Repeated Omniweb crashes
 

Since upgrading my machine to 19GB, I have experienced repeated crashes as shown below. This occurs for all sneaky peak and production builds.

Please let me know if you require any additional information.

- - - - - - - - -

[B]Machine[/B]:

Mac Pro: Xeon
Memory: 19GB
OS X: 10.5.7


[B]Console Error[/B]:

18/05/2009 12:55:48 [0x0-0x54d54d].com.omnigroup.OmniWeb5 OmniWeb(8015,0xb032f000) malloc: *** mmap(size=16777216) failed (error code=12)
18/05/2009 12:55:48 [0x0-0x54d54d].com.omnigroup.OmniWeb5 *** error: can't allocate region
18/05/2009 12:55:48 [0x0-0x54d54d].com.omnigroup.OmniWeb5 *** set a breakpoint in malloc_error_break to debug
18/05/2009 12:55:48 [0x0-0x54d54d].com.omnigroup.OmniWeb5 exception = 0x00000006, crashedThreadStackPointer = 0xb032e470, code = 0x00000002;0x00000000;
18/05/2009 12:55:48 [0x0-0x54d54d].com.omnigroup.OmniWeb5 Starting '/Applications/My Utilities/Internet/OmniWeb.app/Contents/Resources/OmniGroupCrashCatcher.app/Contents/MacOS/OmniGroupCrashCatcher' for pid 8015.
18/05/2009 12:55:48 [0x0-0x54d54d].com.omnigroup.OmniWeb5 OmniWeb(8015,0xb02ad000) malloc: *** mmap(size=16777216) failed (error code=12)
18/05/2009 12:55:48 [0x0-0x54d54d].com.omnigroup.OmniWeb5 *** error: can't allocate region
18/05/2009 12:55:48 [0x0-0x54d54d].com.omnigroup.OmniWeb5 *** set a breakpoint in malloc_error_break to debug
18/05/2009 12:55:48 [0x0-0x54d54d].com.omnigroup.OmniWeb5 OmniWeb(8015,0xa0142720) malloc: *** mmap(size=16777216) failed (error code=12)
18/05/2009 12:55:48 [0x0-0x54d54d].com.omnigroup.OmniWeb5 *** error: can't allocate region
18/05/2009 12:55:48 [0x0-0x54d54d].com.omnigroup.OmniWeb5 *** set a breakpoint in malloc_error_break to debug
18/05/2009 12:55:51 com.apple.launchd ([0x0-0x54d54d].com.omnigroup.OmniWeb5[8015]) Exited with exit code: 1

How much RAM did you have before the upgrade, when everything was working?

[QUOTE=svenners;59993]18/05/2009 12:55:48 [0x0-0x54d54d].com.omnigroup.OmniWeb5 OmniWeb(8015,0xb032f000) malloc: *** mmap(size=16777216) failed (error code=12)[/QUOTE]

Hmm. Error code 12 is a "Cannot allocate memory" error from BSD APIs:

[CODE] 12 ENOMEM Cannot allocate memory. The new process image required more
memory than was allowed by the hardware or by system-imposed mem-
ory management constraints. A lack of swap space is normally
temporary; however, a lack of core is not. Soft limits may be
increased to their corresponding hard limits.
[/CODE]

That seems somewhat strange to me, because the memory OmniWeb is trying to allocate is mapping to a file (that's what "mmap" does) and it's asking for just 16MB (relatively small compared to the 4GB address space which is available*to it in its 32-bit mode—and even the full 4GB address space is small compared to amount of the RAM you now have).

If this was triggered by installing new RAM, I guess I see two possibilities:

One possibility is that some part of OmniWeb is allocating memory based on a percentage of your total RAM, and exhausted your address space (since it's running in 32-bit mode). If that were the case, I'd expect we would have seen other reports from other users who have lots of RAM, which we haven't—but just in case, you might watch Activity Monitor to see what OmniWeb's VSIZE is before it crashes.

The other possibility I can think of is that some of your new RAM might be bad, so when OmniWeb tries to use some of it the allocation fails and reports an error. I haven't done much memory testing myself so I can't recommend a good testing application, but a quick good search for "test Mac memory" seems to turn up some hits. (Be sure to look for a 64-bit tester, since you can't address 19GB of memory with a 32-bit app.)

One last question: Did OmniCrashCatcher launch and offer to report a crash? If so, did you send it in? (I'm not finding a ticket under your email address.) A crash report might help us see exactly where OmniWeb's code is failing to allocate memory.

Ken,

I have just submitted a Crash Report ( OG #438637 )

Thanks for your prompt response.