cancel
Showing results for 
Search instead for 
Did you mean: 

Tape alert 16

Kindermann
Occasional Visitor

Tape alert 16

I have a Proliant ML350 with an internal HP LTO230i. Running on SLE8 with HP DataProtector A5.10 and it works actually very well. But after few days running, I can’t unload the Tape in a normal way. I have to push and hold the unload button for around 10 seconds, then the tape will be unloaded.
At this time I am getting an error when I try to read, write or scan the tape for the first time after changing.
==============================================
[Normal] From: MMA@webdmz1.friatec.de "LTO1-webdmz1" Time: 31.03.2005 13:36:54
STARTING Media Agent "LTO1-webdmz1"

[Critical] From: MMA@webdmz1.friatec.de "LTO1-webdmz1" Time: 31.03.2005 13:37:03
/dev/nst0
Tape Alert [16]: The operation has failed because the tape cartridge was manually ejected while the tape drive was actively writing or reading.

[Normal] From: MSM@hercule.friatec.de "LTO1-webdmz1" Time: 31.03.2005 13:37:03
Data Protector medium "webdmz1_März" found.

[Normal] From: MMA@webdmz1.friatec.de "LTO1-webdmz1" Time: 31.03.2005 13:37:05
COMPLETED Media Agent "LTO1-webdmz1

The second scan I can make without alerts.
=============================================
[Normal] From: MMA@webdmz1.friatec.de "LTO1-webdmz1" Time: 31.03.2005 13:40:40
STARTING Media Agent "LTO1-webdmz1"

[Normal] From: MSM@hercule.friatec.de "LTO1-webdmz1" Time: 31.03.2005 13:40:54
Data Protector medium "webdmz1_März" found.

[Normal] From: MMA@webdmz1.friatec.de "LTO1-webdmz1" Time: 31.03.2005 13:41:04
COMPLETED Media Agent "LTO1-webdmz1"
==============================================

And after a reboot, the system works well for the next few days again. Also the tape can be unload with only a short push at the unload button.
Any ideas?

Thanks!
Herbert Kinderma
9 REPLIES
Curtis Ballard
Honored Contributor

Re: Tape alert 16

Pushing the manual eject button for 10 seconds causes a "Force Eject" which will generate a TapeAlert 16 message every time. The application reads that message the next time it goes to talk to the drive and reports it.

The message really means that a force eject was issued and the decoding or eject during read/write isn't quite correct.

You need to try a normal eject and wait. It is very unusual for these drives to fail to eject normally. The most common reason for an eject failure is that Windows Removable Storage Manager is running and has prevented eject. In rare cases the drive is in the process of an operation and will hold off before processing the eject.
Stuart Whitby
Trusted Contributor

Re: Tape alert 16

Are you sure this is done writing at the time that you try to eject it? I'd go check that the backup actually finished without errors. The other thing which raises warning flags for me here is that 3 letter word, DMZ, which shows up there...

Backups and firewalls don't really mix all that well. If this is a backup through a firewall and no data is sent for a long enough period (probably because the backup process is walking the filesystem and hasn't found anything that's changed within a set period) then the firewall may just note that there's been no activity and shut down that connection, leaving things hanging on each side. You say this happens every few days: does this coincide with any particular level of backup, by any chance?
A sysadmin should never cross his fingers in the hope commands will work. Makes for a lot of mistakes while typing.
Kindermann
Occasional Visitor

Re: Tape alert 16

When I try a normal eject, nothing will happen.
The Windows Removable Storage Manager is also running on at the SuSE Linux Enterprise Server 8 (SLES8)?
I also think that one process is using the drive and hold it, but which one is it and why it is only happen after few days?
Kindermann
Occasional Visitor

Re: Tape alert 16

When I will eject the tape, all backups are finished and the Session Report from DataProtector is without errors.
It is right, the system is placed in the DMZ. But the backup doesnâ t go through the firewall, because we are using an internal LTO Drive. Only the communication with the Cell Manager goes through.
The backup type is always the same. It is a complete backup every da
Curtis Ballard
Honored Contributor

Re: Tape alert 16

It sounds like a application on a host has issued a prevent media removal and you will have to track down which application and host. Since the drive is local it is prabably some application on your primary host has the drive locked.

Some backup applications will always prevent media removal when using a device and you must use an interface provided by the application to remove the media. Is your backup software running when you issue the unload?

You can use L&TT (HP StorageWorks Library and Tape Tools) to pull a support ticket right after you attempt a normal unload then a forced unload. If you zip and attach that ticket there should be additional information that will help explain why. I expect that it will report the drive has been locked by an application.
Stuart Whitby
Trusted Contributor

Re: Tape alert 16

Uhh, so is the connected system Linux or Windows? Or some wierd combo of both? Whatever - shut down RSM!!!

fuser probably exists in Linux, and doing fuser on the device file will tell you which process is using it. You *may* be able to get similar info using handle.exe from www.sysinternals.com (not sure about tape handles though). Figure out which process is using the tape drive and, if it's safe, kill off that process and see if your drive is still hung up.
A sysadmin should never cross his fingers in the hope commands will work. Makes for a lot of mistakes while typing.
Curtis Ballard
Honored Contributor

Re: Tape alert 16

Regarding the previous posting.

Killing off a process using a drive will often leave the drive locked if that process has issued a SCSI prevent media removal command. If you are going to use that method to test make certain that you do not use a kill -9. Some processes will clean up if you use kill -3. If the drive is unlocked if after you kill a process you have found the problem. If the drive is still locked the problem could still be the process you just killed but it didn't clean up.
Kindermann
Occasional Visitor

Re: Tape alert 16

The OS is SuSE Linux Enterprise Server 8.0 (SLES8) and the System is a Live System. So I wouldnâ t make some experiments with this System. I made the mistake and I used the L&TT and the System felt down after using the tools for few minutes. L&TT is only available for Red Hat and maybe this was the reason for that, but maybe I have a really problem with the motherboard or the RAID Controller. Because I already changed the LTO drive. So I will call HP again and make an escalation with this ca
Curtis Ballard
Honored Contributor

Re: Tape alert 16

Unfortunately we don't have support for Suse Linux with LTT yet. A LTT support ticket will be necessary to get to root cause but it is extremely unlikely that the cause is the drive. The support ticket could prove that but wouldn't be able to point to what application has the drive locked.

I would recommend that you talk to your backup application provider and ask about drive reservations.