confused about size of a node

[kongfranon]

I am not sure which section to post this so hopefully this is ok.

So I have a linux host that is currently using 2.4 TB. I am backing it up for the first time on my new TSM server.

I am confused about statistics, I started the backup last night @ around 8 pm, and it is still running now so far backup is taking 16hours+:

Sess Comm. Sess Wait Bytes Bytes Sess Platform Client Name
Number Method State Time Sent Recvd Type
------ ------ ------ ------ ------- ------- ------- -------- ------------------
2,617 Tcp/Ip RecvW 0 S 14.6 K 5.2 T Node Linux NSRV01
x86-64

Since last night it says it sent 5.2 TB is that correct?

I don't see how that is possible since it is 2.4 TB?

I just ran an audit licenses command and then see the following when I type q auditocc:


Node Name Backup Archive Space-Managed Total
Storage Storage Storage Used Storage
Used (MB) Used (MB) (MB) Used (MB)
----------------------------------- --------- --------- ------------- ---------

NSRV01 228,098 0 0 228,098


and now when I run q filespace:


Node Name Filespace FSID Platform Filespace Is Files- Capacity Pct
Name Type pace Util
Unicode?
--------------- ----------- ---- -------- --------- --------- ----------- -----
NSRV01 /vm1fs2 1 Linux EXT4 No 1,098 GB 16.3
x86-64
NSRV01 / 2 Linux EXT4 No 62,365 MB 10.7
x86-64
NSRV01 /boot 3 Linux EXT4 No 484 MB 17.9
x86-64
NSRV01 /offieo- 4 Linux EXT4 No 1,000 GB 5.1
dnfs x86-64
NSRV01 /dbnfs 5 Linux EXT4 No 4,317 GB 50.6


so in theory maximum size of server is roughly 5.4 TB but only 2.4 TB is being used at the moment.


Does any of this make sense? Why would it backup 5 TB when only using 2.4? and then when I run auditocc it shows only 230 GB ?


Am I interpreting this correct?

Thanks

 

[Trident]

Hi,

Do you have any open files, like mysql or similar? The number 5.2tb is bytes transferred, not backed up. Look in the error log for clues.

 

[kongfranon]

Hi,

Do you have any open files, like mysql or similar? The number 5.2tb is bytes transferred, not backed up. Look in the error log for clues.

Oh ok thanks! No this server is just exports of oracle backups, no live db files

This is what I see over and over


06/15/2012 15:32:49 ANS9020E Could not establish a session with a TSM server or client agent. The TSM return code is -53.
06/15/2012 15:32:49 ANS1029E Communication with the TSM server is lost.

 

[kongfranon]

Hi,

Please check one of the following:

-dsm.sys/dsm.opt for the SERVERNAME
-dsm.sys Verify that the TCPSERVERADDRESS and TCPPORT settings have the correct values for your Tivoli Storage Manager server
-dsm.sys for the correct registered NODENAME is being used

Please make sure that it is coded correctly.

Thanks that is all verified, and if i run dsmc at the command line it logs in correct and can run commands like q filespace or incr. If those were set in correct I would think I would not be able to do that from the client right?

 

[smajl]

since I see in list of filespaces /vmwarenfs2 /devdbnfs I assume there 2 filesystems contains NFS shares with files opened by some clients/applications.
dsmc backup up those files but if file is changed during processing it tries to backup them again. see parameter chaingingretries 0-3 in dsm.sys or specification of copygroup. q se counts sum of transferred bytes even for incompletly saved files.

 

[kongfranon]

since I see in list of filespaces /vmwarenfs2 /devdbnfs I assume there 2 filesystems contains NFS shares with files opened by some clients/applications.
dsmc backup up those files but if file is changed during processing it tries to backup them again. see parameter chaingingretries 0-3 in dsm.sys or specification of copygroup. q se counts sum of transferred bytes even for incompletly saved files.

That makes sense...I will take a look in the mean time I have opened a ticket with IBM, they keep thinking it a network issue, I did every single test I can think of transferring files directly outside of TSM and had no issues at all. I looked at the cisco switch ports, no errors or mis config with duplex / speed

I am still waiting to hear back from them.

 

[smajl]

you should see something like this in dsmsched.log:
06/20/12 14:27:44 Normal File--> 576,229,376 /data/mail/filename.nsf Changed
06/20/12 14:29:45 Retry # 1 Normal File--> 576,229,376 /data/mail/filename.nsf Changed
06/20/12 14:33:50 Retry # 2 Normal File--> 576,229,376 /data/mail/filename.nsf Changed
06/20/12 14:39:03 Retry # 3 Normal File--> 576,229,376 /data/mail/filename.nsf Sent

This would cause in summary Backed up data ~550GB but Transferred data ~1.5TB (depends on activity). This is normal behavior.

If during "send to tsm" operation file is changed (by OS/DB/app...) this change is being detected by TSM client and it tried to re-send it again (whole file). This repeats until file is successfully backed up (=no change in file during backup) or changingretries is reached and file si reported as "Failed"

 

[kongfranon]

Finally got it working where it it completes the backup in 6 hours instead of 18+...it was vmware that was changing the vmdk files. The vmware admin changed some settings on his side, and now no more changed vmdk files.

Thanks everyone for the help

 

[kongfranon]

I have another question related to a different node...and I realized my original question never got answered since we got sidetracked.

So I have an oracle db server now that shows 3 TB in the auditocc and then in the filespace shows different. I really don't know which one to go by? It doesn't make sense to me that it has 3 TB, when the server does not have that much space?


Node Name Filespace FSID Platform Filespace Is Files- Capacity Pct
Name Type pace Util
Unicode?
--------------- ----------- ---- -------- --------- --------- ----------- -----
DB01 /mnt/raid1 1 Linux EXT3 No 492 GB 69.9
x86-64
DB01 / 2 Linux EXT3 No 19,838 MB 75.0
x86-64
DB01 /a1_01 3 Linux EXT3 No 59,515 MB 27.2
x86-64
DB01 /opt 4 Linux EXT3 No 19,838 MB 38.6
x86-64
DB01 /boot 5 Linux EXT3 No 493 MB 11.1
x86-64

tsm: BACKUP02>q auditocc qadb01
License information as of last audit on 07/02/2012 at 08:00:02 AM.

Node Name Backup Archive Space-Managed Total
Storage Storage Storage Used Storage
Used (MB) Used (MB) (MB) Used (MB)
----------------------------------- --------- --------- ------------- ---------
DB01 3,077,994 0 0 3,077,994



Maybe the better question is what is difference between auditocc and filespace when it shows size? Since they don't seem to match up?

 

[rmazzon]

The filespace table (q filespace comamnd) shows the drive/file system information of the client. The q auditoccupancy commands returns the amount of backup data being managed for that client by the TSM server. Your occupancy should always be larger than the sum of the filespace sizes (capacity*pct_util/100) under most circumstances. I am ignore the effects of deduplication.

Use the q auditocc command, or the occupancy table itself to determine how much data you are managing for any node.