Oh … I’d say this was the question I’ve got asked the second most often in the past months (after a potential extension for Oracle 12.2.0.1 support): When will Oracle certify OL8 and RHEL8? And I read this question internally even more often. Now thanks to my team mate Daniel Overby Hansen who spotted it on Twitter on the weekend, we can tell you that … finally … drum roll … Oracle Database 19c is certified on OL8 and RHEL8.

Photo by Derek Oyen on Unsplash
How do you get the information?
Go to MyOracle Support and hit the “Certification” tab.
Then choose “Oracle Database“, “19.0.0.0.0” and either “Linux x86-64 Oracle Linux 8.x” or “Linux x86-64 Red Hat Enterprise Linux 8“.
You may also recognize that no other database version has been certified with OL8 or RHEL8.
What’s the fine print?
I know that we all tend to NOT read fine prints. But you should in this case. You will need to have database RU 19.7 as minimum database patch level. Or you will need to add extra patches on top of 19.6. But life will be much easier with 19.7.0, believe me.
Oracle Linux 8
These are the minimum requirements OL8:
Minimum RU: 19.7 (or 19.6 with patches, refer to aforementioned certification notes)
Minimum kernel versions:
-
- Oracle Linux 8.1 with the Unbreakable Enterprise Kernel 6: 5.4.17-2011.0.7.el8uek.x86_64 or later,
or - Oracle Linux 8.0 with the Red Hat Compatible kernel: 4.18.0-80.el8.x86_64 or later
- Oracle Linux 8.1 with the Unbreakable Enterprise Kernel 6: 5.4.17-2011.0.7.el8uek.x86_64 or later,
ACFS: Please refer to ACFS Supported on OS Platforms for Oracle ACFS certification details (MOS Note: 1369107.1)
Red Hat Enterprise Linux 8
For RHEL8, these are the minimum requirements:
Minimum RU: 19.7 (or 19.6 with patches, refer to aforementioned certification notes)
Minimum kernel versions:
-
- Minimum kernel version: 4.18.0-80.el8.x86_64 or later [corrected on Nov 9, 2020 as the information on MOS is now different]
- Minimum kernel version: 4.18.0-80.el8.x86_64 or later [corrected on Nov 9, 2020 as the information on MOS is now different]
ACFS: Please refer to ACFS Supported on OS Platforms for Oracle ACFS certification details (MOS Note: 1369107.1)
Requirements
Thanks to Chris for opening the SR to ask for more information regarding patch requirements etc.
This is the document you will need to look at:
Further Links and Information
- Certification OL8 for Oracle Database 19c
- Certification RHEL8 for Oracle Database 19c
- Patching all my environments with the April 2020 Bundles (includes 19.7.0)
- Limited Error Correction Support for Oracle Database 12.2.0.1
–Mike
Hello Mike,
Thank you very much for sharing this information.
I have 2 questions related to this certification :
1) Is it foreseen that Oracle 12.2.0.1 is also going to be certified against OL8 and RHEL8 or is Oracle 19.7 (or 19.6) going to be the minimum version for all Oracle DB’s on OL8 and RHEL8 ?
2) Would it be possible to mention the needed patches for Oracle 19.6 as I have not been able to identify the needed patches for 19.6 on Metalink while trying to find them with the following certification notes :
– Database Certification Information (Doc ID 1306539.1).
– Certification Information for Oracle Database on Linux x86-64 (Doc ID 1304727.1)
Thank you very much in advance !
Greetings,
Chris
Hi Chris,
I doubt that 12.2 will be certified on OL8 as well. But to be 100%, I’m not involved in such things.
And please ask Support for needed patches. I know, my answer sounds odd and boring, but I guess there may be either a note already, or somebody may be triggered to publish the information.
Thanks,
Mike
Hello Mike,
Thank you for the feedback. Regarding the additional patches to be installed I created Service Request “SR 3-23091076241 : Additional patches required for Oracle 19.6.x on Red Hat Enterprise Linux 8”. Unfortunately one week later I still do not know which additional patches are required for Oracle 19.6.x :-(.
