[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

Jim Jagielski  changed:

   What|Removed |Added

 Status|NEEDINFO|RESOLVED
 Resolution|--- |FIXED

--- Comment #44 from Jim Jagielski  ---
Fixed in truck and will be in 2.4.26

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

not@gmail.com changed:

   What|Removed |Added

 CC||not@gmail.com

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #43 from Endika  ---
Hello,

Something new about when this patch will be included in the next release?

Best regards

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #42 from Endika  ---
Hello,

When will the solution be included in a release?

Thanks

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

Szőgyényi Gábor  changed:

   What|Removed |Added

 CC||szg0...@freemail.hu

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

cbarb...@okta.com changed:

   What|Removed |Added

 CC||cbarb...@okta.com

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

Re: [Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[Jim Jagielski]

Cool!

Yann, I see this as a show-stopper for 2.4.26... Do you
agree? Would you like to propose this (and the 2 other
commits) for backport to close this bug.

Thx!

> On Jan 20, 2017, at 2:21 AM, bugzi...@apache.org wrote:
> 
> https://bz.apache.org/bugzilla/show_bug.cgi?id=60071
> 
> --- Comment #41 from Endika  ---
> Hello!
> 
> Great news!!! With the last patch after 21 hours and 38 tests i've got no
> errors!!!
> 
> So it seems to be fixed!!
> 
> Good work!!!
> 
> Thanks
> 
> -- 
> You are receiving this mail because:
> You are the assignee for the bug.
> -
> To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
> For additional commands, e-mail: bugs-h...@httpd.apache.org
> 


-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #41 from Endika  ---
Hello!

Great news!!! With the last patch after 21 hours and 38 tests i've got no
errors!!!

So it seems to be fixed!!

Good work!!!

Thanks

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #40 from Endika  ---
Hi!!

After 18 hours and 305000 requests i haven't obtained any error with original
module and "ProxyHCTPsize 0"

Now I'll prepare the test with the last patch

Thanks!!!

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #39 from Jim Jagielski  ---
My VERY simple testing looks good!

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

Yann Ylavic  changed:

   What|Removed |Added

  Attachment #34620|0   |1
is obsolete||

--- Comment #38 from Yann Ylavic  ---
Created attachment 34642
  --> https://bz.apache.org/bugzilla/attachment.cgi?id=34642=edit
Threads safety + single pool/bucket_alloc

The previous patch did not handle APR_EOF correctly while reading response
body, now fixed.

After the test with unpatched 2.4.25 and "ProxyHCTPsize 0" (proposed by Jim) to
valide that multithreading is the issue, could you please try this new patch
(with default/no ProxyHCTPsize this time) to verify that multithreading would
work (without locking)?

Multithreading is nice when (some) backends start to respond slowly/timeout.
BTW, we probably shouldn't start a new check for a backend that did not finish
with the previous one (should its timeout be above the watchdog's pediod), but
that's a different story we won't ask you to test here ;)

Thanks Endika for all this testing.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #37 from Endika  ---
Ok, i am making tests with original module and 

ProxyHCTPsize 0

Thanks

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #36 from Jim Jagielski  ---
OK... I see that it was moved from hc_check_http and hc_check_tcp to the actual
baton creation.

I still wonder if we are overcomplicating things by such a drastic
restructuring... having a hard time groking the various changes :)

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #35 from Yann Ylavic  ---
(In reply to Jim Jagielski from comment #33)
> Hold on... These patches are wrong since they prevent the health checks from
> being changed via the balancer-manager application (because hc_get_hcworker
> is only called once per start...

Hmm, could you elaborate?

hc_get_hcworker() is called for each worker on each watchdog run, this part did
not change AFAICT.

BTW, it looks like the "method" only can be changed with the manager, always
been so...

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #34 from Jim Jagielski  ---
If possible, go back to the original, unmatched version of the module and see
if setting ProxyHCTPsize 0 resolves the problem.

If so, then that conclusively indicates a thread concurrent issue.

Thanks!

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #33 from Jim Jagielski  ---
Hold on... These patches are wrong since they prevent the health checks from
being changed via the balancer-manager application (because hc_get_hcworker is
only called once per start...

I think we need to rethink this...

Using the orig code as a start, can we just create a mutex to make the allocs
thread-safe now? That seems like a "cleaner" solution, assuming that that is
the root cause.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #32 from Endika  ---
Double checked the patch applied and it's the correct one

