12.7 Case studies

 < Day Day Up > 

12.7 Case studies

In this section, we include two practical examples for using reported data to solve a problem.

12.7.1 Linux console goes to HOLDING state

Suppose you have the Linux console is logged on and a terminal is attached to the console. If the terminal is in HOLDING mode and the CP run option is OFF (which is the default), all processing will be stopped until the screen is cleared. As shown in Figure 12-12, there is a gap in the data received, because the processing for this VMID has been suspended until the Hold was cleared. The hold started in the 15:48 minute and lasted till 15:54. A look at any other screen, including Linux reports, will show that nothing was happening in either VM or Linux for this Linux guest.

click to expand
Figure 12-12: Showing a gap in data

This is important to know, as this is something that can happen when the Linux console is a 3270 terminal.

To correct this problem, the CP SET RUN ON command can be issued in the Linux VM ID profile. SET RUN ON prevents the virtual machine from going into CP READ when the virtual console is reconnected.

If you want to prevent the screen from going into HOLDING, you can use the TERM settings. However, if you set TERM MORE 0, you may be unable to enter commands on the tn3270 session when a runaway process floods the console. TERM HOLD OFF may be a good command to issue also.

12.7.2 DASD hot spot

When a DASD has a "hot spot" (that is, a great deal of I/O to the same VM volume), it can be difficult to trace the VM volume to a Linux mount point to determine what can be done to resolve the issue. In this section, we explain how you can use ESAMON to track down a hot spot.

Use the ESAMON ESADSD2 screen to locate a hot spot on the disk. We found one on the 8220 device. Using the zoom PF key, the screen in Figure 12-13 was produced.

click to expand
Figure 12-13: ESADSD2 and ESAUSEK screens

Device 8222 was where most of the I/O was taking place. By using the split function and going to the ESAUSEK screen, we can see that the 8222 has a vaddr of 211 and is owned by LinuxA.

Now that the specific machine and disk are known, the actual Linux filesystem can be found. The first step is to find out what the Linux device name is for the VM minidisk. This is done by looking at the DASD device table to identify the device. The way to look at the DASD device table is to use the Linux cat /proc/dasd/devices command shown in Figure 12-14. The output from this command shows that vaddr 211 is dasdr; in Linux terms, this is /dev/dasdr.

start figure

  linuxa:~# cat /proc/dasd/devices  0200(ECKD)at ( 94:  0) is  dasda     :active at blocksize:4096,600840 blocks,2347 MB  0201(ECKD)at ( 94:  4) is  dasdb     :active at blocksize:4096,600840 blocks,2347 MB  0202(ECKD)at ( 94:  8) is  dasdc     :active at blocksize:4096,600840 blocks,2347 MB  0203(ECKD)at ( 94: 12) is  dasdd     :active at blocksize:4096,600840 blocks,2347 MB  0204(ECKD)at ( 94: 16) is  dasde     :active at blocksize:4096,600840 blocks,2347 MB  0205(ECKD)at ( 94: 20) is  dasdf     :active at blocksize:4096,600840 blocks,2347 MB  0206(ECKD)at ( 94: 24) is  dasdg     :active at blocksize:4096,600840 blocks,2347 MB  0207(ECKD)at ( 94: 28) is  dasdh     :active at blocksize:4096,600840 blocks,2347 MB  0208(ECKD)at ( 94: 32) is  dasdi     :active at blocksize:4096,600840 blocks,2347 MB  0209(ECKD)at ( 94: 36) is  dasdj     :active at blocksize:4096,600840 blocks,2347 MB  020a(ECKD)at ( 94: 40) is  dasdk     :active at blocksize:4096,600840 blocks,2347 MB  020b(ECKD)at ( 94: 44) is  dasdl     :active at blocksize:4096,600840 blocks,2347 MB  020c(ECKD)at ( 94: 48) is  dasdm     :active at blocksize:4096,600840 blocks,2347 MB  020d(ECKD)at ( 94: 52) is  dasdn     :active at blocksize:4096,600840 blocks,2347 MB  020e(ECKD)at ( 94: 56) is  dasdo     :active at blocksize:4096,600840 blocks,2347 MB  020f(ECKD)at ( 94: 60) is  dasdp     :active at blocksize:4096,600840 blocks,2347 MB  0210(ECKD)at ( 94: 64) is  dasdq     :active at blocksize:4096,600840 blocks,2347 MB  0211(ECKD)at ( 94: 68) is  dasdr     :active at blocksize:4096,600840 blocks,2347 MB 

