Hi Manu

Yes the Assert is in the codepath where the registration was to be  
preserved. So everything except the registration is preserved.

Sumit

On Feb 12, 2008, at 10:11 PM, manu sawkar wrote:

> Hey Sumit,
>
> Thanks for the quick reply. Still havent been able to actually use  
> a client
> and connect to verify but i will be doing disk tests with my raid  
> and a client
> tomorrow so hopefully i'll have some updates, however i can provide  
> you
> with some more info related to this bug... (that you may or may not  
> already
> know about)
>
> When i reboot my system, the LU's that point to files are  
> preserved... but
> the ones i've made that point to raw discs are not (perhaps this is  
> due
> to the failed assert?)
>
> the views appear to be preserved though (makes sense i think)
> this maybe of interest:
> (the unregistered correspond to a rawdisk i added after it disappeared
> upon rebooting)
>
> stmfadm list-lu -v
> LU Name: 6000AE4066000000000047B3342D0001
>     Operational Status: unregistered
>     Provider Name     : unregistered
>     Alias             : -
>     View Entry Count  : 1
> LU Name: 6000AE4066000000000047B334A50002
>     Operational Status: Online
>     Provider Name     : sbd
>     Alias             : /lun0
>     View Entry Count  : 1
> LU Name: 6000AE4066000000000047B33B9B0001
>     Operational Status: unregistered
>     Provider Name     : unregistered
>     Alias             : -
>     View Entry Count  : 1
>
>
> On Feb 13, 2008 12:46 AM, Sumit Gupta <[EMAIL PROTECTED]> wrote:
>> Hi Manu
>>
>> Thanks for your feedback. Despite the assert It did work in this case
>> as you suspected. Regarding the CLI assert, it has been fixed and the
>> fix will be included in the next opensolaris update along with rest
>> of the changes and fixes. We are looking into some other issues at
>> the moment and will update the package and sources soon.
>>
>> Thanks
>> Sumit
>>
>> On Feb 12, 2008, at 6:32 PM, Manu Sawkar wrote:
>>
>>> First off, thanks to Sumit et al. for the great work on this
>>> amazing project. I'm a total solaris newbie and have enjoyed
>>> tinkering with the OS the past couple days. The purpose of this
>>> whole endeavor is to try to create my own (cheap) 4GBs fc target
>>> and before comstar was available it seemed my options were non-
>>> existent. Anyways, on to my bug.
>>>
>>> I have been able to create and register 2 luns from a 1g mkfile and
>>> a dynamic growing file  successfully and exported them with no
>>> problem (i dont have access to an initiator client till tomorrow,
>>> so i'm assuming that works).
>>>
>>> however, i then added an unformated sata disk to the equation
>>> because my ultimate goal is to export raw disks from this machine
>>> with the hopes that those disks could be removed from the solaris
>>> box and mounted elsewhere if need be (this is possible right? if
>>> they are exported as raw devices to hosts... i should be able to
>>> put that disk in the host machine without it knowing that anything
>>> happened)
>>>
>>> anyways, here is what happened:
>>>
>>> a format shows the new disk is up. a prtpart shows its 320GB
>>> i already have 2 LU's registered and i can create the raw disk as a
>>> 3rd LU. However, when i register the 3rd LU i get that assertion
>>> failure and core dump.  It does appears to have registered
>>> correctly, i can export the view... and subsequently prtpart says
>>> the device is busy... so it looks like it worked. hopefully it did!
>>>
>>> Thanks,
>>> -Manu
>>>
>>> below is copied from the terminal.
>>>
>>> # format
>>> Searching for disks...done
>>>
>>>
>>> AVAILABLE DISK SELECTIONS:
>>>        0. c1d0 <DEFAULT cyl 30385 alt 2 hd 255 sec 63>
>>>           /[EMAIL PROTECTED],0/[EMAIL PROTECTED],2/[EMAIL PROTECTED]/[EMAIL 
>>> PROTECTED],0
>>>        1. c4d0 <DEFAULT cyl 38911 alt 2 hd 255 sec 63>
>>>           /[EMAIL PROTECTED],0/[EMAIL PROTECTED],2/[EMAIL PROTECTED]/[EMAIL 
>>> PROTECTED],0
>>> Specify disk (enter its number): ^C
>>>
>>>
>>> # prtpart
>>>
>>> Fdisk information for device /dev/rdsk/c4d0p0
>>>
>>> Block Size : 512 bytes
>>> Controller : ide
>>> Disk       : cmdk
>>> Capacity   : 298 GB
>>> FDISK table NOT VALID
>>> Reminder: dont forget to use either 'p0' or 'D0' device
>>>
>>>
>>> # sbdadm list-lu
>>>
>>> Found 2 LU(s)
>>>
>>>             GUID                    DATA SIZE           SOURCE
>>> --------------------------------  -------------------
>>> ----------------
>>> 6000ae4066000000000047b2f7080002      53687025664      /lun1
>>> 6000ae4066000000000047b2f6160001      1073676288       /lun0
>>>
>>> # sbdadm create-lu /dev/rdsk/c4d0p0
>>> LU created succesfully.
>>> Assigned GUID is 6000ae4066000000000047b303e00001.
>>> # sbdadm register-lu /dev/rdsk/c4d0p0
>>>
>>> Success! Registered the following LU:
>>>
>>>             GUID                    DATA SIZE           SOURCE
>>> --------------------------------  -------------------
>>> ----------------
>>> 6000ae4066000000000047b303e00001      320070287360     /dev/rdsk/
>>> c4d0p0
>>> Assertion failed: h->rh_holder == pthread_self(), file ../common/
>>> lowlevel.c, line 126
>>> Abort - core dumped
>>>
>>> # sbdadm list-lu
>>>
>>> Found 3 LU(s)
>>>
>>>             GUID                    DATA SIZE           SOURCE
>>> --------------------------------  -------------------
>>> ----------------
>>> 6000ae4066000000000047b303e00001      320070287360     /dev/rdsk/
>>> c4d0p0
>>> 6000ae4066000000000047b2f7080002      53687025664      /lun1
>>> 6000ae4066000000000047b2f6160001      1073676288       /lun0
>>>
>>> # prtpart /dev/rdsk/c4d0p0
>>> cannot open file: Device busy
>>>
>>>
>>> This message posted from opensolaris.org
>>> _______________________________________________
>>> storage-discuss mailing list
>>> [email protected]
>>> http://mail.opensolaris.org/mailman/listinfo/storage-discuss
>>
>>

_______________________________________________
storage-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/storage-discuss

Reply via email to