Exadata - some helpful MOS notes

I travel all over the place working on Exadata Database Machines. I often do patching and other kinds of work. As you might expect, patching is usually a nervous time for the customer. It's a nervous time for me too! Sitting there and waiting for the storage servers to reboot, or the compute nodes to reboot is nerve wracking.... but it does provide a moment or two for a good conversation with the people sitting around you.

I've found that over and over there are certain MOS notes that I reference time and time again. Some of them I use so much I actually have the note numbers memorized. Isn't that sad? What I decided to do about six months ago was to create a cheat sheet of notes that I use often. That way I can just cut and paste the note number into MOS and *poof* - there it is. I also find this handy when MOS has timed out my session and the note I previously had up, is no longer there (which I find irritating).

I've removed the internal notes from this list, since you would not have access to them anyway. If you look for one of these notes and don't have access to them because they are not public or the note does not appear to exist, please let me know as a comment here.

I hope you find these notes helpful. If you have any you use a lot, please leave a comment here and I'll take a look at it and update the list if it's list worthy! Also, I'll update this list from time to time here so it stays current.

Here is my Exadata Note cheat sheet...

As my friend Tim Hall says on his site oracle-base.com - Hope this helps!

Information Center

1306791.2 – Information Center: Oracle Exadata Database Machine

Best Practices

757552.1 - Oracle Exadata Best Practices 

 

Operational

1473002.1 – Using dbserver_backup.sh to backup compute nodes. 

(Section 7 – run backup – Section 8 – roll back)

1538068.1 – Remove partition if dbserver_backup.sh fails.

1428394.1 – Password stuff (pam_talley2)

1269133.1 – Password stuff (pam_talley2 – resetting the /etc/pam.d/sshd file.

1093890.1 – Shutdown and startup Exadata and Compute nodes on rack.

1446274.1 – ILOM command reference (startup and shutdown Exadata from ILOM).

1520896.1 – DBFS Configuration Health Check

1054431.1 – Configure DBFS on Exadata Checklist

Cell Related

1306635.1 – Replacement of flash – how to check firmware and status. Resetting status.

Patching Notes

1262380.1 – Master note on Exadata patching.

888828.1 – Master note on Exadata supported versions. Really important note!

1473002.1 – Using ULN to install server patches with YUM. This covers both the compute nodes and cell node kernels.

1545789.1 – ISO install Cheat Sheets

1136544.1 – Relinking notes.

1553103.1 – Exadata Database Server Patching using the DB Node Update Utility

Notes on patching related problems I’ve encountered

1.       Node rebooting after backup of compute node or after patch is complete.

# This note discusses and issue with compute node backups. The method of backup in the documentation and the recommended method of backup

# of the compute nodes before a patch (as per 1473002.1 – see above in operational section) are different. You can run into this problem if they have been

# using the manual backup (in the docs) and then change to using db_server_backup.ksh.

1538068.1 -  Exadata compute node continuously booting with kernel panic

2.       Node won’t boot after starting patching process.

# Related to the problem above – if they are doing manual backups they have to drop and re-create a partition. I’ve had one customer who accidently dropped the wrong partition (the boot partition). If they started patching and all of a sudden a node panics or won’t boot, check on this as the likely issue. You can log onto the ILOM and use lvdisplay as seen here:

lvdisplay|grep 'LV Name'
  LV Name                /dev/VGExaDb/LVDbSys1  -- boot partition
  LV Name                /dev/VGExaDb/LVDbSys2 – backup boot partition
  LV Name                /dev/VGExaDb/LVDbSwap1
  LV Name                /dev/VGExaDb/LVDbOra1

if LVDBSys1 does not show up, or it’s empty – then you have hit this problem. Open SR and you will need to do a restore of that node.

Current patches notes (As of 9/25/2013) on patches from 888828.1

Latest Releases and Patching News

• Oracle Database 11.2.0.4 is supported on Exadata Database Machine.
• New 11.2.0.3 Database release - Patch 16869210 - 11.2.0.3 recommended Quarterly Database Patch for Exadata Jul 2013 - 11.2.0.3.20
• New QFSDP release - Patch 16784347 - Quarterly Full Stack Download Patch (QFSDP) Jul 2013
• New 11.2.0.2 Database release - Patch 16824987 - 11.2.0.2 Bundle Patch 21 for Exadata
• Oracle Database 12c is supported on Exadata Database Machine.  See Document 1537407.1 for details.
• In Exadata release 11.2.3.2.0 and 11.2.3.2.1, the Oracle Linux Unbreakable Enterprise Kernel (UEK) and non-UEK kernels were both shipped and switching between them was supported on 2-socket database servers. The large majority of Exadata customers installed the default UEK kernel.  UEK was the only option on 8-socket database servers.  Starting with the next Exadata release after 11.2.3.2.1,  the only installable kernel will be the UEK2 kernel. The UEK2 kernel is specifically tuned for performance and availability for Engineered Systems.
• New OneCommand release - Patch 16921136 - Supports 11.2.0.3.19 and 11.2.0.2 BP20
• New Exadata Storage Server release - Patch 14522699 - Exadata Storage Server software 11.2.3.2.1 (Note 1485475.1)
• Minimum version required if using Write-back Smart Flash Cache
• Oracle Solaris 11.1 is supported on database servers.