Greetings,
Chris
Hello Mike,
I got an update on my Service Request saying that the information about the additional patches needed for 19.6 on OL8/RHEL8 can be found in the document “Requirements for Installing Oracle Database 19c on OL8 or RHEL8 64-bit (x86-64) (Doc ID 2668780.1)”.
While having a look at that document then I conclude this is the list of additional (One-Off) patches :
– 30484981 – OJVM RELEASE UPDATE 19.6.0.0.0
– 30150710 – EM express OneOff patch 19.6.0.0.0
– 30489227 – OCWRU Patch 19.6.0.0.0
Remark : For the OCWRU patch it is mentioned that users may choose to update the OCW version
Greetings,
Chris
Thanks a lot, Chris – I’ll add this to the blog post.
Cheers,
Mike
Hi Mike,
I noticed the support note 13044727.1 isn’t update about Linux 8 with Database 19c.
Could you confirm this please ?
Thanks.
Hi Eric,
I guess this will happen soon – and actually I’m not surprised – but if you’d like to push the guys, please log an SR.
Otherwise it typically takes a bit …
Cheers,
Mike
Hi Mike,
the ACFS fine print (MOS Note: 1369107.1) means that there is no ASM certification for now, or do I miss something?
Hi Ben,
no, this is not the case. i just assume that the note hasn’t been updated yet for OL8/RHEL8.
You may open an SR if you want to push an update – or otherwise it may take a bit.
Thanks and kind regards,
Mike
Hi Mike,
thanks for your replay.
I think there are some pakages missing, or I miss more Informations then I expected.
I can’t find the asm kernel modul to install oracleasm and the there are no asm rpms (oracleasmlib / oracleasm-support ) on the OEL 8 Repos.
Best regards,
Ben
Hi Ben,
does this answer your question:
https://twitter.com/AnuragMishra50/status/1260302194949021696
It has the link mentioned to download the packages from OTN.
Mike
Hi Mike,
thanks for your replay again. 🙂
Install OL7 RPMs on an OL8 release. Is this Oracle’s preferred way? I hope not, and if we do, will we still get Oracle support?
Best regards,
Ben
Ben,
this note has been published/updated.
“Requirements for Installing Oracle Database 19c on OL8 or RHEL8 64-bit (x86-64) (Doc ID 2668780.1)”.
Thanks,
Mike
Are there special binaries to download ? I cant see to get past ‘supportedOSCheck’ failure on OL8, with 19.7.0?
The preinstall package should do this:
https://yum.oracle.com/repo/OracleLinux/OL8/baseos/latest/x86_64/getPackage/oracle-database-preinstall-19c-1.0-1.el8.x86_64.rpm
And this is the secret (or otherwise you need to invoke the installer with ./runInstaller -ignoreSysPreReqs
export CV_ASSUME_DISTID=OEL7.6
Cheers,
Mike
Note that the runInstaller program fails on the OS check for on-premise 19.3 database (have to install before you can patch to 19.7). You can avoid this by using something like the following before running runInstaller:
export CV_ASSUME_DISTID=OEL7.6
Hi Mike,
given it is now certified for RHEL8/EL8 (including hopefully ASMFD / ACFS required for Clusterware / RAC), shouldn’t it also be supported for RHEL/EL versions 7,6, 7,7, 7.8 ? I have an SR on this, because most of the MOS Docs only talks about kernels for up to 7.6 ?
We need to get away from the [currently running] 7.5 asap, and to a highest possible 7.x version due to EUS support issues.
Thanks!
Hi Zarko,
all OS patch levels are ALWAYS supported. This is true since the stone ages. Oracle never certifies extra for an OS patch level after the initial base OS version has been given green light. It could be, like in this case here, that we require a minimum OS patch level (such as SLES 12 SP3+) or now, OL8.1+ (whereas RHEL 8.0 is fine). Hence, when OL7 is certified (in this case it was 7.4+ for OL, but 7.5+ for RHEL) all subsequent patch sets for the OS are automatically included.
Not sure what supported needs to find out there 🙁 Sorry for this inconvenience …
Cheers,
Mike
Hi Mike,
and thanks for the very quick reply! What you say is comforting, however… I was never 100% sure about it and there is confusion (to say the least). Having a minimum OS kernel/patch required level for a given Oracle GI / RDBMS release is logical and well understood.
