Thursday, April 17, 2014

11.2.0.3 VIP/SCAN VIPs unreachable causing connection issue

There are always new things I find on My Oracle Support, but one interesting find was "11.2.0.3 VIP/SCAN VIP is Not Pingable After Failover Leads to Connection Issue" from note 1379498.1 because it reminds me of the problems we had with our ZFS connection after upgrading to 11.2.0.3 on Exadata.  Of course there is a workaround and the next patch of 11.2.0.3 contains the fix, as well as 11.2.0.4, but it would not hurt at all to make sure your platform is on the right version to avoid bug 13440962.

Wednesday, April 16, 2014

R12 Payables: Running Accounts Payable Trial Balance Report by Customer

From the "did you know?" file, comes an article on My Oracle Support titled "Customers Are Available In The "Accounts Payable Trial Balance Report" Parameter" in note 1629022.1 explaining that the new Payables architecture in R12 allows the report to pull in Customers as well as Suppliers. Of course this makes sense with the changes, but since I haven't had much of a chance to get hands-on experience with this change I am sure that I will keep running into these tidbits of information that I will be sharing with you along the way!

Tuesday, April 15, 2014

R12: Filling up the $APPLTMP directory

The other day we had our mount volume monitoring went off indicating that we had a mount starting to fill up, and they were a bunch of randomly named files all with the extension of .t for some reason.  So off to My Oracle Support our DBA went to create a ticket and the answer came back that we should look at note 730342.1 for "Files Get Created In The $APPLTMP Directory And Fills The Drive" to take the requested actions to resolve this issue.  While yes following this advice will help get us out of the weeds, I think is a band-aid for what the root cause might be.  I'll follow up with more on that when I research it a bit more, and maybe another issue I might have caught while investigating this.  :}

Monday, April 14, 2014

Heartbleed, OpenSSL, Exadata, R12 & You

Do you know what Oracle products are affected by the Heartbleed OpenSSL bug?  Well, you do now after reading this public article on the Oracle Technology Network!  The good news is that some large products such as Exadata and EBS R12 are not affected at their core…the caveat to that I see is unless they are running on a version of the Linux OS that is listed as affected.

Sunday, April 13, 2014

Weekend Learning: GoldenGate replication after flashing back

Getting further into the guts of what makes the Oracle Database product, this article "How to recover GoldenGate replication after flashing back (or recovering to a precious point at) source database?" in MOS note 1626736.1 helps to get a bit more understanding.  It is a very good step by step explanation of how to flashback to a specific transaction SCN and then changing the GoldenGate extract so that it catches back up from that point.

Saturday, April 12, 2014

R12 Receivables: Unapplied Receipts Journal Report has wrong data

In what seems like a fairly important issue, in R12 the Unapplied Receipts Journal Report shows as off from the R11 counterpart.  In My Oracle Support note 1628501.1, the cause is listed as a query missing a where clause but luckily it also lists the patch which has the fix available as well.

Friday, April 11, 2014

R12: Expense Reports get stuck after approval

Today I got to deal with a new A/P issue that we've reported to Oracle development, where APEXP workflows will spawn off the AME approval process successfully but after approval the workflow doesn't report back to the parent workflow.  Instead of having the child process report back, the parent workflow sits there at the Block stage until you retry the most recent Active stage which was the AP Approval Process and then the parent workflow goes all the way to the end allowing the invoice to be imported and paid.  We're not sure why certain expense reports fail in this way, as most succeed without issue, but I wanted to put this out there so you can try the workaround and I'll follow up with what we find as a resolution or cause.

Thursday, April 10, 2014

Hyperion: Planning PSU application breaks Financial Reporting Studio

Our team had applied a Hyperion Planning PSU in DEV to try and resolve a report issue, but they reported that they started receiving an error in Financial Reporting Studio starting with "HARSnapin Initialize() Error -2147467259 - ; nested exception is: java.rmi.UnmarshalException: error unmarshalling return" when they tried to open any report.  The key here was the PSU that was deployed, as My Oracle Support note 1518423.1 was applied to "refresh" Financial Reporting Studio but if there wasn't a new PSU deployed note 1311069.1 could have applied where a file was corrupted on the server that needed replacing.