These are the logs with LogLevel trace7

[Wed Jan 18 16:20:58.439498 2017] [proxy_hcheck:trace2] [pid 20265:tid
139992545253120] mod_proxy_hcheck.c(912): Checking balancer://cluster_namenode
worker: http://ejld1131.ejgvdns:8170  [4] (18e4c70)
[Wed Jan 18 16:20:58.439541 2017] [proxy_hcheck:debug] [pid 20265:tid
139992445744896] mod_proxy_hcheck.c(802): AH03256: Threaded Health checking
http://ejld1132.ejgvdns:8170
[Wed Jan 18 16:20:58.439552 2017] [proxy:debug] [pid 20265:tid 139992445744896]
proxy_util.c(2156): AH00942: HCOH: has acquired connection for
(ejld1132.ejgvdns)
[Wed Jan 18 16:20:58.439724 2017] [proxy:trace2] [pid 20265:tid
139992445744896] proxy_util.c(2850): HCOH: fam 2 socket created to connect to
ejld1132.ejgvdns
[Wed Jan 18 16:20:58.439773 2017] [proxy_hcheck:debug] [pid 20265:tid
139992437352192] mod_proxy_hcheck.c(802): AH03256: Threaded Health checking
http://ejld1131.ejgvdns:8170
[Wed Jan 18 16:20:58.439819 2017] [proxy:debug] [pid 20265:tid 139992445744896]
proxy_util.c(2884): AH02824: HCOH: connection established with
10.190.40.107:8170 (ejld1132.ejgvdns)
[Wed Jan 18 16:20:58.439827 2017] [proxy:debug] [pid 20265:tid 139992437352192]
proxy_util.c(2156): AH00942: HCOH: has acquired connection for
(ejld1131.ejgvdns)
[Wed Jan 18 16:20:58.439841 2017] [proxy:debug] [pid 20265:tid 139992445744896]
proxy_util.c(3051): AH00962: HCOH: connection complete to 10.190.40.107:8170
(ejld1132.ejgvdns)
[Wed Jan 18 16:20:58.439859 2017] [proxy_hcheck:trace7] [pid 20265:tid
139992445744896] mod_proxy_hcheck.c(609): GET /webhdfs/v1/data?op=GETFILESTATUS
HTTP/1.0\r\nHost: ejld1132.ejgvdns:8170\r\n\r\n
[Wed Jan 18 16:20:58.439870 2017] [core:trace6] [pid 20265:tid 139992445744896]
core_filters.c(525): [remote 10.190.40.107:8170] core_output_filter: flushing
because of FLUSH bucket
[Wed Jan 18 16:20:58.439933 2017] [proxy:trace2] [pid 20265:tid
139992437352192] proxy_util.c(2850): HCOH: fam 2 socket created to connect to
ejld1131.ejgvdns
[Wed Jan 18 16:20:58.440223 2017] [proxy:debug] [pid 20265:tid 139992437352192]
proxy_util.c(2884): AH02824: HCOH: connection established with
10.190.40.106:8170 (ejld1131.ejgvdns)
[Wed Jan 18 16:20:58.440260 2017] [proxy:debug] [pid 20265:tid 139992437352192]
proxy_util.c(3051): AH00962: HCOH: connection complete to 10.190.40.106:8170
(ejld1131.ejgvdns)
[Wed Jan 18 16:20:58.440276 2017] [proxy_hcheck:trace7] [pid 20265:tid
139992437352192] mod_proxy_hcheck.c(609): GET /webhdfs/v1/data?op=GETFILESTATUS
HTTP/1.0\r\nHost: ejld1131.ejgvdns:8170\r\n\r\n
[Wed Jan 18 16:20:58.440282 2017] [core:trace6] [pid 20265:tid 139992437352192]
core_filters.c(525): [remote 10.190.40.106:8170] core_output_filter: flushing
because of FLUSH bucket
[Wed Jan 18 16:20:58.443916 2017] [proxy_hcheck:debug] [pid 20265:tid
139992445744896] mod_proxy_hcheck.c(628): AH03254: HTTP/1.1 403 Forbidden
[Wed Jan 18 16:20:58.443962 2017] [proxy_hcheck:trace7] [pid 20265:tid
139992445744896] mod_proxy_hcheck.c(662): Cache-Control: no-cache
[Wed Jan 18 16:20:58.443975 2017] [proxy_hcheck:trace7] [pid 20265:tid
139992445744896] mod_proxy_hcheck.c(662): Expires: Wed, 18 Jan 2017 15:20:58
GMT
[Wed Jan 18 16:20:58.443983 2017] [proxy_hcheck:trace7] [pid 20265:tid
139992445744896] mod_proxy_hcheck.c(662): Date: Wed, 18 Jan 2017 15:20:58 GMT
[Wed Jan 18 16:20:58.443990 2017] [proxy_hcheck:trace7] [pid 20265:tid
139992445744896] mod_proxy_hcheck.c(662): Pragma: no-cache
[Wed Jan 18 16:20:58.444020 2017] [proxy_hcheck:trace7] [pid 20265:tid
139992445744896] mod_proxy_hcheck.c(662): Expires: Wed, 18 Jan 2017 15:20:58
GMT
[Wed Jan 18 16:20:58.444030 2017] [proxy_hcheck:trace7] [pid 20265:tid
139992445744896] mod_proxy_hcheck.c(662): Date: Wed, 18 Jan 2017 15:20:58 GMT
[Wed Jan 18 16:20:58.444034 2017] [proxy_hcheck:trace7] [pid 20265:tid
139992445744896] mod_proxy_hcheck.c(662): Pragma: no-cache
[Wed Jan 18 16:20:58.444045 2017] [proxy_hcheck:trace7] [pid 20265:tid
139992445744896] mod_proxy_hcheck.c(662): Content-Type: application/json
[Wed Jan 18 16:20:58.444085 2017] [proxy_hcheck:trace7] [pid 20265:tid
139992445744896] mod_proxy_hcheck.c(662): Server: Jetty(6.1.26.hwx)
[Wed Jan 18 16:20:58.444103 2017] [proxy:debug] [pid 20265:tid 139992445744896]
proxy_util.c(2171): AH00943: HCOH: has released connection for
(ejld1132.ejgvdns)
[Wed Jan 18 16:20:58.444165 2017] [proxy_hcheck:debug] [pid 20265:tid
139992445744896] mod_proxy_hcheck.c(559): AH03251: Health check GET Status (1)
for 18e4690.
[Wed Jan 18 16:20:58.445331 2017] [proxy_hcheck:debug] [pid 20265:tid
139992437352192] mod_proxy_hcheck.c(628): AH03254: HTTP/1.1 200 OK
[Wed Jan 18 16:20:58.445353 2017] [proxy_hcheck:trace7] [pid 20265:tid
139992437352192] mod_proxy_hcheck.c(662): Cache-Control: no-cache
[Wed Jan 18 16:20:58.445359 2017] [proxy_hcheck:trace7] [pid 20265:tid
139992437352192] 

