Hello,

I have a problem with TSM Server 5.3.1.4 on z/Linux Suse SLES 9.
I recently updated from SLES 8 TSM 5.2.x to that version.

The server seems to run fine, except for two things:
- server 5.3 or SLES9 seems to cause much more load than TSM 5.2 under
SLES8 during nightly backups. (Scripts that nagios runs on that machine
over ssh for monitoring purposes are getting timeouts)

- I see following scary entries in /var/log/messages:
(the same error showed up a couple of days ago)

> Aug 23 00:21:32 tsmserv kernel: 0000000019a43980 0000000000407a00 
> 000000000049697c 0000000000355510
> Aug 23 00:21:32 tsmserv kernel:        0000000000130fe4 0000000019a43890 
> 0000000019a43898 0000000000000000
> Aug 23 00:21:32 tsmserv kernel:        00000000000000d0 0000000000000040 
> 0000000000000001 0000000000000000
> Aug 23 00:21:32 tsmserv kernel:        0000000000000000 00000000003580e8 
> 0000000000106fde 0000000019a438e0
> Aug 23 00:21:32 tsmserv kernel:        0700000000040000 0700000000040000 
> 07000040000000d0 07000040000000d0
> Aug 23 00:21:32 tsmserv kernel: Call Trace:
> Aug 23 00:21:32 tsmserv kernel:  [<000000000016a9ca>] 
> __alloc_pages+0x4ce/0x4d4
> Aug 23 00:21:32 tsmserv kernel:  [<000000000016aace>] 
> __get_free_pages+0x4e/0x8c
> Aug 23 00:21:32 tsmserv kernel:  [<0000000028deca00>] 
> IBMTapeReadWriteBuffer+0x90/0x11c [IBMtape]
> Aug 23 00:21:32 tsmserv kernel:  [<0000000028ded246>] 
> IBMTapeDriveWrite+0x2ca/0x18c8 [IBMtape]
> Aug 23 00:21:32 tsmserv kernel:  [<0000000028de7d0a>] IBMTapeWrite+0xee/0x1a0 
> [IBMtape]
> Aug 23 00:21:32 tsmserv kernel:  [<0000000000192b92>] vfs_write+0x102/0x1ac
> Aug 23 00:21:32 tsmserv kernel:  [<0000000000192e28>] sys_write+0xa8/0x118
> Aug 23 00:21:32 tsmserv kernel:  [<000000000011f35c>] sysc_noemu+0x10/0x16
> Aug 23 00:21:32 tsmserv kernel:
> Aug 23 00:21:32 tsmserv kernel: dsmserv: page allocation failure. order:6, 
> mode:0xd0
> Aug 23 00:21:32 tsmserv kernel: 0000000019a438e0 0000000000407a00 
> 000000000049697c 0000000000355510
> Aug 23 00:21:32 tsmserv kernel:        0000000000130fe4 0000000019a437f0 
> 0000000019a437f8 0000000000000000
> Aug 23 00:21:32 tsmserv kernel:        00000000000000d0 0000000000000040 
> 0000000000000001 0000000000000000
> Aug 23 00:21:32 tsmserv kernel:        0000000000000000 00000000003580e8 
> 0000000000106fde 0000000019a43840
> Aug 23 00:21:32 tsmserv kernel:        0700000019a43840 0700000019a43840 
> 07000040000000d0 07000040000000d0
> Aug 23 00:21:32 tsmserv kernel: Call Trace:
> Aug 23 00:21:32 tsmserv kernel:  [<000000000016a9ca>] 
> __alloc_pages+0x4ce/0x4d4
> Aug 23 00:21:32 tsmserv kernel:  [<000000000016aace>] 
> __get_free_pages+0x4e/0x8c
> Aug 23 00:21:32 tsmserv kernel:  [<0000000028decbd4>] 
> IBMTapeInitScatterGather+0x148/0x3b8 [IBMtape]
> Aug 23 00:21:32 tsmserv kernel:  [<0000000028decf5c>] 
> IBMTapeScatterGather+0x118/0x138 [IBMtape]
> Aug 23 00:21:32 tsmserv kernel:  [<0000000028ded286>] 
> IBMTapeDriveWrite+0x30a/0x18c8 [IBMtape]
> Aug 23 00:21:32 tsmserv kernel:  [<0000000028de7d0a>] IBMTapeWrite+0xee/0x1a0 
> [IBMtape]
> Aug 23 00:21:32 tsmserv kernel:  [<0000000000192b92>] vfs_write+0x102/0x1ac
> Aug 23 00:21:32 tsmserv kernel:  [<0000000000192e28>] sys_write+0xa8/0x118
> Aug 23 00:21:32 tsmserv kernel:  [<000000000011f35c>] sysc_noemu+0x10/0x16
> Aug 23 00:21:32 tsmserv kernel: dsmserv: page allocation failure. order:5, 
> mode:0xd0
> Aug 23 00:21:32 tsmserv kernel: 0000000019a438e0 0000000000407a00 
> 000000000049697c 0000000000355510
> Aug 23 00:21:32 tsmserv kernel:        0000000000130fe4 0000000019a437f0 
> 0000000019a437f8 0000000000000000
> Aug 23 00:21:32 tsmserv kernel:        00000000000000d0 0000000000000020 
> 0000000000000001 0000000000000000
> Aug 23 00:21:32 tsmserv kernel:        0000000000000000 00000000003580e8 
> 0000000000106fde 0000000019a43840
> Aug 23 00:21:32 tsmserv kernel:        0700000019a43840 0700000019a43840 
> 07000040000000d0 07000040000000d0
> Aug 23 00:21:32 tsmserv kernel: Call Trace:
> Aug 23 00:21:32 tsmserv kernel:  [<000000000016a9ca>] 
> __alloc_pages+0x4ce/0x4d4
> Aug 23 00:21:32 tsmserv kernel:  [<000000000016aace>] 
> __get_free_pages+0x4e/0x8c
> Aug 23 00:21:32 tsmserv kernel:  [<0000000028decbd4>] 
> IBMTapeInitScatterGather+0x148/0x3b8 [IBMtape]
> Aug 23 00:21:32 tsmserv kernel:  [<0000000028decf5c>] 
> IBMTapeScatterGather+0x118/0x138 [IBMtape]
> Aug 23 00:21:32 tsmserv kernel:  [<0000000028ded286>] 
> IBMTapeDriveWrite+0x30a/0x18c8 [IBMtape]
> Aug 23 00:21:32 tsmserv kernel:  [<0000000028de7d0a>] IBMTapeWrite+0xee/0x1a0 
> [IBMtape]
> Aug 23 00:21:32 tsmserv kernel:  [<0000000000192b92>] vfs_write+0x102/0x1ac
> Aug 23 00:21:32 tsmserv kernel:  [<0000000000192e28>] sys_write+0xa8/0x118
> Aug 23 00:21:32 tsmserv kernel:  [<000000000011f35c>] sysc_noemu+0x10/0x16
> Aug 23 00:21:32 tsmserv kernel:

The server doesn't show anything special during that time in the activity log.
Memory looks fine ( no swapping, still enough memory free ). Filesystem is ok.

Following tape/library packages installed:
ibmatl-6.2.4.0-0
IBMtape-2.0.8-0

We have a 3494 Libray with 3592 tape drives.

Does anyone know if this could be a serious problem?
Has anyone experienced increase of system load after migration to either or 
both SLES9 and TSM 5.3?

thanks in advance,

Volker