Login or Register to Ask a Question and Join Our Community


Solaris

flarcreate disk error 28

Login to Discuss or Reply to this Discussion in Our Community

 
Thread Tools Search this Thread
Operating Systems Solaris flarcreate disk error 28
# 8  
Old 12-28-2006
You have a problem there, your root filesystem is full and since its a production server that needs to be up 24/7 you better clean it up.
Look in /var/adm for old log files that you might be able to remove. Check for any core dumps in /var/crash aswell. syslogs can also get big.

If you want to flare the root partition you can not use /tmp as it is only 7GB and you root partition is 10GB.

Use absdata5 or absdata6 as you have 51GB available there.
Tornado
# 9  
Old 12-28-2006
I just had a look at your first post again..
You were using /var/tmp which is on your root partition, I guess thats why your root partition is full.

If I was you I would cleanup /var/tmp ASAP and in future before trying something like this on a 24/7 production server make sure you know what it is you are doing..
Tornado
# 10  
Old 12-28-2006
I agree with the others. Root (/) is full and you must clean it up. Check /var/tmp and make sure you clear out old flarcreate files, unneeded temp files (it is a temporary directory; anything not reasonably current should be able to be deleted). Obvious additional choices are /var/adm and /var/log. Check your logs. Make sure they're being rolled (moved daily to something like syslog.0 or syslog.20061227) then make sure they're compressed (looking like syslog.0.gz or syslog.20061227.gz). If they are not being rolled and compressed, check your scripts to get that automatically taken care of. Also remove old log files. That'll depend on company policy. Generally 45 to 90 days is what's used. Next are home directories. Run a du -ks * | sort -n in /export/home or /home depending on your setup. That'll give you the largest directories at the bottom and you can address them for cleanup.

Another issue is your planning on flarcreate. In my location, we use flarcreate to create a bare-metal OS level backup, not a backup of the entire system. So we'll exclude all the oracle file systems for instance. In looking at your setup, you have about 230 GB of data on this system and you're trying to put it all into a filesystem that can have about 8 GB free at best. Regular backups need to be done on the production data and not by using flarcreate. An external tape drive for instance.

Another piece of advice. When posting to the forums, please enclose your output in a code tag (that's the # tag in the bar above your message). It takes the fixed font screen output you have and leaves it fixed font rather than proporational. For example, I whipped up a quick perl script to massage your df -k output:

Code:
                 Filesystem     kbytes       used      avail   capacity                   Mounted
          /dev/dsk/c1t0d0s0   10080200   10072392          0       100%                         /
                      /proc          0          0          0         0%                     /proc
                         fd          0          0          0         0%                   /dev/fd
                     mnttab          0          0          0         0%               /etc/mnttab
                       swap    7787888        136    7787752         1%                  /var/run
          /dev/dsk/c1t1d0s4   20645791   18699829    1739505        92%                   /oracle
          /dev/dsk/c1t2d0s5   20645791   13964517    6474817        69%                     /work
                       swap    7788440        688    7787752         1%                      /tmp
          /dev/dsk/c1t4d0s6   35009161   14362368   20296702        42%                    /oper2
          /dev/dsk/c1t3d0s6   35009161    5064115   29594955        15%                    /oper1
          /dev/did/dsk/d6s4     288509       4525     255134         2%   /global/.devices/node@1
 /dev/vx/dsk/utldg/oradata1   10321884    2433607    7785059        24%                 /oradata1
 /dev/vx/dsk/utldg/oradata2   10321884    5102382    5116284        50%                 /oradata2
 /dev/vx/dsk/utldg/oradata4   10321884     734421    9484245         8%                 /oradata3
 /dev/vx/dsk/utldg/absdata1   51609487    5173737   45919656        11%                 /absdata1
 /dev/vx/dsk/utldg/absdata2   51609487   41031377   10062016        81%                 /absdata2
 /dev/vx/dsk/utldg/absdata3   40254451    5162473   34689434        13%                 /absdata3
 /dev/vx/dsk/utldg/absdata4   51609487   33419515   17673878        66%                 /absdata4
 /dev/vx/dsk/utldg/absdata5   51609487      51217   51042176         1%                 /absdata5
 /dev/vx/dsk/utldg/absdata6   51609487      51217   51042176         1%                 /absdata6
 /dev/vx/dsk/utldg/absdata7   51609487    4486385   46607008         9%                 /absdata7
/dev/vx/dsk/utldg/absindex1   30965686   20520817   10135213        67%                /absindex1
/dev/vx/dsk/utldg/absindex2   30965686      30737   30625293         1%                /absindex2
/dev/vx/dsk/utldg/absindex3   30965686      30737   30625293         1%                /absindex3
/dev/vx/dsk/utldg/orabackup   10321884    8120756    2097910        80%                /orabackup
/dev/vx/dsk/utldg/operdata1   51609487   32162386   18931007        63%                /operdata1
/dev/vx/dsk/utldg/operdata2   51609487   39924869   11168524        79%                /operdata2
/dev/vx/dsk/utldg/operdata3   51609487   10907848   40185545        22%                /operdata3
         /dev/did/dsk/d34s4     288509       4430     255229         2%   /global/.devices/node@2