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #31 from Yann Ylavic  ---
(In reply to Endika from comment #30)
> Applying the last patch 

This attachment https://bz.apache.org/bugzilla/attachment.cgi?id=34620 right?
It's not cumulative, applies directly on 2.4.25's
modules/proxy/mod_proxy_hcheck.c.

> both workers of the balancer are in status (1), but
> one should be ok

Could you set "LogLevel trace7" please?

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #30 from Endika  ---
Applying the last patch both workers of the balancer are in status (1), but one
should be ok

> HTTP/1.1 200 OK



[Wed Jan 18 15:43:51.225390 2017] [proxy:debug] [pid 28953:tid 140683774117632]
proxy_util.c(2884): AH02824: HCOH: connection established with
10.190.40.107:8170 (ejld1132.ejgvdns)
[Wed Jan 18 15:43:51.225389 2017] [proxy_hcheck:debug] [pid 28953:tid
140683782510336] mod_proxy_hcheck.c(802): AH03256: Threaded Health checking
http://ejld1131.ejgvdns:8170
[Wed Jan 18 15:43:51.225440 2017] [proxy:debug] [pid 28953:tid 140683782510336]
proxy_util.c(2156): AH00942: HCOH: has acquired connection for
(ejld1131.ejgvdns)
[Wed Jan 18 15:43:51.225445 2017] [proxy:debug] [pid 28953:tid 140683774117632]
proxy_util.c(3051): AH00962: HCOH: connection complete to 10.190.40.107:8170
(ejld1132.ejgvdns)
[Wed Jan 18 15:43:51.226028 2017] [proxy:debug] [pid 28953:tid 140683782510336]
proxy_util.c(2884): AH02824: HCOH: connection established with
10.190.40.106:8170 (ejld1131.ejgvdns)
[Wed Jan 18 15:43:51.226047 2017] [proxy:debug] [pid 28953:tid 140683782510336]
proxy_util.c(3051): AH00962: HCOH: connection complete to 10.190.40.106:8170
(ejld1131.ejgvdns)
[Wed Jan 18 15:43:51.228620 2017] [proxy_hcheck:debug] [pid 28953:tid
140683774117632] mod_proxy_hcheck.c(628): AH03254: HTTP/1.1 403 Forbidden
[Wed Jan 18 15:43:51.228662 2017] [proxy:debug] [pid 28953:tid 140683774117632]
proxy_util.c(2171): AH00943: HCOH: has released connection for
(ejld1132.ejgvdns)
[Wed Jan 18 15:43:51.228719 2017] [proxy_hcheck:debug] [pid 28953:tid
140683774117632] mod_proxy_hcheck.c(559): AH03251: Health check GET Status (1)
for 226c690.
[Wed Jan 18 15:43:51.230431 2017] [proxy_hcheck:debug] [pid 28953:tid
140683782510336] mod_proxy_hcheck.c(628): AH03254: HTTP/1.1 200 OK
[Wed Jan 18 15:43:51.230481 2017] [proxy:debug] [pid 28953:tid 140683782510336]
proxy_util.c(2171): AH00943: HCOH: has released connection for
(ejld1131.ejgvdns)
[Wed Jan 18 15:43:51.230508 2017] [proxy_hcheck:debug] [pid 28953:tid
140683782510336] mod_proxy_hcheck.c(559): AH03251: Health check GET Status (1)
for 226cc70.