However, there seem to be issues when going to more recent, or should I say, most recent OS versions. I understand this may primarily be due to ACFS / ASMFD support which require kernel modues. I had an SR in the past (will not post numbers publicly here), where with Jan 2020 RU, we requested a confirmation whether we can upgrade the RHEL OS from 7.5 to a newer version, in that case it was 7.7. I was told the kernel was not supported (at that time), and the maximum we could do is to go from 7.5 to 7.6. Now, the current issue is similar, but taking into account the Apr 2020 RU. If we could go from 7.5 to 7.8, that would be ideal, but I really need a firm (certified) confirmation. The Doc 1369107.1 does not (yet) reflect any of this.
I can just hope the guys behind the SR will do it…
Thanks again, kind regards,
-Zarko
Hi Zarko,
I wouldn’t say that every answer from a support person you’ll receive is always true.
Thanks,
Mike
Hi Mike,
Installing our gold image of 19.7 on RHEL 8 gets the following error during runInstaller:
[WARNING] [INS-13001] Oracle Database is not supported on this operating system. Installer will not perform prerequisite checks on the system.
CAUSE: This operating system may not have been in the certified list at the time of the release of this software.
ACTION: Refer to My Oracle Support portal for the latest certification information for this operating system. Proceed with the installation if the operating system has been certified after the release of this software.
[FATAL] [INS-30060] Check for group existence failed.
CAUSE: Unexpected error occurred while trying to check for group existence.
ACTION: Refer to the logs or contact Oracle Support Services. Note for advanced users: Launch the installer by passing the following flag ‘-ignoreInternalDriverError’.
Setting -ignoreInternalDriverError flag allows the install to run without using CV_ASSUME_DISTID but this is not documented anywhere.
All documentation is around applying 19.7 RU to an earlier version so maybe need to get some comments added to consider fresh 19.7 installs that use gold images?
Will this -ignoreInternalDriverError flag be required for 19.8 RU?
Thanks,
Chris
Hi Chris,
thanks for the information – I haven’t seen or used this flag before.
You may need to check with Support whether this will be required for 19.8 as well – I don’t think so but I can’t tell you.
Cheers,
Mike
Having the same issue with our 19.16 gold image…
So the database version 19c is certified on RHEL 8, what about Grid Infrastructure ?
It seems oracleasm from Red Hat is not available in an el8 version, and ASMFD is not certified by Oracle on RHEL8, so how does one run ASM ? Oracle RAC _is_ certified on RHEL according to the Certification pages on MOS.
https://mikedietrichde.com/2020/05/14/of-course-oracle-clusterware-is-certified-on-ol8-rhel8-as-well/
Cheers,
Mike
Thanks Mike, so my followup question is, if ASMFD is not yet supported on RHEL8, and I must therefore use asmlib, is it supported to use the el7 versions of oracleasmlib, oracleasm-support and kmod-oracle-asm ? These must have been used by Oracle to certify Clusterware and RAC so I guess it must be OK ?
I guess so as well – and I think people reported this either in the comments section or on twitter (not sure anymore).
But to be on the very very safe side, you may need to open an SR and check back with Oracle Support.
Cheers,
Mike
Hi Mike,
I am setting up a cluster 2 nodes with RHEL 8.2 and Oracle 19.7 . The installation went well , but when i created a rac database , a got troubles with instance eviction even the creation went well too. But few minutes later one of the instances is evinced and message lake ORA-27508: IPC error sending a message appears and ora-29740 evicted by instance number 2. Did have meet this kind of behavior ? any help or recommendations will be appreciated. The cluster ware work perfectly.
Thanks . Best regards.
Ali Langar
Ali,
please understand that I can’t solve such questions on the blog. You may need to open an SR please.
Thanks,
Mike
Thanks Mike, i already opened an SR.
Ali.
Mike,
I already opened an SR , it was just to find out if you experienced this kind of behavior.
Thanks
Ali
Thanks!
Ali,
You say you are running RHEL 8.2 – you can’t be using ASMFD as this is not yet certified,so I guess you mus be using the RHEL ASM RPMs ? Where did you get those from ? I didn’t know those components were yet ready for RHEL 8.
