Corrupt Configs
-
@stephenw10 We have a bunch of these out in the wild and so updates don't make it to all in a timely fashion. The last one this happened on was 2.4.4-p1. This one is 2.3.4-p1 but I'd like to wait until I figure this out to update it so that nothing important to the diagnosis gets lost.
-
The only possible way I could see that happening (other than someone actually manually doing that) would be if the box rebooted at that point and it somehow pulled in a new config via ECL which had those packages in it.
Do you have system logs going back to that?
Steve
-
@stephenw10 Is there a way to view the logs from the command line or to export the existing logs to a text file?
-
You can get the complete system log by running:
clog /var/log/system.log > /tmp/systemlog.txtthen downloading /tmp/systemlog.txt.
You can do both those things from Diag > Command Prompt in the webgui.The default log size is 500k though, it may no longer contain that data.
Steve
-
@stephenw10 Interesting. The logs show that the filesystem was full! But right now it shows:
Filesystem Size Used Avail Capacity Mounted on /dev/ufsid/57758f7e2e26ac75 50G 42G 4.6G 90% /I think we're on to something. Is there a way to have the units send an email if the disk gets over a certain percentage used?
-
Not that I'm aware of. I thik Zabbix can do it if you had that configured.
What's using 42GB there though?
Usually that only happens if some package is not behaving, usually failing to rotate it's logs. I would guess it's Suricata.
Steve
-
I see Squid listed as an installed/re-installed package in the first post. The Squid cache can eat up a lot of disk space. The Suricata package did have an issue with log rotation not working properly, but that was fixed a few versions back. So unless you have a really old version of Suricata, its log rotation problem should not be impacting you now.
-
16G ./suricata_igb128922 15G ./suricata_igb252880 31G .igb128922 is an http.log files that goes back to 7/13/2017.
igb252880 is http.log and alerts.log that goes back only to 5/1/2019. -
@bmeeks Yeah, I got bit by the stats log not rotating on some other units. Not sure if this is the same. The Squid cache is 8.7GB as well.
-
The Suricata log rotation fix was this year I believe, but I don't remember precisely when. One aggravation with the Suricata binary is that not all of the log output options have built-in size limits. So I have to use a separate outside process to try and trim the files. What was happening is that Suricata would continue to write to the old file handle and thus the old file would continue to grow even though a new one was started.
-
That probably never made a 2.3.X package though....
Make sure you set a total log directory size limit as well as the individual log sizes.
Steve
-
Yeah, that fix was definitely for pfSense-2.4.x.
-
I've updated to 2.4.4-p3 but the available packages page is blank:
-
Try running
pkg updatesee what error it gives.Steve
-
Updating pfSense-core repository catalogue... pkg: Repository pfSense-core load error: access repo file(/var/db/pkg/repo-pfSense-core.sqlite) failed: No such file or directory pkg: https://pkg.pfsense.org/pfSense_v2_4_4_amd64-core/meta.txz: Network is unreachable repository pfSense-core has no meta file, using default settings pkg: https://pkg.pfsense.org/pfSense_v2_4_4_amd64-core/packagesite.txz: Network is unreachable Unable to update repository pfSense-core Updating pfSense repository catalogue... pkg: Repository pfSense load error: access repo file(/var/db/pkg/repo-pfSense.sqlite) failed: No such file or directory pkg: https://pkg.pfsense.org/pfSense_v2_4_4_amd64-pfSense_v2_4_4/meta.txz: Network is unreachable repository pfSense has no meta file, using default settings pkg: https://pkg.pfsense.org/pfSense_v2_4_4_amd64-pfSense_v2_4_4/packagesite.txz: Network is unreachable Unable to update repository pfSense Error updating repositories!EDIT: WAN is up. I'm connecting out through the unit right now.
-
Can the firewall itself connect out? Ping? DNS lookup?
If you're policy routing it may have lost it's default route coming from 2.3.X.