Thanks

Endika

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #29 from Endika  ---
ok, just for clearify

I´m patching the original version of .c file (version 2.4.25). 
orig.c + attach 34601

Trying to do comulative patches, with the second one i obtained an error 
orig.c + attach 34601 + attach 34618

so i patched the original file with the second patch file...

orig.c + attach 34618.

>From that moment i've always made the same proccess. I've patched the original
.c file with the next patch.

What is the exact proccess to follow for patching the files?

Thanks 

Endika

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #28 from Endika  ---
Sorry, my error...i've patched with the wrong code. 

I'll try again

Endika

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #27 from Yann Ylavic  ---
(In reply to Endika from comment #26)
> We have tested the patch today and in less than a hour we've obtained 8
> Segmentation Faults...

I'm not sure which patch you tested, attachment 34620 ?

> 
> 
> Core dump
> 
> (gdb) bt full
[...]
> #1  0x7fd6b7976c03 in hc_send (ctx=0x2085678, ctx=0x2085678,
> backend=0x7fd69c026e80, out=, ptemp=0x7fd6880008e8)
> at mod_proxy_hcheck.c:612
> tmp_bb = 0x7fd688000a60

There is no tmp_bb is the latest patch, which precisely addresses possible
issues with ctx->ba.

Any chance this test was with attachment 34619 or earlier (now obsolete/hidden
above) ?

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #26 from Endika  ---
Hello,
We have tested the patch today and in less than a hour we've obtained 8
Segmentation Faults...

Logs in error log:

[Wed Jan 18 11:02:01.358056 2017] [proxy:debug] [pid 32592:tid 140559909840640]
proxy_util.c(2209): [client 10.136.101.140:62325] AH00944: connecting
http://ejld1131.ejgvdns:8170/webhdfs/v1/data?op=GETFILESTATUS=root to
ejld1131.ejgvdns:8170
[Wed Jan 18 11:02:01.358067 2017] [proxy:debug] [pid 32592:tid 140559909840640]
proxy_util.c(2418): [client 10.136.101.140:62325] AH00947: connected
/webhdfs/v1/data?op=GETFILESTATUS=root to ejld1131.ejgvdns:8170
[Wed Jan 18 11:02:01.359566 2017] [proxy:debug] [pid 32592:tid 140559909840640]
proxy_util.c(2171): AH00943: http: has released connection for
(ejld1131.ejgvdns)
[Wed Jan 18 11:02:01.359590 2017] [proxy_balancer:debug] [pid 32592:tid
140559909840640] mod_proxy_balancer.c(688): [client 10.136.101.140:62325]
AH01176: proxy_balancer_post_request for (balancer://cluster_namenode)
[Wed Jan 18 11:02:01.373675 2017] [core:notice] [pid 29786:tid 140560333223744]
AH00051: child pid 4685 exit signal Segmentation fault (11), possible coredump
in /tmp
[Wed Jan 18 11:02:10.776046 2017] [proxy_balancer:debug] [pid 9243:tid
140560027338496] mod_proxy_balancer.c(688): [client 10.136.101
.140:62342] AH01176: proxy_balancer_post_request for
(balancer://cluster_namenode)
[Wed Jan 18 11:02:10.794876 2017] [watchdog:debug] [pid 3222:tid
140560165148416] mod_watchdog.c(159): AH02972: Singleton Watchdog
(_proxy_hcheck_) running
[Wed Jan 18 11:02:10.794991 2017] [proxy_hcheck:debug] [pid 3222:tid
140560165148416] mod_proxy_hcheck.c(888): AH03258: _proxy_hcheck_ watchdog
started.
[Wed Jan 18 11:02:10.796926 2017] [proxy_hcheck:debug] [pid 3222:tid
140560165148416] mod_proxy_hcheck.c(902): AH03313: apr_thread_pool_create()
with 16 threads succeeded
[Wed Jan 18 11:02:10.979604 2017] [authz_core:debug] [pid 9243:tid
140560027338496] mod_authz_core.c(835): [client 10.136.101.140:62342] AH01628:
authorization result: granted (no directives)
[Wed Jan 18 11:02:12.800263 2017] [proxy:debug] [pid 3222:tid 140560165148416]
proxy_util.c(1856): AH00930: initialized pool in child
 3222 for (ejld1132.ejgvdns) min=0 max=25 smax=25
[Wed Jan 18 11:02:12.800440 2017] [proxy_hcheck:debug] [pid 3222:tid
140560165148416] mod_proxy_hcheck.c(457): AH03248: Creating hc worker 20a0690
for http://ejld1132.ejgvdns:8170
[Wed Jan 18 11:02:12.800479 2017] [proxy:debug] [pid 3222:tid 140560165148416]
proxy_util.c(1779): AH00925: initializing worker 20a0690 shared
[Wed Jan 18 11:02:12.800488 2017] [proxy:debug] [pid 3222:tid 140560165148416]
proxy_util.c(1821): AH00927: initializing worker 20a0690 local
[Wed Jan 18 11:02:12.800513 2017] [proxy:debug] [pid 3222:tid 140560165148416]
proxy_util.c(1856): AH00930: initialized pool in child 3222 for
(ejld1132.ejgvdns) min=0 max=25 smax=25
[Wed Jan 18 11:02:12.800571 2017] [proxy_hcheck:debug] [pid 3222:tid
140560165148416] mod_proxy_hcheck.c(457): AH03248: Creating hc worker 20a0c70
for http://ejld1131.ejgvdns:8170
[Wed Jan 18 11:02:12.800595 2017] [proxy:debug] [pid 3222:tid 140560165148416]
proxy_util.c(1779): AH00925: initializing worker 20a0c70 shared
[Wed Jan 18 11:02:12.800603 2017] [proxy:debug] [pid 3222:tid 140560165148416]
proxy_util.c(1821): AH00927: initializing worker 20a0c70 local
[Wed Jan 18 11:02:12.800619 2017] [proxy:debug] [pid 3222:tid 140560165148416]
proxy_util.c(1856): AH00930: initialized pool in child 3222 for
(ejld1131.ejgvdns) min=0 max=25 smax=25
[Wed Jan 18 11:02:12.800623 2017] [proxy_hcheck:debug] [pid 3222:tid
140559523972864] mod_proxy_hcheck.c(819): AH03256: Threaded Health checking
http://ejld1132.ejgvdns:8170
[Wed Jan 18 11:02:12.800710 2017] [proxy_hcheck:debug] [pid 3222:tid
140559641470720] mod_proxy_hcheck.c(819): AH03256: Threaded Health checking
http://ejld1131.ejgvdns:8170
[Wed Jan 18 11:02:12.800874 2017] [proxy:debug] [pid 3222:tid 140559523972864]
proxy_util.c(2156): AH00942: HCOH: has acquired connection for
(ejld1132.ejgvdns)
[Wed Jan 18 11:02:12.800900 2017] [proxy:debug] [pid 3222:tid 140559641470720]
proxy_util.c(2156): AH00942: HCOH: has acquired connection for
(ejld1131.ejgvdns)
[Wed Jan 18 11:02:12.801249 2017] [proxy:debug] [pid 3222:tid 140559523972864]
proxy_util.c(2884): AH02824: HCOH: connection established with
10.190.40.107:8170 (ejld1132.ejgvdns)
[Wed Jan 18 11:02:12.801293 2017] [proxy:debug] [pid 3222:tid 140559523972864]
proxy_util.c(3051): AH00962: HCOH: connection complete to 10.190.40.107:8170
(ejld1132.ejgvdns)
[Wed Jan 18 11:02:12.801412 2017] [proxy:debug] [pid 3222:tid 140559641470720]
proxy_util.c(2884): AH02824: HCOH: connection established with
10.190.40.106:8170 (ejld1131.ejgvdns)
[Wed Jan 18 11:02:12.801445 2017] [proxy:debug] [pid 3222:tid 140559641470720]
proxy_util.c(3051): AH00962: HCOH: connection complete to 10.190.40.106:8170
(ejld1131.ejgvdns)
[Wed Jan 18 

