Re: iscsi on Ubuntu - write of large files failed

2009-04-22 Thread Ulrich Windl 

On 21 Apr 2009 at 23:22, tomorrowflow wrote:

 
 Hi,
 
 i'm running open-iscis and ocfs2 on ubuntu 8.10 (open-iscsi 0.865) and
 having much trouble...
 when i'm trying to write large files (5GB) to the openfiler target, my
 system reboots after 800MB - there are thre boxes all configured the
 same. i also tryed open-iscsi 0.870 and kernels between 2.6.24 and
 2.6.27.

Is there any kernel message before reboot (/var/log/*), or is there some XEN 
message (xm dmesg (as long as it's still running))?

 there are VMs on xen running on ocfs2, so they are destroyed after
 some time - also the concrete failure is not reproducable because i
 can't see when it will happen or has happend.
 the only way to reproduce a failure is the above.

Maybe changing the syslog to write messages synchonously and putting messages 
on a 
synchronously writing filesystem would be helpful to capture the last messages. 
Maybe something like
while xm dmesgxen.log
do
  sleep 1 # or longer
done

will be helpful.

I'm not using file backends for XEN (yet); how did you configure those?

Regards,
Ulrich


--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
open-iscsi group.
To post to this group, send email to open-iscsi@googlegroups.com
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/open-iscsi
-~--~~~~--~~--~--~---

Re: iscsi on Ubuntu - write of large files failed

2009-04-22 Thread tomorrowflow 

On 22 Apr., 08:43, Ulrich Windl ulrich.wi...@rz.uni-regensburg.de
wrote:
 On 21 Apr 2009 at 23:22, tomorrowflow wrote:



  Hi,

  i'm running open-iscis and ocfs2 on ubuntu 8.10 (open-iscsi 0.865) and
  having much trouble...
  when i'm trying to write large files (5GB) to the openfiler target, my
  system reboots after 800MB - there are thre boxes all configured the
  same. i also tryed open-iscsi 0.870 and kernels between 2.6.24 and
  2.6.27.

 Is there any kernel message before reboot (/var/log/*), or is there some XEN
 message (xm dmesg (as long as it's still running))?
The syslog is going to a syslog-server, but there is nothing to see.
Where can i set the debug level for the scsid? I started the iscsid
from the command-line and the system rebooted urgently.

I do not use the iscsi-backend of xen/libvirt - the VMs are still in
img-files. The option is still there through libvirt, but i did't try
it out yet.

Thank you,

Florian
--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
open-iscsi group.
To post to this group, send email to open-iscsi@googlegroups.com
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/open-iscsi
-~--~~~~--~~--~--~---

Re: Persistent connections across initiator reboot

2009-04-22 Thread Mike Christie 

HIMANSHU wrote:
 Thank you really.But I have not got the expected results.
 
 Steps
 
 1. iscsiadm -m discovery -t sendtargets -p 192.168.30.12 -o new -o
 delete
 
 2. iscsiadm -m node -T iqn.2009-04.com.systems:Target1 -p
 192.168.30.12:3260 -o update -n node.startup -v automatic
 
so '/var/lib/iscsi/nodes/iqn.2009-04.com.systems
 \:Target1/192.168.30.12\,3260\,1/default' entry changed from
 node.startup=manual(as specified in iscsid.conf) to
 node.startup=automatic
 
 3. iscsiadm -m node -T iqn.2009-04.com.systems:Target1 -p
 192.168.30.12:3260 --login
 
 4. Then if I again perform step 1(i.e. discovery),node.startup gets
 overwritten as 'manual'.
 
 I have specified 'node.startup' entry to be 'manual' and will update
 it to 'automatic' only of those targets which are logged in.
 
 What is going wrong.Am I performing the same operations as you told?
 I am using iscsi-initiator-utils-6.2.0.868-0.7.el5 and it seems to
 have discovery switches.
 Do I need to upgrade it?


Yeah, just upgrade to:
iscsi-initiator-utils-6.2.0.868-0.18.el5

 
 
 On Apr 21, 7:24 pm, Mike Christie micha...@cs.wisc.edu wrote:
 HIMANSHU wrote:
 Yeah.This worked for me.I specified 'node.startup=manual' in
 'iscsid.conf'.
 so running update command to change it to 'automatic' only on login(or
 I will give option for persistent connection across reboot).
 But my problem these changes vanish after another discovery command.
 I login to one target say TAR1, from 30.51 and fired update command to
 change 'node.startup' to 'automatic'.
 and then after some time,If I want to login to another target from
 30.51,I need to fire discovery command,which will make 'node.startup'
 of TAR1 changed to 'manual' again due to 'iscsid.conf'.
 You can pass the discovery commands some flags to indicated how you want
 it to manage the db.

 iscsiadm -m discovery -t st -p ip -o new

 will just add new target portals that are not in the db.

 iscsiadm -m discovery -t st -p ip -o delete
 will delete target portals that are in the db but no longer returned by
 the target in sendtargets discvoery.

 iscsiadm -m discovery -t st -p ip -o update
   was supposed to update target portal records that are in the db and
 returned by sendtargets using the iscsid.conf info. I checked the code
 today and it actually updates existing records and will add new ones
 using the iscsid.conf info.

 Then you can pass in combos. I think you want
 iscsiadm -m discovery -t st -p ip -o new -o delete
 this will add new records for new target portals, and delete records for
 targets no longer returned by the target, and for existing portals it
 will not update the records so your existing record settings will not be
 overwritten.
  


--~--~-~--~~~---~--~~
You received this message because you are subscribed to the Google Groups 
open-iscsi group.
To post to this group, send email to open-iscsi@googlegroups.com
To unsubscribe from this group, send email to 
open-iscsi+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/open-iscsi
-~--~~~~--~~--~--~---