[Oisf-users] Suricata Hungs

[Peter Manev]

On Wed, Dec 5, 2018 at 10:58 AM Michael Tsukanov <zukinzin at gmail.com> wrote:
>
> Yes, it's only inline related.
> I use kill -USR2 `cat /var/run/suricata.pid` for reloading.
>
> I removed whole snort ruleset which caused these errors in log during reload, but it didn't help.
>

So it seems you can reproduce it consistently.
Can you share the output of "suricata --build-info" ? Also when it
happens  - it would be great if you could share stats.log or similar.

Thank you

> пт, 30 нояб. 2018 г. в 10:04, Peter Manev <petermanev at gmail.com>:
>>
>> On Tue, Nov 20, 2018 at 1:19 PM Michael Tsukanov <zukinzin at gmail.com> wrote:
>> >
>> > I didn't think so, it will be a huge file since we have around 200 people in location and we can't predict when it fails. Moreover with passive mode we haven't any issues with suricata
>> >
>>
>> Ok - so this is only inline related.
>> How do you do your rule reloads? (do you use unix socket/hup or cold
>> restart of Suricata?)
>>
>> Thank you
>>
>> > вт, 20 нояб. 2018 г. в 14:24, Peter Manev <petermanev at gmail.com>:
>> >>
>> >> On Tue, Nov 20, 2018 at 11:51 AM Michael Tsukanov <zukinzin at gmail.com> wrote:
>> >> >
>> >> > Yes, these errors are related to the rules from snort rulesset (which is "not optimized" for suricata)
>> >> > But we also have locations where suricata work fine with these rules...
>> >>
>> >> In that case it seems it is related to some traffic condition  - is it
>> >> possible to narrow it down to a pcap ?
>> >>
>> >> > I'll try to use ET only, but I would like to have some "hooks" if it will fails again...
>> >> >
>> >> > вт, 20 нояб. 2018 г. в 12:29, Peter Manev <petermanev at gmail.com>:
>> >> >>
>> >> >> On Mon, Nov 19, 2018 at 7:08 PM Michael Tsukanov <zukinzin at gmail.com> wrote:
>> >> >> >
>> >> >> > Hi Peter,
>> >> >> > Yes we also had the same with 4.1.0 and rolled back to 4.0.5
>> >> >> >
>> >> >> > Stats.log - https://pastebin.com/sKmLwVJP
>> >> >> > Suricata.log - https://pastebin.com/q9Z3z0Zg
>> >> >> > suricata.yaml - https://pastebin.com/EEGHz4M4
>> >> >> > start line: /usr/local/bin/suricata -D --netmap --pidfile /var/run/suricata.pid -c /usr/local/etc/suricata/suricata.yaml
>> >> >> >
>> >> >> > no any unusual rules are triggered in that moment
>> >> >> > We use 114 alert and 6357 drop rules from Snort ruleset and 7314 alert and 3626 drop rules from ET rulesset + 1929 IP addresses from reputations lists
>> >> >> >
>> >> >>
>> >> >> It seems there are a lot of errors during loading similar to  -
>> >> >>
>> >> >> 19/11/2018 -- 02:23:07 - <Error> - [ERRCODE:
>> >> >> SC_ERR_RULE_KEYWORD_UNKNOWN(102)] - unknown rule keyword
>> >> >> 'http_raw_cookie'.
>> >> >> 19/11/2018 -- 02:23:07 - <Error> - [ERRCODE:
>> >> >> SC_ERR_INVALID_SIGNATURE(39)] - error parsing signature "alert tcp
>> >> >> $EXTERNAL_NET any -> $HOME_NET $HTTP_PORTS (msg:"SERVER-WEBAPP
>> >> >> Multiple products DVR admin password leak attempt";
>> >> >> flow:to_server,established; content:"/device.rsp"; fast_pattern:only;
>> >> >> http_uri; content:"uid="; http_raw_cookie; content:"cmd=list";
>> >> >> metadata:policy balanced-ips drop, policy max-detect-ips drop, policy
>> >> >> security-ips drop, service http; reference:cve,2018-9995;
>> >> >> classtype:web-application-attack; sid:46825; rev:1;)" from file
>> >> >> /usr/local/etc/suricata/rules/snort.rules at line 11386
>> >> >>
>> >> >> It may be somehow be related to some rules maybe -  but you say in
>> >> >> af-packet you may have  a problem once every two moths or so.
>> >> >> Is it possible to narrow it down  a bit - for example - load ET only
>> >> >> rules and see if any difference?
>> >> >>
>> >> >> Thank you
>> >> >>
>> >> >> > Sorry, I can't provide the details for AF_PACKETS right now - it may works for 1-2 months without any issues and restarts
>> >> >> >
>> >> >> >
>> >> >> >
>> >> >> > пн, 19 нояб. 2018 г. в 20:36, Peter Manev <petermanev at gmail.com>:
>> >> >> >>
>> >> >> >> On Mon, Nov 19, 2018 at 6:35 PM Peter Manev <petermanev at gmail.com> wrote:
>> >> >> >> >
>> >> >> >> >
>> >> >> >> > On Mon, Nov 19, 2018 at 6:25 PM Michael Tsukanov <zukinzin at gmail.com> wrote:
>> >> >> >> > >
>> >> >> >> > > Friends,
>> >> >> >> > > we've faced an issue with suricata running in inline mode.
>> >> >> >> > >
>> >> >> >> > > Could you please help us to find the root cause of the issue or determinate any useful  metrics which we may use for investigation.
>> >> >> >> > >
>> >> >> >> > > It may works 1-3 days, then we loose the access to switch behind the Suricata and Internet in the office.
>> >> >> >> > >
>> >> >> >> >
>> >> >> >> > Is it possible some rule triggers that condition ?
>> >> >> >> >
>> >> >> >> > > Suricata is placed between ASA and root switch
>> >> >> >> > > We use FreeBSD 11.2, Suricata 4.0.5 with Netmap (but also faced this situation with Ubuntu and AF_Packets in other location). The server has I350 Ethernet adapters, 16Gb RAM, i5 cpu.
>> >> >> >> >
>> >> >> >> > Could you share a bit more information with regards to the set up (ex config/start line etc...) and logs when that hapens - stats.log/suricata.log - for the af-packet set up for example ?
>> >> >> >> >
>> >> >> >>
>> >> >> >> Also (sent out the previous mail too fast - apologies ) - do you have
>> >> >> >> the same problem with Suricata 4.1  ?
>> >> >> >>
>> >> >> >> > > We use one /16 net as HOME_NET in suricata.yaml. The Internet channel is 80Mbps
>> >> >> >> > >
>> >> >> >> > > Thank you in advance
>> >> >> >>
>> >> >> >>
>> >> >> >>
>> >> >> >> --
>> >> >> >> Regards,
>> >> >> >> Peter Manev
>> >> >>
>> >> >>
>> >> >>
>> >> >> --
>> >> >> Regards,
>> >> >> Peter Manev
>> >>
>> >>
>> >>
>> >> --
>> >> Regards,
>> >> Peter Manev
>>
>>
>>
>> --
>> Regards,
>> Peter Manev



-- 
Regards,
Peter Manev