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #25 from Jim Jagielski  ---
Can you try w/ the latest patch and let us know?

Thanks!!

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

Yann Ylavic  changed:

   What|Removed |Added

  Attachment #34619|0   |1
is obsolete||

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #24 from Yann Ylavic  ---
Created attachment 34620
  --> https://bz.apache.org/bugzilla/attachment.cgi?id=34620=edit
Updated patch w/ method + single pool and bucket_alloc

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

Jim Jagielski  changed:

   What|Removed |Added

  Attachment #34618|0   |1
is obsolete||

--- Comment #23 from Jim Jagielski  ---
Created attachment 34619
  --> https://bz.apache.org/bugzilla/attachment.cgi?id=34619=edit
Updated patch w/ method

Updated thread-safety patch w/ method and req pre-prepared

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #22 from Yann Ylavic  ---
(In reply to Jim Jagielski from comment #20)
> Great idea!

Also it seems that we should get rid of ctx->ba, and probably use
r->connection->bucket_alloc instead.

We could also probably avoid a subpool for r, ptemp is fine.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #21 from Endika  ---
okso, i'll wait to the last change

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #20 from Jim Jagielski  ---
Great idea!

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #19 from Yann Ylavic  ---
(In reply to Jim Jagielski from comment #17)
> Created attachment 34618 [details]
> Updated thread patch

Nice, maybe put the method string in wctx too (at init time)?

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

Jim Jagielski  changed:

   What|Removed |Added

  Attachment #34601|0   |1
is obsolete||
 CC||j...@apache.org

--- Comment #17 from Jim Jagielski  ---
Created attachment 34618
  --> https://bz.apache.org/bugzilla/attachment.cgi?id=34618=edit
Updated thread patch

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

Jim Jagielski  changed:

   What|Removed |Added

 Status|NEW |NEEDINFO

--- Comment #18 from Jim Jagielski  ---
Can you check out the updated patch?

