Oracle Solaris Reviews

Reliability, safety, support offered, speed, endurance. Overall, if I’m forced to make an automotive industry comparison, from my perspective Oracle Solaris is the “Volvo” of operating systems.

Reliability – because referenced operating system was always trustful to host assigned applications, in the most desirable way

Safety – penetration test performed against related platforms never get the chance to pick any vulnerabilities. Attempts were logged on system log, daily check scripts pick those and send to monitoring team, investigations was performed to pinpoint the offending host. All the times the host was the one used to perform penetration testing. Other layers of infrastructure security prevented the attackers to rich the Solaris environments, but never heard about proper configured OS like this to get on its knees due to such.

Support Offered – Oracle SPARC platforms come with free support for Solaris operating system, and in contrast with the saying “there is no free lunch”, the support offered is at the highest standards. In some of the rare occasion when high severity incidents ( with root cause proven to be triggered from application side), the customer on organized conference requested a Service request to be opened with Oracle, so the issue to be cornered from both directions. I can say that have nevered encountered a new or known bug on OS side during those experiences.
Speed – Living the hardware and storage capabilities out of the equation, the responsiveness and agility of Solaris operating system is one in the top of my preferences.

Endurance - Years of uptime (except planned downtime related to recommended patch set applying) say it all.

After all of the above, hope that the comparison from automotive industry will not be considered forced at all.

Using Oracle Solaris OS in conjunction with Oracle Solaris Cluster, we are able to offer our organization full OS support so they can achieve the best possible Oracle database performance, security and failover capabilities. This way, their business doesn’t suffer due to unexpected and unplanned IT infrastructure-related downtime.

Since Solaris Volume Manager is the obvious choice for a shared filesystem, I would love to see improvements in SVM so that filesystems can be increased or migrated without downtime for the environment - in a similar way that ZFS is capable of.

Offering more GUI applications might help adoption of Oracle Solaris by professionals coming from the Windows “world”.

Also, adaptation of behavior of commands similar to those from Linux would add familiarity for Linux system administrators aiming to use Oracle Solaris as well. (E.g., in Linux, the ifconfig command with no parameters returns the output for all interfaces. In Oracle Solaris, it causes an error and usage help).

I have used it for 3.5 years.

Encountered issues were back in the time when not enough experience was available for related OS, and some due to bugs that were fixed by patches released.

I will rate technical support as excellent.

IBM AIX was also considered, but x86 platform support from Oracle Solaris (for laboratory, test, DEV and QaS environments) added an advantage that leaned the scales toward Solaris. Proprietary hardware came with an added list of tasks you need to consider and be aware of during live use.

The initial migration project came with lots of organizational challenges, but from a technical perspective, the well-organized “My Oracle Support” and attached community provided the needed answers to go along with a successful pilot environment. From that point on, we needed only to be patient, trigger and organize interactions between various teams, and solve performance issues related to layers other than the OS.

There was an in-house implementation. My advice is to boldly adapt Oracle Solaris OS if the cost fits the organization budget, because this will save lots of future costs related to fixing issues and malfunctions. During implementation, adopters can count on professional support from the vendor using the “My Oracle Support” tool.

ROI, pricing and licensing are variables in an equation that changes from one case to the other. In my case, I wasn’t involved in the accounting part of this; therefore, I cannot provide more details.

Before adopting Oracle Solaris, extensive pre-testing is needed in order to gain the necessary experience. Ideally, certification on Oracle Solaris for related IT team members is desirable.

In Solaris 11, Network Virtualization, LDOMs, Zones, ZFS, and Live Upgrade are the most valuable features to me. In Solaris 10, the most valuable features are the Live Upgrade, ZFS, Zones, and SVC Feature.

It helps a lot in data center consolidation, P2V, and with LDOM live migration. It reduces required overall downtime, and is highly scalable, especially with T5 architecture.