Make sure you have a default gateway set in System > Routing > Gateways rather than set to 'automatic' there.
Steve
-
Using the force command fixed it, though.
/root: pkg update -f Updating pfSense-core repository catalogue... pkg: Repository pfSense-core load error: access repo file(/var/db/pkg/repo-pfSense-core.sqlite) failed: No such file or directory Fetching meta.txz: 100% 944 B 0.9kB/s 00:01 Fetching packagesite.txz: 100% 2 KiB 1.7kB/s 00:01 Processing entries: 100% pfSense-core repository update completed. 7 packages processed. Updating pfSense repository catalogue... Fetching meta.txz: 100% 944 B 0.9kB/s 00:01 Fetching packagesite.txz: 100% 138 KiB 140.9kB/s 00:01 Processing entries: 100% pfSense repository update completed. 518 packages processed. All repositories are up to date. -
Ah, interesting. And it sees packages now?
-
Yes. So far so good. I don't know why it ran out of space while in service but after taking it offline suddenly there is 4.6GB free. Certainly need to clear it out but where did that other space run off to? I wonder...
-
I've also found an oddity in /var/squid/logs:
/var/squid/logs: ls -lah | grep access -rw-r----- 1 squid proxy 0B Aug 13 00:00 access.log -rw-r----- 1 squid proxy 6.1G Jun 8 05:14 access.log.0 -rw-r----- 1 squid proxy 52M May 5 00:00 access.log.1 -rw-r----- 1 squid proxy 0B Apr 25 00:00 access.log.10 -rw-r----- 1 squid proxy 0B Apr 24 00:00 access.log.11 -rw-r----- 1 squid proxy 0B Apr 23 00:00 access.log.12 -rw-r----- 1 squid proxy 0B Apr 22 00:00 access.log.13 -rw-r----- 1 squid proxy 0B Apr 21 00:00 access.log.14 -rw-r----- 1 squid proxy 0B Apr 20 00:00 access.log.15 -rw-r----- 1 squid proxy 0B Apr 19 00:00 access.log.16 -rw-r----- 1 squid proxy 0B Apr 18 00:00 access.log.17 -rw-r----- 1 squid proxy 0B Apr 17 00:00 access.log.18 -rw-r----- 1 squid proxy 0B Apr 16 00:00 access.log.19 -rw-r----- 1 squid proxy 56M May 4 00:00 access.log.2 -rw-r----- 1 squid proxy 0B Apr 15 00:00 access.log.20 -rw-r----- 1 squid proxy 0B Apr 14 00:00 access.log.21 -rw-r----- 1 squid proxy 0B Apr 13 00:00 access.log.22 -rw-r----- 1 squid proxy 0B Apr 12 00:00 access.log.23 -rw-r----- 1 squid proxy 0B Apr 11 00:00 access.log.24 -rw-r----- 1 squid proxy 0B Apr 10 00:00 access.log.25 -rw-r----- 1 squid proxy 0B Apr 9 00:00 access.log.26 -rw-r----- 1 squid proxy 0B Apr 8 00:00 access.log.27 -rw-r----- 1 squid proxy 0B Apr 7 00:00 access.log.28 -rw-r----- 1 squid proxy 0B Apr 6 00:00 access.log.29 -rw-r----- 1 squid proxy 35M May 3 00:00 access.log.3 -rw-r----- 1 squid proxy 9.6M May 2 00:00 access.log.4 -rw-r----- 1 squid proxy 1.9M May 1 00:00 access.log.5 -rw-r----- 1 squid proxy 0B Apr 29 00:00 access.log.6 -rw-r----- 1 squid proxy 0B Apr 28 00:00 access.log.7 -rw-r----- 1 squid proxy 0B Apr 27 00:00 access.log.8 -rw-r----- 1 squid proxy 0B Apr 26 00:00 access.log.9with access.log.0 being 6.1GB in size, would that indicate that it isn't rotating?