Thx again!

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #16 from Endika  ---
Hello,

We have applied the patch and apparently it works fine!!!

After making jmeter tests during 27 hours and 485500 request the Segmentation
Fault error hasn`t appeared and all request have gone ok.

Thanks and Good work!!!

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #15 from Jim Jagielski  ---
As a quick check, see if setting

   ProxyHCTPsize 0

resolves the issue. This avoids thread pools for the hcheck workers

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #14 from Jim Jagielski  ---
What we need is a pool that each HC worker can use.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #13 from Yann Ylavic  ---
(In reply to Jim Jagielski from comment #10)
> We are generating the request for each and every health check, so yeah, it's
> a fast path.

(In reply to Jim Jagielski from comment #11)
> Plus, it should never change for each HC worker.

OK, so let's allocate it once at init time, my first goal here is to verify if
it solves the issue(s)...

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #12 from Yann Ylavic  ---
(In reply to Jim Jagielski from comment #8)
> Also, from what I can see, the common logic path and usage for
> hc_determine_connection() simply checks that worker->cp->addr isn't null and
> then assigns worker->cp->addr to worker->cp->addr... This seems wasteful and
> inefficient as well.

This is the same here AFAICT, hc_init_worker() calls hc_determine_connection()
exactly once, otherwise it's called from hc_get_backend() which should either
find an existing/valid worker->cp->addr, or NULL (in which case it must create
a new (backend->)addr for each connection and be thread-safe too.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #11 from Jim Jagielski  ---
Plus, it should never change for each HC worker.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #10 from Jim Jagielski  ---
We are generating the request for each and every health check, so yeah, it's a
fast path.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #9 from Yann Ylavic  ---
(In reply to Jim Jagielski from comment #7)
> Seems crazy and inefficient that we are doing the below every single time:
> 
>req = apr_psprintf(ptemp,...
> 
> Why??

Because ctx->p is not thread-safe here, and req is the request-line + Host
header, not that huge...

Is that a fast path, really?

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #8 from Jim Jagielski  ---
Also, from what I can see, the common logic path and usage for
hc_determine_connection() simply checks that worker->cp->addr isn't null and
then assigns worker->cp->addr to worker->cp->addr... This seems wasteful and
inefficient as well.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #7 from Jim Jagielski  ---
Seems crazy and inefficient that we are doing the below every single time:

   req = apr_psprintf(ptemp,...

Why??

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

Yann Ylavic  changed:

   What|Removed |Added

 CC||egandar...@gfi.es

--- Comment #6 from Yann Ylavic  ---
*** Bug 60553 has been marked as a duplicate of this bug. ***

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #5 from Yann Ylavic  ---
Created attachment 34601
  --> https://bz.apache.org/bugzilla/attachment.cgi?id=34601=edit
Fix thread-safety in mod_proxy_hcheck

Could you try this patch to address possible threads concurrency issues?

-- 
You are receiving this mail because:
You are the assignee for the bug.
-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

6731...@163.com changed:

   What|Removed |Added

 Status|NEEDINFO|NEW

--- Comment #4 from 6731...@163.com ---
apache debug log:
[Fri Oct 21 15:34:18.030981 2016] [proxy_balancer:debug] [pid 8773:tid
140072177952512] mod_proxy_balancer.c(132): AH01158: Looking at
balancer://xxx.xxx.com -> http://x.x.x.x:8080 initialized?
[Fri Oct 21 15:34:18.031014 2016] [proxy_balancer:debug] [pid 8774:tid
140072177952512] mod_proxy_balancer.c(132): AH01158: Looking at
balancer://xxx.xxx.com -> http://x.x.x.x:8080 initialized?
[Fri Oct 21 15:34:18.031056 2016] [proxy_balancer:debug] [pid 8775:tid
140072177952512] mod_proxy_balancer.c(132): AH01158: Looking at
balancer://xxx.xxx.com -> http://x.x.x.x:8080 initialized?
[Fri Oct 21 15:34:18.031212 2016] [proxy_balancer:debug] [pid 8773:tid
140072177952512] mod_proxy_balancer.c(132): AH01158: Looking at
balancer://xxx.xxx.com -> http://x.x.x.x:8080 initialized?
[Fri Oct 21 15:34:18.031236 2016] [proxy_balancer:debug] [pid 8774:tid
140072177952512] mod_proxy_balancer.c(132): AH01158: Looking at
balancer://xxx.xxx.com -> http://x.x.x.x:8080 initialized?
[Fri Oct 21 15:34:18.031262 2016] [proxy_balancer:debug] [pid 8775:tid
140072177952512] mod_proxy_balancer.c(132): AH01158: Looking at
balancer://xxx.xxx.com -> http://x.x.x.x:8080 initialized?
[Fri Oct 21 15:34:49.075240 2016] [core:notice] [pid 8439:tid 140072177952512]
AH00051: child pid 8775 exit signal Segmentation fault (11), possible coredump
in /tmp/test

