Can't backup, volume 1 has been used 100 time(s)

[Harry Yu]

Hi everybody,

I'm having a problem with our HP-UX backup. We automated our backup to tape everyday and recently the backup failed. I received the following message in the log...
-------------------------------------------------
Full backup failed on orlyhp : Sun Jun 6 21:05:17 PDT 2004 .
For more information, see the log file /var/sam/log/br_log.

Output from fbackup:

br_backup: Invoking fbackup. See /var/sam/log/br_log for details.
fbackup(1004): session begins on Sun Jun 6 21:05:01 2004
fbackup(1517): /net not backed up - 'n' option (NFS) not specified
fbackup(3203): volume 1 has been used 100 time(s)
fbackup(3204): volume used more than maximum, do you want to use this volume an?
fbackup(3004): writer aborting
fbackup(1002): Backup did not complete : Reader or Writer process exit
-------------------------------------------------
I didn't setup this backup so I don't know how to override the 100 times limitation. I believe the backup was setup from SAM and is in the cron also. Someone please tell me how can I override this limitation. Also, the log is set to send to the pervious admin who is no longer with us. What should I do to change the system to notify me instead? Thanks alot!!!

[Sridhar Bhaskarla]

Hi Harry,

I would try initializing the medium and see if it works.

mt -t /dev/rmt/0m rew
mediainit /dev/rmt/0m

and then try fbackup.

-Sri

[Harry Yu]

Thanks, Sri.

I tried "mt -t /dev/rmt/0m rew" and it seems fine, but when I do "mediainit /dev/rmt/0m", it returns the following...
-----------------------------------------------
Please wait. The initialisation will take approx. 2 minutes
error in ioctl SET CMD MODE: mode select: I/O error
-----------------------------------------------
well, I assume that first command is to renew the media and not sure what's the 2nd command do. If you've no other suggestion, I'll wait for the night backup to kick in since users are in the database right now. Thanks alot.


- Harry

[Bill Hassell · Accepted solution]

Don't bother with medinint--it could take a very long time. Just write over the tape with tar:

tar cvf /dev/rmt/0m /tmp

That erases to fbackup use counter. The next time you run a backup, fbackup will give you a warning that it could not read a volume header which means the tape was not written by fbackup (which is true). fbackup will continue and start the counter at 1.

Another point: never run fbackup without a config file. You are getting the default of 100 uses for eachy tape and more importantly, you are getting the default settings for the tape drive records. Unless you are using 1/2" reel-to-reel tape (the kind used in 1960's scifi movies), the defaults are very bad for DDS, DLT, Ultrium and other modern tape drives. The option in fbackup is -c. Here is a recommended config file which includes setting the max use count to 200 (more reasonable):

blocksperrecord 512
records 64
checkpointfreq 1024
readerprocesses 6
maxretries 5
retrylimit 5000000
maxvoluses 200
filesperfsm 2000

Create a file with these lines, then add them to the fbackup command line found in root's crontab. Edit root's crontab this way:

cd
crontab -l > crontab.root
vi crontab.root
crontab crontab.root

You will likely find your backup now taking less than 1/2 the time.

[Michael Tully]

If this is a DAT/DDS type tape, do yourself a favour and throw it away and get another tape.

[malay boy]

Hi,
Couldn't agree more on Michael Tully.

regards
mB

[Thierry Poels_1]

Hi,

as said by Bill fbackup uses a counter on the tape to keep track of its used. But once you use another utility to write on the tape (tar, pax, cpio, dd, ...) the fbackup header is gone, and the counter is removed. But you don't realy want to do this, then you loose track of the tape's age, and you'll only get an idea of its age when you get I/O errors ;-)

Depending on the type of tape used you can increase the limit, but within reason: remember that stupid tape can be worth millions in case of disaster.
Personally I opt for a very large pool of tapes: instead of replacing your tapes "often" you cover a large period of time.

regards,
Thierry.

[Shaikh Imran]

Hi,
It's advisable to change the tape and try for better reliability on the backup.

