Guest User!

You are not Sophos Staff.

This discussion has been locked.
You can no longer post new replies to this discussion. If you have a question you can start a new discussion

IPS Engine Dead

Hi all.

I've been running SFOS on an SG125w here, and i've noticed my IPS engine is dead, I cannot restart it and reboots don't change anything.

If I try and conduct a Pattern Update, the log shows the IPS update downloading, pass its checksum but fail to install.

Here's a snippet of the U2D.log

<code>

SG125w_XN02_SFOS 15.01.0# tail -f u2d.log
DEBUG     Nov 13 14:58:39 [2458]: Response length : 401
DEBUG     Nov 13 14:58:39 [2458]: Received name : ips_10.0_3.12.48.tar.gz.gpg
DEBUG     Nov 13 14:58:39 [2458]: Received location : d30ncyzaneb4q0.cloudfront.net/ips_10.0_3.12.48.tar.gz.gpg
DEBUG     Nov 13 14:58:39 [2458]: Received version : 3.12.48
DEBUG     Nov 13 14:58:39 [2458]: Received size : 946452
DEBUG     Nov 13 14:58:39 [2458]: Received md5sum : 3a028f4f6cdb3e43c8f56732ed3dfed6
DEBUG     Nov 13 14:58:39 [2458]: Received module : ips
DEBUG     Nov 13 14:58:39 [2458]: Received cv : 10.0
DEBUG     Nov 13 14:58:39 [2458]: Received type : full
Fri Nov 13 14:59:30 2015 Starting download for file ips_10.0_3.12.48.tar.gz.gpg
Fri Nov 13 15:00:30 2015 Download completed for file ips_10.0_3.12.48.tar.gz.gpg
gpg: Signature made Fri Nov  6 00:12:45 2015 EST using RSA key ID 6A20EB0B
gpg: NOTE: trustdb not writable
gpg: Good signature from "Sophos Up2Date Server <updates@sophos.com>"
Fri Nov 13 15:00:30 2015 Download for file ips_10.0_3.12.48.tar.gz.gpg passed integrity and gpg checks
Fri Nov 13 15:00:30 2015 Current ips patterns are at /content/ips_10.0/3.12.38
Fri Nov 13 15:00:30 2015 New updated  patterns are now at /content/ips_10.0/3.12.48
Fri Nov 13 15:00:44 2015 Callback u2d_pt_installed failed for ips, version = 3.12.48.
Fri Nov 13 15:00:44 2015 Setting status 'fail' in DB and reverting link for ips to old version = 3.12.38.
Fri Nov 13 15:00:44 2015 ips patterns are again at /content/ips_10.0/3.12.38

</code>

Here's a snippet of the IPS.log that is pointing to an 'Invalid CPU Number'.

<code>

INFO[4163]:Nov 11 15:38:03:spo_alert_garnersock.c:250:GarnerStart:Start called
INFO[4163]:Nov 11 15:38:03:spo_alert_garnersock.c:291:GarnerStart:fd 3 host 127.0.0.1 sport 2929 dport                                                                                             198
INFO[4163]:Nov 11 15:38:03:snort.c:1574:main:SnortInit() done
INFO[4163]:Nov 11 15:38:03:snort.c:1605:main:Snort master started
INFO[4163]:Nov 11 15:38:03:snort.c:434:get_cpu_cnt:got cores = 2 from '/proc/interrupts'
INFO[4163]:Nov 11 15:38:03:snort.c:511:put_cpulist:cpunum=2 snort_cpulist = 2
fd 5 size 528384
size 1022 maxapp 4096 counter 2 bytesize 512
IPS: invalid cpu number 2
ERROR[4161]:Nov 11 15:38:03:snort.c:1045:notify_newmaster:read_full failed: ret: 0 Success
INFO[4161]:snort.c:2873:child_handler:child 4163 dead
INFO[4161]:Nov 11 15:38:03:snort.c:1009:kill_snortmaster:exited(4163): exited, status=255
ERROR[4161]:Nov 11 15:38:03:snort.c:1566:main:Snort exited with '-1'
INFO[4161]:Nov 11 15:38:03:snort.c:2936:CleanExit:killing snort master
svc_set_status: svc_init not done 

</code>

Anyone else seen this?



This thread was automatically locked due to age.
  • have fixed this..

    I needed to clear the IPS instances within the console of the XG. Using the command ‘set ips ips-instance clear’ I was able to purge the config that had 1 ip instance with 2 CPU.

    After the clear i then set a new IPS Instance using the command ‘set ips ips-instance add IPS CPU 1'

    Once I did this I was able to start the IPS Engine and then perform and Pattern Update which downloaded the correct sigs.

    I’ll be interested to see of others ‘set ips ips-instance add IPS CPU 1’ - using TAB to autocomplete the line after the CPU and before entering an number, I’d be keen to see what CPU configuration it allows… my SG125w says the options are 0 (zero) and 1.

    ==

    When in doubt, Script it out.

  • I was unable to start IPS and receiving the same error in an esxi 6 install configured with 2 virtual sockets and 1 core per socket. After changing the number of virtual sockets to 1 with multiple (4) cores per socket, the error went away and IPS started successfully and updated normally.

    Will

    --------------------
    Sophos UTM / Sophos XG Firewall home user

    Private: Virtual UTM 9.3, ESXi 6.0: 2 vCPUs, 8GB RAM, 120GB vHDD, 3 vNICs

    Private: Virtual Sophos XG Firewall, ESXi 6.0: 1 vCPU, 4vCores, 6GB RAM, 80/4 GB vHDD, 3 vNICs

    ESXi 6.0 Host: SuperMicro C7Z97-OCE, Core I5-4690 3.5 GHz, 32GB RAM, NICS: I217-V, I210, I340-T4

Share Feedback
×

Submitted a Tech Support Case lately from the Support Portal?