As you can see, it's a lot more readable this way.

Good luck.

Carl
# 11  
Old 02-15-2007
So I want OS level volume backup
In my location, we use flarcreate to create a bare-metal OS level backup,

I want just OS level root (it means internal disk). What will be the command

to volume backup just OS level to tape or drive. (it means when system broke, I

recover OS level by flarcreate and use another tape for DB recovery).

Is it ok I just flarcreate only internal disk? what is the actual meaning OS level

backup? It will be ok I can recover by two step, 1 flarcreate 2 coldbackup

tape.

Is it something like

flarcreate -n "utlabsflar" -c -S -R / -x /oradata1 -x /oradata2 -x /oradata3 -x /absdata1 -x /absdata2 -x /abadata3 -x /absdata4 -x /absdata5 -x /absdata6 -x /absdata7 -x /absindex1 -x /absindex2 -x /absindex3 -x /orabackup -x /operdata1 -x /operdata2 -x /operdata3 -t /dev/rmt/0n
Last edited by 197oo302; 03-01-2007 at 01:59 AM..
# 12  
Old 03-01-2007
I complete flarcreate
I finished backup of my server only internal disk.

some question is why external disk has error message during backup?

and other message is "too large to archive in current mode" how to avoid this?

This is my log of flarcreate and flar info.

absapp@nepalabs1 # sh volumebackup.sh
WARNING: hash generation disabled when using tape (-t)
Determining which filesystems will be included in the archive...
Creating the archive...
cpio: cpio: absdata1/nccbs_ref.dbf: too large to archive in current mode
cpio: cpio: absdata2/nccbs_app.dbf: too large to archive in current mode
cpio: cpio: absdata2/nccbs_app6.dbf: too large to archive in current mode
cpio: cpio: absdata4/expdat.dmp: too large to archive in current mode
cpio: cpio: absdata6/vol072017.flar: too large to archive in current mode
cpio: cpio: absdata7/nccbs_app_temp.dbf: too large to archive in current mode
cpio: cpio: absindex1/nccbs_idx3.dbf: too large to archive in current mode
cpio: Error with lstat() of "oper1/absopr/cpm/ST_TFD/UTL_cef988_cefb3f_20070226_235004.TXT", errno 2, No such file or directory
cpio: Error with lstat() of "oper1/absopr/cpm/ST_TFD/UTL_cef5a0_cef793_20070226_221148.TXT", errno 2, No such file or directory
errno 2, No such file or directory
of "oper1/absopr/cpm/ST_TFD/UTL_cdf389_cdf57c_20070226_100108.TXT", errno 2, No such file or directory
cpio: cpio: oper1/temp/utlapp_0209.dmp: too large to archive in current mode
cpio: cpio: operdata1/nccbs/log/cpm/cpm_up_20631010_104709.log: too large to archive in current mode
cpio: cpio: operdata1/nccbs/log/cpm/cpm_up_20631011_124141.log: too large to archive in current mode
cpio: cpio: operdata2/orabackup/export/expdat.dmp: too large to archive in current mode
cpio: File size of "oracle/ora9i/920/network/log/listener.log" has increased by 183
cpio: cpio: oracle/ora9i/seo/nepal/20040209/utlapp_0209.dmp: too large to archive in current mode
cpio: cpio: oradata2/ORA9/undotbs01.dbf: too large to archive in current mode
155845519 blocks
199 error(s)
Archive creation complete

nepalabs1:/oper1/absapp> flar info -t /dev/rmt/0
files_archived_method=cpio
creation_date=20070228032613
creation_master=nepalabs1
content_name=utlabsflar
creation_node=nepalabs1
creation_hardware_class=sun4u
creation_platform=SUNW,Sun-Fire-880
creation_processor=sparc
creation_release=5.8
creation_os_name=SunOS
creation_os_version=Generic_108528-18
files_compressed_method=compress
content_architectures=sun4u
# 13  
Old 03-08-2007
Why not just use ufsdump ?
You can use something like rkbackup to make it easier..
http://linuxmafia.com/faq/Admin/tape-backup.html
Tornado
# 14  
Old 03-08-2007
I mean solaris volume backup
Your thing is linux backup.