Wednesday, April 9, 2014

Conference reminders!

Don't forget to check Twitter for the hashtag #C14LV or click here for the Twitter link to keep up with this year's Collaborate conference from OAUG!

Looking to the future, the call for proposals deadline for Oracle Open World this year is fast approaching on April 15th so you only have a few more days to get your abstracts polished and submitted!

Tuesday, April 8, 2014

Memory leak in ZFS Storage Appliance driver

As you may remember, one of my first posts was on an outage with our 7420 ZFS array, so you may need to know about MOS note 1635002.1 for "A Memory Leak in the nxge(7D) Driver for ZFS Storage Appliance Software May Cause ZFS Storage Appliance Service Outages" as it sounds a LOT like the situation I encountered years ago.  The good news?  It affects several versions of the ZFS appliance, but they all end in 10 and not 20, so I know this isn't something new that we have to worry about for our installation!

Monday, April 7, 2014

Trying to commit and getting ORA-25408 instead

Today finds me needing to change data in QA as part of a project, and running what I was asked to I received this error back when I went to issue a commit:

ORA-25408: can not safely replay call

Obviously I thought something was wrong with what they were doing with their project, until I was asked what user I was connected to the database as.  Switching from one schema owner to the other, even though the original had access to read the data (and I thought write as well), I ran the same command then issued my commit.  No problem.  Interesting!  The ORA-code website suggests that maybe I lost my connection while trying to do the commit, but that doesn't make a whole lot of sense even though that could be the final answer.

Sunday, April 6, 2014

Getting proactive!

Ever wonder how to get proactive with regards to High Availability, Backup and Recovery, RMAN, or Corruption?  Look no further than My Oracle Support note 1626745.1 to prevent, resolve and upgrade on these topics!

Saturday, April 5, 2014

In defense of regular alerting - Part 2

Following up on the first installment of this idea, today I bring you yet another reason to put in regular alerting on your system.

How many sessions do you have coming into your system?  Is your EBS system a spider in the middle of an ever increasingly complex web?  If you are not sure about what is coming into your system, can you swear that the system is working as intended?

In my newest alert to monitor the sessions coming in to our EBS from other servers in our ecosystem, a pleasant side effect is that we could see what server/user combinations had 0 connections.  What does that mean for us?  Well, in this instance we could actually tell when the other server had been bounced and the services did not come back up automatically.  This means we can get to another level of proactivity by seeing an issue before the business reports it AND help out our IT partners by alerting them to this condition if they were not already aware of it.

Friday, April 4, 2014

R12: Create Accounting and scaling options

Coming fresh off the presses, yesterday we had a problem with our Create Accounting program where it gave us the following error:

****Warning!!! Due to high volume of data, got out of memory exception...***
****Please retry with scalable option or modify the Data template to run in scalable mode...***

Obviously this is not something we wanted (or expected) to see so we compared the run parameters to what was run before, and it did not appear to be anything out of the ordinary so off to My Oracle Support I went.  I found note 1470317.1 titled "PAXACCPB PRC: Create Accounting Completes With Error: Due to high volume of data, got out of memory exception...Please retry with scalable option or modify the Data template to run in scalable mode" which had steps for resolving our issue but the interesting part of this was we had already followed the note EXCEPT for a critical part of it where we made the scalable flag default to N instead of Y.  Once we made that change, after regression testing in QA first, the business re-ran the report for the same org that had the same volume of data, and the report completed successfully this time!

Thursday, April 3, 2014

Weekend Learning: Exadata Energy Storage Module approaching end-of-life

I talk quite a bit about being on Exadata, but I rarely get to deal directly the hardware or architecture so the article ""Energy Storage Module Is Approaching End-of-life" error message via ASR or remote monitoring" found in My Oracle Support note 1583496.1 led me to learn something about Exadata.  The Energy Storage Module associated with the flash cards is actually designed to be replaced after several years, depending on which type of Exadata machine you have, but this note shares the fact that it can be misreported as needing replacement.  If you don't know how to check this, visit the note and try it out for yourself!