end figure

Figure 12-14: cat /proc/dasd/devices

From this point we can look at the filesystem table in /etc/fstab to see if /dev/dasdr is defined in it. If it is, then the search for the mount point is complete. The fstab would indicate where the device is mounted, and a guess as to what it is being used for can be made.

However, in this case the device is not in the filesystem table. So the next place to look is in the Logical Volume Manager to find out which logical volume the device is assigned to. This can be somewhat challenging, as the device can be used in many different logical volumes.

To find out where the device is in the logical volume, the lvdisplay command is used, as shown in Figure 12-15 on page 313. From this command, we can tell the following information:

  • We know from the ESAMON ESADSD2 screen in Figure 12-13 on page 311 that the device is a 3390-3. This means that it has 585 physical extents (PEs).

  • The lvdisplay command shows that /dev/dasdr is used in the /dev/domino/transloga logical volume.

  • It also shows that all 585 physical extents are being used by this logical volume, and that there are no other devices used in this volume.

start figure

  linuxa:~#lvdisplay -v /dev/domino/*|less  :  :  ---Logical volume ---  LV Name              /dev/domino/transloga  VG Name              domino  LV Write Access      read/write  LV Status            available  LV #                 4  #open                1  LV Size              2.29 GB  Current LE           585  Allocated LE         585  Allocation           next free  Read ahead sectors   1024  Block device         58:3      ---Distribution of logical volume on 1 physical volume  ---      PV Name                 PE on PV     reads      writes      /dev/dasdr1             585          187985     1805150  :  : 

end figure

Figure 12-15: lvdisplay command

In this case, we were fortunate that the logical volume maps very easily to the physical device; we had only a single logical volume.

The final step is to return to the filesystem table (/etc/fstab) and see where the logical volume is mounted. As shown in Figure 12-16, the logical volume with /dev/dasdr (LinuxA's 211 minidisk) is mounted as /domserva/notesdata/translog.

start figure

  linuxa:~ # cat /etc/fstab  /dev/dasda1          /                    ext3      defaults            1 1  /dev/dasdb1          /opt                 reiserfs  defaults            1 2  /dev/domino/domserva /domserva            ext2      defaults            0 2  /dev/domino/transloga /domserva/notesdata/translog ext2 defaults        0 2  /dev/domino/mail1    /notesmail/nbmail1   ext2      defaults            0 2  /dev/domino/mail2    /notesmail/nbmail2   ext2      defaults            0 2  devpts               /dev/pts             devpts    mode=0620,gid=5     0 0  proc                 /proc                proc      defaults            0 0  /dev/dasdu1          swap                 swap      defaults            0 0  /dev/dasdv1          swap                 swap      defaults            0 0 

end figure

Figure 12-16: Filesystem table

From the name of the mount point, a guess can be made as to the purpose of the device and to why the I/O rate is high. In this case, the mount point is exactly as it has been described, and it is the transaction log disk for Domino.

 < Day Day Up > 

IBM Lotus Domino 6. 5 for Linux on zSeries Implementation
IBM Lotus Domino 6.5 for Linux on Zseries Implementation
ISBN: 0738491748
EAN: 2147483647
Year: 2003
Pages: 162
Authors: IBM Redbooks

flylib.com © 2008-2017.
If you may any questions please contact us: flylib@qtcs.net