strace:
8439  <... clone resumed> child_stack=0,
flags=CLONE_CHILD_CLEARTID|CLONE_CHILD_SETTID|SIGCHLD,
child_tidptr=0x7f651868e9d0) = 8775
8775  set_robust_list(0x7f651868e9e0, 24 
8775  <... set_robust_list resumed> )   = 0
8775  rt_sigaction(SIGTERM, {0x46fb10, [], SA_RESTORER|SA_INTERRUPT,
0x7f65175707e0}, {0x46b620, [], SA_RESTORER, 0x7f65175707e0}, 8) = 0
8775  geteuid( 
8775  <... geteuid resumed> )   = 0
8775  setgid(2 
8775  <... setgid resumed> )= 0
8775  open("/proc/sys/kernel/ngroups_max", O_RDONLY 
8775  <... open resumed> )  = 11
8775  read(11,  
8775  <... read resumed> "65536\n", 31) = 6
8775  close(11 
8775  <... close resumed> ) = 0
8775  open("/etc/group", O_RDONLY|O_CLOEXEC 
8775  <... open resumed> )  = 11
8775  fstat(11,  
8775  <... fstat resumed> {st_mode=S_IFREG|0644, st_size=529, ...}) = 0
8775  mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0

8775  <... mmap resumed> )  = 0x7f6518697000
8775  lseek(11, 0, SEEK_CUR 
8775  <... lseek resumed> ) = 0
8775  read(11,  
8775  <... read resumed> "root:x:0:\nbin:x:1:bin,daemon\ndae"..., 4096) = 529
8775  read(11,  
8775  <... read resumed> "", 4096)  = 0
8775  close(11 
8775  <... close resumed> ) = 0
8775  munmap(0x7f6518697000, 4096 
8775  <... munmap resumed> )= 0
8775  setgroups(4, [2, 1, 4, 7] 
8775  <... setgroups resumed> ) = 0
8775  geteuid( 
8775  <... geteuid resumed> )   = 0
8775  setuid(2 
8775  <... setuid resumed> )= 0
8775  prctl(PR_SET_DUMPABLE, 1 
8775  <... prctl resumed> ) = 0
8775  futex(0x7f6516dc5a80, FUTEX_WAKE_PRIVATE, 2147483647 
8775  <... futex resumed> ) = 0
8775  semop(3702787, {{0, -1, SEM_UNDO}}, 1 
8775  <... semop resumed> ) = 0
8775  semop(3702787, {{0, 1, SEM_UNDO}}, 1 
8775  <... semop resumed> ) = 0
8775  semop(3702787, {{0, -1, SEM_UNDO}}, 1 
8775  <... semop resumed> ) = 0
8775  semop(3702787, {{0, 1, SEM_UNDO}}, 1) = 0
8775  write(8, "[Fri Oct 21 15:34:18.031056 2016"..., 210 
8775  <... write resumed> ) = 210
8775  write(8, "[Fri Oct 21 15:34:18.031262 2016"..., 210 
8775  <... write resumed> ) = 210
8775  semop(3702787, {{0, -1, SEM_UNDO}}, 1 
8775  <... semop resumed> ) = 0
8775  semop(3702787, {{0, 1, SEM_UNDO}}, 1) = 0
8775  mmap(NULL, 10489856, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_ANONYMOUS|MAP_STACK, -1, 0) = 0x7f6512935000
8775  mprotect(0x7f6512935000, 4096, PROT_NONE 
8775  <... mprotect resumed> )  = 0
8775  clone( 
8775  <... clone resumed> child_stack=0x7f6513334ff0,
flags=CLONE_VM|CLONE_FS|CLONE_FILES|CLONE_SIGHAND|CLONE_THREAD|CLONE_SYSVSEM|CLONE_SETTLS|CLONE_PARENT_SETTID|CLONE_CHILD_CLEARTID,
parent_tidptr=0x7f65133359d0, tls=0x7f6513335700, child_tidptr=0x7f65133359d0)
= 8776
8775  futex(0x23d46d8, FUTEX_WAIT_PRIVATE, 2, NULL 
8775  <... futex resumed> ) = -1 EAGAIN (Resource temporarily
unavailable)
8775  futex(0x23d46d8, FUTEX_WAKE_PRIVATE, 1 
8775  <... futex resumed> ) = 0
8775  rt_sigprocmask(SIG_SETMASK, ~[ILL TRAP ABRT BUS FPE SEGV USR2 PIPE SYS
RTMIN RT_1],  
8775  <... rt_sigprocmask resumed> NULL, 8) = 0
8775  mmap(NULL, 10489856, PROT_READ|PROT_WRITE,

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

Ruediger Pluem  changed:

   What|Removed |Added

 Status|NEW |NEEDINFO

--- Comment #3 from Ruediger Pluem  ---
It would be helpful to have the following additional information:

1. The full stack trace of the crashing thread not just the last frame.
2. Having debug logging enabled. You should see an error message with the
number 03248 in the log just before the crash.
3. In gdb please execute the following commands for the crashed thread:

print wptr
print *wptr
print hc
print hc->s
print hc->s->name
print *worker->s->name

-- 
You are receiving this mail because:
You are the assignee for the bug.

