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

[9.4xx] Monthly httpproxy segfault

We've been running pre-9.4xx releases on SG450 smoothly since initial deployment. However, after upgrading to 9.402-7 in Jun, unexpected httpproxy segfault comes:

/var/log/kernel/2016/06/kernel-2016-06-15.log.gz:2016:06:15-17:10:43 sophos450-2 kernel: [1582959.343159] httpproxy[6930]: segfault at 80cc3a0 ip 00000000f73f37f9 sp 00000000ecb01f58 error 7 in libglib-2.0.so.0.2200.5[f73df000+c5000]
/var/log/kernel/2016/06/kernel-2016-06-15.log.gz:2016:06:15-17:24:58 sophos450-1 kernel: [1583461.042708] httpproxy[7004]: segfault at 3534314d ip 00000000f748b85e sp 00000000ee78ff60 error 4 in libglib-2.0.so.0.2200.5[f745f000+c5000]

And after we upgraded to 9.404-5, it keeps coming:

/var/log/kernel/2016/07/kernel-2016-07-27.log.gz:2016:07:27-09:04:32 sophos450-2 kernel: [2212858.251975] EpollWorker_25[8452]: segfault at 0 ip 00000000f6a864d1 sp 00000000f2b8ff2c error 4 in libc-2.11.3.so[f6a0c000+16c000]

Such segfault is so nasty that it freezes up the web UI, doesn't restart in timely fashion, doesn't trigger HA failover, and prevents reboot in web UI or SSH. We have to physically attend to the device to reboot. We've contacted support and they've collected some logs already. But my gut feeling tells me that they might not come up with solution soon.

Not sure if anybody have the similar experience on this?



This thread was automatically locked due to age.
Parents Reply Children
  • As I said, I've upgraded to 9.404-5 already and got another segfault. I am not able to find NUTM-3978 you mentioned in the release notes of 9.403-4 or 9.404-5. However, support just told me that it's NUTM-4163, which will be fixed in 9.405 very soon.

    And I was told that I could run the following to restart httpproxy, not sure if it would work next time (I don't hope to see another segfault anyway!)

    /var/mdw/scripts/httpproxy restart

    I'm just curious why HA failover doesn't handle software failure when some daemon like httpproxy is so fragile. [:(]

  • Hi,

    Apologies, I missed to add the second part, NUTM-3978 effected v9.401  and it was resolved in v9.404. However, later the fix NUTM-4163 effected v9.404 and we are chasing the developers to get a solid fix for this.

    Thanks

    Sachin Gurung
    Team Lead | Sophos Technical Support
    Knowledge Base  |  @SophosSupport  |  Video tutorials
    Remember to like a post.  If a post (on a question thread) solves your question use the 'This helped me' link.

  • That's why I am feeling that httpproxy is far unstable than pre-9.4 [:(]