/www/csrf/csrf-magic.php on line 165
This may have been related to my creation of a list using "host" when it should have been "network". I have removed the errored list, and I do not know if this was the cause.
I will be watching for future occurrences of this.
-
Crash report begins. Anonymous machine information:
arm64
14.0-CURRENT
FreeBSD 14.0-CURRENT aarch64 1400094 #1 plus-RELENG_23_09_1-n256200-3de1e293f3a: Wed Dec 6 20:59:18 UTC 2023 root@freebsd:/var/jenkins/workspace/pfSense-Plus-snapshots-23_09_1-main/obj/aarch64/8ra4gn87/var/jenkins/workspace/pfSense-Plus-snapshots-23_Crash report details:
PHP Errors:
[23-Jan-2024 10:14:58 America/Chicago] PHP Fatal error: str_ireplace(): Cannot use output buffering in output buffering display handlers in /usr/local/www/csrf/csrf-magic.php on line 165No FreeBSD crash data found.
-
That's the full error given?
-
@stephenw10 Yes, afraid so! Without knowledge of what this section of code is doing, not sure what to make of it!
I was assuming from the text that it may be related to creating a backup snapshot XML. I was hoping someone would know!
-
Hmm, that's usually a secondary error such as shown here: https://redmine.pfsense.org/issues/14513
Have you seen this more than once? Can you trigger it deliberately?
-
@stephenw10 So far, this has only happened once and I do not know the cause.
-
Anything CSRF related is usually caused by the user session expiring and refreshing a page. Is it possible that happened>
-
@stephenw10 This was shortly after a reboot, so I would assume this is a possibility.
-
@AudioDave said in Crash Reporter:
@stephenw10 This was shortly after a reboot, so I would assume this is a possibility.
Is your BIOS battery/clock OK?
CSRF tokens are time-based, so if you loaded the GUI "too soon" at boot before NTP had a chance to sync the clock, and then the clock had a significant jump (hours/days), the tokens would appear to have expired immediately.
-
@jimp Good question! The unit is new from netgate, but I have been seeing some warning about lack of sync, even though the time appears correct.
I suppose there's always a chance the battery is old.
-
@AudioDave said in Crash Reporter:
@jimp Good question! The unit is new from netgate, but I have been seeing some warning about lack of sync, even though the time appears correct.
I suppose there's always a chance the battery is old.
Which model?
Some of them don't have an RTC battery so they rely on a clock sync every boot. IIRC that's just certain ARM devices though, not anything recent that's amd64.
-
@jimp Netgate 2100. I have gone ahead and replaced the battery.
The warning I've been seeing is in the NTP System Log:
Jan 23 10:14:11 ntpd 5623 kernel reports TIME_ERROR: 0x41: Clock Unsynchronized
I think this is happening on reboot, though I was NOT removing power, so seems unlikely that the battery is the cause. I had a spare, so I replaced it anyway.
-
That's a normal message at boot when ntpd first starts and hasn't yet decided it's configured sources are valid.
