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

WAF crash (Segmentation fault)

I have an XG system running 16.05.1 MR-1 that is crashing multiple times per day like this in reverseproxy.log:
[Wed Mar 01 10:18:23.038084 2017] [core:notice] [pid 27556:tid 4147414848] AH00052: child pid 23923 exit signal Segmentation fault (11)

In the most recent two cases the error right before the crash has been this (hostname redacted):

[Wed Mar 01 15:06:05.290637 2017] [form_hardening:error] [pid 17275:tid 3758963520] [client 92.63.91.81:55104] Failed to validate form: Received unhardened form data (1)
[Wed Mar 01 15:06:05.290818 2017] [security2:error] [pid 17275:tid 3758963520] [client 92.63.91.81] ModSecurity: Warning. Match of "eq 0" against "REQBODY_ERROR" required. [file "/content/waf/2.7.3/modsecurity_crs_protocol_violations.conf"] [line "151"] [id "960912"] [rev "1"] [msg "Failed to parse request body."] [data "Multipart parsing error: Multipart: Final boundary missing."] [severity "CRITICAL"] [ver "OWASP_CRS/2.2.7"] [maturity "9"] [accuracy "9"] [tag] [tag] [hostname "hostnameredacted.com"] [uri "/wp-admin/admin-ajax.php"] [unique_id "WLZIrH8AAAEAAEN7DMUAAAAs"]
[Wed Mar 01 15:06:05.291425 2017] [security2:error] [pid 17275:tid 3758963520] [client 92.63.91.81] ModSecurity: Warning. Pattern match "(.*)" at TX:960912-OWASP_CRS/PROTOCOL_VIOLATION/INVALID_REQ-REQBODY_ERROR. [file "/content/waf/2.7.3/modsecurity_crs_inbound_blocking.conf"] [line "26"] [id "981176"] [msg "Inbound Anomaly Score Exceeded (Total Score: 5, SQLi=, XSS=): Last Matched Message: Failed to parse request body."] [data "Last Matched Data: 1"] [hostname "uim.net.au"] [uri "/wp-admin/admin-ajax.php"] [unique_id "WLZIrH8AAAEAAEN7DMUAAAAs"]

In both cases the source IP is the same, but the hostname is different (completely different websites).

I'm wondering if I can turn that particular protection rule off, but that seems like a pretty crap workaround, and also with only two occurrences logged it's hard to know if this crash was just a coincidence that it happened to land on that particular rule.

I have a ticket open with Sophos Support but haven't heard back since I last emailed on the 23rd, despite this being a high priority issue for us.

Is anyone else seeing such a thing or could offer a suggestion for a workaround?

thanks

James



This thread was automatically locked due to age.
Parents Reply Children