Haik Aftandilian wrote:
> Alexandre Chartre wrote:
>> What is taking 2 hours? Dumping the crash during a panic or running
>> savecore to generate the unix/vmcore files after the panic?
> 
> Dumping the crash. It isn't this slow all the time. The times I've 
> noticed it have been I was running a memory intensive test and I've 
> stopped the system and issued a "sync".
> 
>> I never had this problem although I regularly generates crash dumps
>> which are more 500Mb large.
> 
> Hmm. Let me try this without my changes. Perhaps my LDC changes are 
> affecting the speed.

I can't reproduce this on a domain without my changes, so disregard my 
question for now.

Thanks,
Haik

> 
> P.S., please let me know if this is not the type of conversation this 
> alias is intended to be used for.
> 
> Thanks,
> Haik
> 
>>
>> alex.
>>
>> Haik Aftandilian wrote:
>>> At present I've configured my LDoms such that each domain has a swap 
>>> disk image and a root disk image. The problem I run into is that for 
>>> large memory images (my domain has 4GB of memory), it takes ~2 hours 
>>> to save the core dump.
>>>
>>> Here's the zfs listing for a single domain.
>>>
>>> # zfs list
>>> NAME                                 USED  AVAIL  REFER  MOUNTPOINT
>>> ldomszpool                          16.2G  31.7G    29K  /ldomszpool
>>> ldomszpool/ldm1-disk                11.2G  31.7G  10.0G  
>>> /ldomszpool/ldm1-disk
>>> ldomszpool/ldm1-swap                1.00G  31.7G  1.00G  
>>> /ldomszpool/ldm1-swap
>>>
>>> Is there a better way to do this? Or is using a real disk instead of 
>>> a file-based disk image for swap going to be the only way to get 
>>> better core dump performance?
>>>
>>> Thanks,
>>> Haik
>>> -- 
>>> This message was posted from opensolaris.org
>>> _______________________________________________
>>> ldoms-discuss mailing list
>>> ldoms-discuss at opensolaris.org
>>> http://mail.opensolaris.org/mailman/listinfo/ldoms-discuss
> 

Reply via email to