-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #2 from 6731...@163.com ---
encountered same bug here

apache info:
Server version: Apache/2.4.23 (Unix)
Server built:   Oct 17 2016 17:26:32
Server's Module Magic Number: 20120211:61
Server loaded:  APR 1.5.2, APR-UTIL 1.5.4
Compiled using: APR 1.5.2, APR-UTIL 1.5.4
Architecture:   64-bit
Server MPM: event
  threaded: yes (fixed thread count)
forked: yes (variable process count)
Server compiled with
 -D APR_HAS_SENDFILE
 -D APR_HAS_MMAP
 -D APR_HAVE_IPV6 (IPv4-mapped addresses enabled)
 -D APR_USE_SYSVSEM_SERIALIZE
 -D APR_USE_PTHREAD_SERIALIZE
 -D SINGLE_LISTEN_UNSERIALIZED_ACCEPT
 -D APR_HAS_OTHER_CHILD
 -D AP_HAVE_RELIABLE_PIPED_LOGS
 -D DYNAMIC_MODULE_LIMIT=256
 -D HTTPD_ROOT="/data/apache"
 -D SUEXEC_BIN="/data/apache/bin/suexec"
 -D DEFAULT_PIDLOG="logs/httpd.pid"
 -D DEFAULT_SCOREBOARD="logs/apache_runtime_status"
 -D DEFAULT_ERRORLOG="logs/error_log"
 -D AP_TYPES_CONFIG_FILE="conf/mime.types"
 -D SERVER_CONFIG_FILE="conf/httpd.conf"

apache config:
#MODULES
LoadFile /usr/lib64/libxml2.so
LoadFile /usr/lib64/liblua-5.1.so
LoadModule security2_module modsecurity/lib/mod_security2.so
LoadModule authz_core_module modules/mod_authz_core.so
LoadModule mime_module modules/mod_mime.so
LoadModule log_config_module modules/mod_log_config.so
LoadModule headers_module modules/mod_headers.so
LoadModule unique_id_module modules/mod_unique_id.so
LoadModule socache_shmcb_module modules/mod_socache_shmcb.so
#REVERSE_PROXY MODULES
LoadModule proxy_module modules/mod_proxy.so
LoadModule proxy_http_module modules/mod_proxy_http.so
LoadModule ssl_module modules/mod_ssl.so
LoadModule unixd_module modules/mod_unixd.so
LoadModule proxy_balancer_module modules/mod_proxy_balancer.so
LoadModule slotmem_shm_module modules/mod_slotmem_shm.so
LoadModule lbmethod_byrequests_module modules/mod_lbmethod_byrequests.so
LoadModule proxy_hcheck_module modules/mod_proxy_hcheck.so
LoadModule watchdog_module modules/mod_watchdog.so

hc config:

BalancerMember "http://xxx:8080; #hcmethod=TCP hcinterval=2 hcpasses=3
hcfails=3
BalancerMember "http://xxx:8080; #hcmethod=TCP hcinterval=2 hcpasses=3
hcfails=3


coredump info:
Core was generated by `/data/apache/bin/httpd -k start'.
Program terminated with signal 11, Segmentation fault.
#0  0x7fd8c4384169 in hc_get_hcworker (ctx=0x117af70, worker=0x1238f28,
p=0x7fd8bc030e58) at mod_proxy_hcheck.c:426
426PROXY_STRNCPY(hc->s->name, wptr);

-- 
You are receiving this mail because:
You are the assignee for the bug.

-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

--- Comment #1 from 6731...@163.com ---
encountered same bug here

coredump info below:

Core was generated by `/data/apache/bin/httpd -k start'.
Program terminated with signal 11, Segmentation fault.
#0  0x7fd8c4384169 in hc_get_hcworker (ctx=0x117af70, worker=0x1238f28,
p=0x7fd8bc030e58) at mod_proxy_hcheck.c:426
426PROXY_STRNCPY(hc->s->name, wptr);

-- 
You are receiving this mail because:
You are the assignee for the bug.

-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org

[Bug 60071] Child httpd processes crash with Segmentation fault when enabling more than 1 healthcheck

[bugzilla]

https://bz.apache.org/bugzilla/show_bug.cgi?id=60071

apa...@fiddaman.net changed:

   What|Removed |Added

 CC||apa...@fiddaman.net

-- 
You are receiving this mail because:
You are the assignee for the bug.

-
To unsubscribe, e-mail: bugs-unsubscr...@httpd.apache.org
For additional commands, e-mail: bugs-h...@httpd.apache.org