Snort inilization failure
-
I didn't see any different behavior until I updated to the latest snapshot. Then snort would get stuck in an infinite initialization loop.
Thats funny. I haven't changed anything… All that the new image contains is FreeBSD 6.2.
-
I can confirm that under the new version, the results are exactly the same.
After a while snort exits with a core dump (it doesn't reload)
-
I can still provide a configuration file to help out, if you wish .
-
I had the same problem for awhile….snort always ending in a core dump. (I have 1 gig of ram installed)...What finally did the trick for me was switching from ac sparsebands to just acs.......no more core dumping. Has been up for the past week with no problems.
-
I had the same problem for awhile….snort always ending in a core dump. (I have 1 gig of ram installed)...What finally did the trick for me was switching from ac sparsebands to just acs.......no more core dumping. Has been up for the past week with no problems.
;D …and after all that, it looks like acs is the key. certainly been up now for longer than normal, AND it's consuming less memory too
-
I changed my snort performance to acs as well. I am still seeing snort failing to bootup completely immediately after a reboot. However, it will start the initialization process over once it fails. On the second initialization, Snort will successfully boot up.
So I've done some more testing.
Using the performance modes ac, ac-std, or ac-sparebands I get snort initialization failures and core dumps when I reboot. I cannot get snort to boot until I click save in settings.
If I use acs or lowmem, they appear to boot up fine.
Question is, why?
-
It's funny I couldn't figure it either, I tried them all but acs worked consistently well. I'm using about 10% less ram also. The alerts are also triggering consistently better also. I'm running snort+squid, I wonder if all this is related to squid. Did anybody have this problem(core dumping) with just snort installed and not squid?
-
Yep, me. I've only been running SNORT.
Although I will add Squid as soon as it becomes available again.
acs uses significantly less memory for me too and also works as intended. Thank coldfusion
-
Question is, why?
No idea. This is now a SNORT issue and this should be asked on their lists.
-
I think the latest version of SNORT (2.6.1.1) fixes this issue
-
Not sure if this fixes our issue.
This is the issue solved in 2.6.1.1: "Fix problem with snort using high CPU and reprocessing the same rebuilt packets at session end or ACK in middle of packet when there are gaps in the packet sequence."
-
certainly seems to :)
Running in sparsebands now, only been up for ~5 minutes, but that's ~3 more than I've managed before
-
I updated to the 2.6.1.1 versions and now I cant even get Snort to boot up on any performance setting. I am getting brand new errors now:
Nov 24 14:53:39 snort2c[1571]: unable to open alertfile - exit
Nov 24 14:53:39 snort2c[1571]: unable to open alertfile - exit
Nov 24 14:53:39 snort2c[1571]: snort2c running in daemon mode pid: 1571
Nov 24 14:53:39 snort2c[1571]: snort2c running in daemon mode pid: 1571I am running the 1.0.1-SNAPSHOT-11-24-2006
-
Oh bugger. I'll check it out a bit later tonite.
-
Thanks Scott.
-
Didn't do that for me, BUT it also didn't trigger any alerts
-
Try running a port scan from http://www.grc.com. You should see it appear in alert as a ping.
-
I'm pretty much screwed with the same alert after I upgraded SNORT. Now the service will not start and stay running.
-
Scott, I see where you reverted snort back to the 2.6.0.2.5. I reverted back to that version and I'm still getting the same error. The trouble must be in the latest snapshot.
-
Try running a port scan from http://www.grc.com. You should see it appear in alert as a ping.
I know, and it doesn't.
I've reverted back to the older version and it still doesn't raise any alerts at all ???