Andy,
I don’t need ASMFD firstly and I’m not using asmlib , just raw devices.
Ali
Hi Ali,
aren’t raw devices unsupported? Or am I wrong here?
https://docs.oracle.com/cd/E11882_01/install.112/e47805/storage.htm#CWSOL275
Cheers,
Mike
Hi Mike,
I ‘am not sure that raw devices are unsupported, but I’m still using them for performances purpose (We are using VMware who supports Raw Device Mapping (RDM) ) . RDM allows a virtual machine to directly access a volume on the physical storage subsystem, and can be used only with Fibre Channel or iSCSI. RDM can be thought of as providing a symbolic link from a VMFS volume to a raw volume. The mapping makes volumes appear as files in a VMFS volume. The mapping file, not the raw volume, is referenced in the virtual machine configuration .
Best regards.
Ali.
Hi Ali,
that’s why I copy/pasted the link to the docs.
Cheers,
Mike
Hi Mike,
I’m sorry ,for clarification I am using raw devices through oracle ASM instance. Besides, I’ve been doing this since the 10g version.
Best Regards.
Ali.
Hi Mike,
Finally my problème was resolved. My trouble was caused by the Strict Reverse Path filtering in Redhat 8.The reason behind strict mode is to prevent IP spoofing from Distributed Denial-of-service (DDos) attacks. However, having strict mode enabled on the private interconnect of an Oracle RAC database cluster may cause disruption of interconnect communication. So two things must be done :
1- Adding in my case my interconnect devices ( ens1f1 , ens2f1) in /etc/sysctl.conf
net.ipv4.conf.ens1f1.rp_filter = 2
net.ipv4.conf.ens2f1.rp_filter = 2
sysctl -p /etc/sysctl.conf
2- Creating /etc/sysctl.d/98-oracle.conf then adding same parameters
net.ipv4.conf.ens1f1.rp_filter = 2
net.ipv4.conf.ens2f1.rp_filter = 2
sysctl -p /etc/sysctl.d/98-oracle.conf
Obviously this et done on all nodes of the cluster. After that every thing works perfectly.
The value 2 for rp_filter means that Loose mode as defined in RFC3704 Loose Reverse Path Each incoming packet’s source address is also tested against the FIB and if the source address is not reachable via any interface .
Thanks!
Ali.
Thanks Ali!
Cheers,
Mike
I’ve just installed this platform, upgraded to 19.8, with the following patches:
p6880880_190000_Linux-x86-64.zip (OPatch)
p31281355_190000_Linux-x86-64.zip (db patch)
p31219897_190000_Linux-x86-64.zip (jre patch)
The instructions aren’t clear about getting all of this on.
What worked for me was to unpack 19.3, move the bundled OPatch out of the way and unzip the new one, then call runInstaller with the unzipped db patch:
export CV_ASSUME_DISTID=OL7
./runInstaller -applyRU ~/hold/db/31281355
Immediately exit the installer when it starts, then:
./runInstaller -applyRU ~/hold/jre/31219897
My question is the post-patch processing on these sets – do the post-patch instructions (if any) for each patch need to be run on any new or upgraded databases?
I have an SR open on a related subject, so I’m going to ask there also.
Hi Charlie,
no – upgrade automatically executes datapatch. DBCA does the same when you create a new database. Hence, no additional “post” steps need to be done when you upgrade your database to this home.
Thanks,
Mike
Hi Mike,
Just got to test installing a Gold Image I created on RHEL 7 and patched to 19.7, onto the latest patch of RHEL 8.
Two issues I hit:
1. It states there is a missing package compat-libcap1-1.10[CHECK_PACKAGE_EXISTENCE_compat-libcap1]:TASK_SUMMARY:FAILED:IGNORABLE:VERIFICATION_FAILED. I don’t see this listed in the known errors in 2668780.1.