Regards,

[Trond Haugen]

I would reccomend following Pharao Bills recomendation on fbackup config file. Removing the cound dosn't make the tape better.
And give him points for his good advise.

Regards,
Trond

[twang]

In my experience, normally I use tape cartridge for ~50 times in our critical machines.

[Jose Mosquera]

Hi,

This is a warning to consider, please remember that the integrity of your information is always more valuable than the new cost of a tape.

However you can restart this internal counter of the fbackup cleaning the heads of of the tape. You can make it initializing the tape with mediainit, or with a simple "tar" of a file into the tape, then when your you use fbackup the counter again it will begin of zero.

Rgds.

[Harry Yu]

Hi Bill,

Thanks for the help. The commend "tar cvf /dev/rmt/0m /tmp" seems to do the trick. Now, about adding the lines into root crontab. I know how to edit crontab, but I'm not sure where should I be adding those lines. So, I decided to paste the crontab here so you can point it out to me. Thanks alot!

-----------------------------------------------
00 21 * * * /usr/bin/stop.ALL >/dev/null 2>&1
00 00 * * * /usr/bin/start.ALL >/dev/null 2>&1
#00 01 * * * /usr/local/bin/autoall.sh >/dev/null 2>&1
00 02 * * 1-5 /usr/local/bin/mrp.sh >/dev/null 2>&1
00 03 * * 1-5 /usr/local/bin/daily.sh >/dev/null 2>&1
00 03 * * 0 /usr/local/bin/weekly.sh >/dev/null 2>&1
#40 07 * * * /usr/local/bin/excel.sh >/dev/null 2>&1
00 08,10,12,14,16 * * 1-5 /users/prog-list >/dev/null 2>&1
05 21 * * * /usr/sam/lbin/br_backup DAT FULL Y /dev/rmt/0m /etc/sam/br/graphLCAa
20927 admin Y 1 N > /var/tmp/SAM_br_msgs 2>&1 #sambackup
~
---------------------------------------------

- Harry

[Harry Yu]

Thierry,

We uses DLT tapes and we have 3 sets 5 tapes for weekly rotation. I am not very familiar with DLT tapes so I don't know how long it should last. But they seems to be very reliable. Do you have any guideline as how many times should DLT, DDS or any other type of backup media (such as OnStream) be used? Thanks for your reply.


- Harry

[Bill Hassell]

DDS tapes are really gone after about 15-200 passes, assuming you use the cleaning tape regularly. ALWAYS run the STM/EMS monitoring software so you can see tape errors that don't make it to fbackup.

DLT, S-DLT and Ultrium are fairly new and reliability looks good EXCEPT: large temperature changes and dropping the tapes can damage the tapes, recorded or not. It is not clear whether 100 or 200 or 500 backups are the limit, and if you don't keep the tape drive busy, it will reposition excessively, thus reducing the tape's (and drive's) lifetime. From the HP web page on DLT warranty:

"Owing to variations in individual usage, there is no definitive lifetime period that will apply equally to all customers. For warranty purposes, "normal working life" for DLT equates to 500,000 passes or 100 full back ups (whichever occurs first). Although tapes may continue to function beyond these parameters, prolonged usage is not recommended or warranted by HP."

(see http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?objectID=lpg50101)

I'm very unclear as to the difference between a 'pass' and a 'backup'.

[Harry Yu]

Thanks Bill,

I still need your input as where I should insert the lines you gave me earlier. Thanks!

- Harry

[Bill Hassell]

Since you are using SAM, it has (deeply) hidden the fbackup command with a lengthy script called br_backup. It has no man page so you have to look for -c in conjunction with fbackup, then see what it points to. It turns out that br_backup creates a default file in /etc/sam/br and once created, you can edit it. Just change the contents to what you saw in the above posting. Then check your run information tomorrow. You should see a significant performance increase.

[Harry Yu]

Thanks Bill! I found the fbackup_config in /etc/sam/br/ and edited the lines as you advised. I will see how it goes tonight. Thanks again!

- Harry