06-24-2011 12:29 AM
Hi All,
Recently we are facing tape drive fluctuation issue at our site.
A total of 20 tape drives were zoned and was visible on master & the LAN media server.
We are receiving plug and play error in event viewer of the LAN Media server(which is a windows server) and the number of tape drives in device management continuosly change.
But we are unable to convince the same to SUN Support Team as we are unable to find these tape fluctuation events on SUN OS.
When asked the SUN support team as weel as SAN Team; they provide us with the output of mt -f <drivepath> status; and stating that all drives are visible on the OS side and there are no error logs at their end .
sgscan tape and tpautoconf -t provides varied output each time.
Is there any Sun OS command where tape drive visibility can be checked.
Due to this fluctuation issue; tapes get stuck in the drive resulting in failure of backups...
Setup: Master Server (Solaris 10)
LAN Media Server(Windows Server 2003)
500+ clients & approx 160 Media servers.
Library: SL500 and L180 each with 10 tape drives(Robot control host Master server)
Solved! Go to Solution.
06-24-2011 08:32 AM
I have issues everytime the switch gets "work" and I lose a few drives to a media server, I usually have to delete the drive from NB then re-add it.
06-24-2011 02:42 AM
06-24-2011 04:02 AM
Is it fluctuating just from the OS point of view?
Have seen some situation in Windows device configuration when involving RSM service, check this out:
http://www.symantec.com/docs/TECH37205
06-24-2011 06:14 AM
Windows RSM is in disabled state..
06-24-2011 06:15 AM
But why does tpautoconf -t show varied output each time on Sun Solaris as well.
I'am not that good at solaris but like windows solaris wont scan for hardware changes where as in windows it does.
In windows if the drives disappear and reapper it will rescan the hardware.
I also observed that "mt -f <> status" provides a slower O/P for certain drives.
06-24-2011 06:57 AM
Any loss of fiber channel connectivity should be logged in /var/adm/messages.
List of devices : prtconf or cfgadm -al
How are the tape drives zones ?. I recommand a 1:1 zoning where a each tape drives is located in is own zone along with the HBA.
But the issue to me sound like SAN issues.
06-24-2011 08:32 AM
I have issues everytime the switch gets "work" and I lose a few drives to a media server, I usually have to delete the drive from NB then re-add it.
06-24-2011 11:41 AM
You should only ever run tpautoconf and/or sgscan when NO drives are in use. It is quite normal for these commands not being able to report on drives that are in use/reserved on this or another media server.
Drives normally get stuck in robots when there is a mismatch between OS device name and drive position in robot.
Do you have Persistent Binding in place on Windows and Solaris media servers?
mt -f <drivepath> status is a good indication that device is visible/accessible at OS level. If not, output would be 'no such device or address'.
Which tape drives are installed on the Windows media server?