I want solaris volume backup.

I use ufsdump as coldbackup of my DB and directory.

And I need volume of solaris OS level backup. not linux.

I almost done. What I ask is my script is ok or not?

I haven't recover from backup. I think it is ok.

Can you confirm that if my system crash I can recover from backup?
Login or Register to Ask a Question

Previous Thread | Next Thread

10 More Discussions You Might Find Interesting

1. Solaris

Flarcreate and excluding directories with software installed

Got a solaris 10 server that I want to take a flar of in order to use to build a new server. Works fine. Only trouble is there is software on the original server that is not needed on the new server - it will be uninstalled. I was thinking of creating the flar from the original server and... (0 Replies)
Discussion started by: psychocandy
0 Replies

2. Solaris

Save flarcreate on another server through ssh

hi, i would like to know if it is posible to create the flash archive and in the same line add another command to save it in another server through SCP. flarcreate -n "sv756flash" -c -X excludefilelist |... i appreciate your help. thanks a lot. (4 Replies)
Discussion started by: pabloli150
4 Replies

3. Solaris

Cloning with flarcreate .... and back

Hello Forum, I'd like to clone a Solaris 10 System. Well, I read a lot about flarcreate. Anyway I only read how to _create a flash archive by using flarcreate. But I don't know how to copy the data back to another harddisk from an ar:confused:chive to get a second runing system. ... (4 Replies)
Discussion started by: borsti007
4 Replies

4. Solaris

flarcreate help.

Hi, I have to clone a system and i would like to use flarcreate. My question is regarding excluding some NFS mounts and some directories under /. If somebody have done this earlier please advice on the following procedure. Create a file like /exclude_list and put the following lines in it ... (1 Reply)
Discussion started by: uxadmin007
1 Replies

5. Solaris

How to take solaris OS backup with "flarcreate" command

I am trying to take Solaris OS backup with the following command "flarcreate -S -n / /server11/inblr.flar", but this is taking complete root / (which includes my othe mount points also), Can any one suggest how to take only solaris OS backup and restore proceedure Thanks in advance (4 Replies)
Discussion started by: raj.chinnu
4 Replies

6. Solaris

System copy with flarcreate

Hi, i'm trying to make a system copy from one server to another (solaris 9), they have the same configuration. I've tried to create a flash archive using the flarcreate but the flash file size is 5.06 approx. and the method the flarcreate is using is CPIO and he cannot handle sizes over 4 GB. Is... (2 Replies)
Discussion started by: pasalagua
2 Replies

7. Solaris

flarcreate

found my notes got it flarcreate -n jumpstart -x /jumpstart/flash /jumpstart/flash/jumpstart.flar i think this will omit that directory and save it there correct? (3 Replies)
Discussion started by: deaconf19
3 Replies

8. Solaris

Flarcreate question (/etc/hosts missing)

Hi there I have created a flash archive using flarcreate and rolled out via jumpstart and everything seems fine, however when I roll out the archive to another box, everything is fine apart from the /etc/hosts file which is the same as the masters hosts file (i.e. it still has the original... (1 Reply)
Discussion started by: hcclnoodles
1 Replies

9. Solaris

flarcreate to remote tape device

Hi, Someone knows hot to create a flash archive to a remote tape device? Here is my try: $flarcreate -n nereus -S -t lisdev:/dev/rmt/1n WARNING: hash generation disabled when using tape (-t) current filter settings Creating the archive... ERROR: Unable to work archive. Regards (4 Replies)
Discussion started by: spacewalker
4 Replies

10. Solaris

flarcreate: command not found

Hi, I have a solaris 5.6 box, and i need to create a Flash Archive for DRP porposes. When i execute the falarcreate commnad the system returns "flarcreate: command not found" . I search the entire system (find . -name *flar*) for that comand and i can find it. Solaris 5.6 donīt have this... (1 Reply)
Discussion started by: spacewalker
1 Replies
Login or Register to Ask a Question