2. I verified that the 19.7 bug patch is included in the $ORACLE_HOME:
$ORACLE_HOME/OPatch/opatch lsinventory|grep 30869156
Patch 30869156 : applied on Sun Apr 26 13:06:06 PDT 2020
Patch description: “Database Release Update : 19.7.0.0.200414 (30869156)”
However, I still had to set the following env variables “export CV_ASSUME_DISTID=OL7” to get past the [WARNING] [INS-08101] Unexpected error while executing the action at state:’supportedOSCheck’, which was supposed to fixed in the above patch per this MOS note.
So two questions:
1. Is it acceptable to take a Gold Image created on RHEL 7 and patched to 19.7 and deploy on RHEL 8.
2. Have you or anyone else hit the two issues I pointed out above?
Thanks!
Hi Brian,
you please need to check with Oracle Support. I’m not an expert for this topic.
Cheers,
Mike
Hi Mike, I would appreciate your take on RHEL7 versus RHEL8 for a new Data Guard standby host for a production X7-2-HA ODA. At this point we intend to upgrade the ODA from OEL6 to OEL7 (31010832) to allow us to upgrade to 19c and better use the remaining 3 years of service life of the ODA. And we are just about to replace our RHEL6 DG standby server, so would we be best to go with RHEL7, to more closely align with the OS version on the ODA? Is there any reason to avoid RHEL8 that you can think of, for this scenario? Any advice would be appreciated, thanks Mike.
Hi Randall,
if it would be only “me”, I would go with OL7. I just receive a number of questions from customers regarding OL8.
OL7 for me does the job as it is long enough supported, and takes all releases from 11.2 to 19c.
Cheers,
Mike
Hi Mike …. agreed, I was leaning that way myself. I appreciate the response!
Having gone through a couple of SRs to get GI and RDBMS running on RHEL8, we have decided to stick with RHEL7 for now. Due to the requirements of ACFS you end up being restricted to a certain kernel version, make a fresh build of the os beyond that version and GI will not install. I’m sure this will get easier over time but for now RHEL7 is the better option.
Hi Mike,
Do you have an update regarding the installation of oracle19c on RHEL 8. Is this now a good option?
Kr
Thierry
Hi Thierry,
I think the situation hasn’t changes so much. Please check the comments sections in both blog posts – the feedback so far isn’t enthusiastic I’d say ..
Cheers,
Mike
Hi,
I have yesterday installed Oracle 19c (with 19.11 RU) on RHEL 8.2 VMs. Installation went well.
(As Oracle ASM Kernel Module driver is not available from RH for RHEL 8.2 I used Udev rules instead of ASM Lib)
Regards
Thiru
Hi,
I meant Oracle 19c Grid Infrastructure (with 19.11 RU) on RHEL 8.2.
Regards,
Thiru
Hello Mike,
Currently I am facing issues with Oracle 19c(19.11) on OEL 8.x on Azure on ASM . We are facing REDO log corruption issue ,however when REDO logs are placed on File system no such issue is been faced. Is this kind of configuration been setup successfully anywhere or same issue been faced ?
Hi Kashyap,
never seen or heard this before.
Can you try please in your spfile:
_use_single_log_writer=true
and in all cases, log an SR for it?
Cheers,
Mike
Hallo Mike,
still no luck when I try to deploy 19.15 GoldenImage created on RHEL7 with EM Fleet Maintenance.
No idea how to set this CV_ASSUME_DISTID= for the Provisioning Job, also this -ignoreInternalDriverError flag is not possible to add this to an existing provisioning profile, however it is not documented, so I can’t sell this to my customers anyways. So we’ll need to create a new GoldenImage with the documented workaround, manually changing the cvu_config file, which is also not really handy to be forced to repeat this every three months.
One could expect, that two years after certification, it should be possible to install the software as is, without working around “bugs”, at least I do; or is the canon, to just refer to the docs and state “but it is documented in the ‘known bugs section'” is eough? IMO this doesn’t free the developers from fixing bugs?
Other than this, keep up the good work you’re doing here with this blog,
Gerrit
Hi Gerrit,
sorry, I’m not really familiar with the EM feature. Only an SR may help you I guess 🙁
Thanks
Mike
Hi Mike — do you know what the plans might be for certification of 19c on RHEL/OL 9?
Hi Steve,
there are plans, and I think it will happen someday – but I am unfortunately not aware of any timeline or commitments yet.
Cheers,
Mike