I have noticed very frequent HBA and NIC card failure in T5-1B or T4-1B blade servers. More stability is required. I have experience multiple instances where QUAD HBA+NIC Port, for T4-1B, T5-1B was failed. And we had some downtime, as replacing HBA card, potentially needs server to be powered-off. Though this was noticed in case of Blade servers only, not for T5-8 or T5-2 or T4-2 etc.

We have been using Solaris 11 for 3-4 years, and Solaris 10 for 10 years.

I didn’t encounter any issues with deployment.

I didn’t encounter any issues with scalability.

I rate technical support at 9.

I have had the opportunity to work with various other UNIX as well as Linux operating systems, but among all of them, I found Solaris to be a very stable operating system. Now with the evolution of LDOMs, ZFS and zones, it’s providing a perfect platform for virtualization solutions.

To me, installation doesn’t seem complex, but at the same time comprehensive understanding of technology is required.

We performed implementation ourselves, except some time when we requested Oracle professional services request.

I recommend Oracle solutions such as T5 Super-Cluster, T5-8 and the Oracle Solaris 11 operating system, as they are highly scalable and fault-tolerant systems.

I rate Solaris 10 a 7.

Built-in virtualisation (Zones / LDOMS), ZFS, SMF, and FMA.

By standardising the infrastructure, we were able to reduce the number of physical hosts. We got rid of file system corruption (thanks to ZFS). Enlarging the file system was as easy as eating the candy.

The product itself is great. I would like to see whether other companies start developing for Solaris.

I have been using Solaris 11 for four years and Sun Cluster 4.1 for three years.

I’ve never had a problem with deployment (well-configured Jumpstart and AI). We have noted almost no hardware failures. Due to Solaris Cluster 4.1’s Live Migration future, we have achieved 99.9999% service uptime.

The level of customer service varies depending on the geo zone, but mostly Oracle support is very good and quick with providing the resolution for problems.

Because we have designed our boxes in a very complex way, the setup took some time and it’s not a straightforward (click, click) setup. It requires skills to create a well-functioning environment. That is why people who have never worked with Solaris or had “one date” with it should never set up any Solaris box.

We implemented the product in-house. I recommend reading the documentation and not assuming that if you know Linux, you know Solaris.

Concurrently we have used HP blades with VMware for our Linux environment. Due to fact that not all of the applications were running on Solaris, it has cost us much more effort to make those three work together. It is more time consuming to deploy VM on that environment than on Solaris.

If you have never done it or you do not have a sufficient amount of experience, hire an external consultant or an Oracle consultant to do the job.

The following are the most valuable features to me:

• Solaris zones: Allows for better server utilization. Drives number of physical servers down. Very lean virtualization technology which I generally prefer to others.
• Live upgrade: In Solaris 10, it allowed us to decrease downtime.
• ZFS: Was and is still the best logical volume manager / file system in my opinion. There is still nothing like it production-ready in the Linux world. Favorite features:
  
  • dedup
  • snapshots
  • checksums and self healing

Dtrace is also pretty useful. However, now Oracle Linux has it and also in RedHat Linux, there is “systemtap” which closely mimics dtrace.

It’s pretty much irrelevant right now as we switched to RedHat several years ago. At least in the telecom solutions I work with.

I would probably like to see it open sourced once again, as was situation with Open Solaris.

Right now, I see less and less organizations using Solaris and, at least from my point of view, there is not much active development around it.

I used this solution for eight years.

Deployment was never an issue.

However, Solaris 10 Jump Start and Solaris 11 AI were somewhat harder to use than the analogue PXE boot + kickstart technology used in RedHat.

Stability was never an issue either.

The last time I had to deal with tech support was probably five years ago.

I find that Solaris had a more difficult learning curve compared to Linux. Partly because there was a much wider Linux community, which is still true today.

I wouldn’t call it complex, but definitely when I started to work with Solaris in 2003, I found a few surprises. At that time, I had Linux administration experience.

We implemented it in-house.

As I've mentioned, it’s hard for me to make any recommendations as I have been working with RedHat for a long time now.

Robust kernel: The heart of an OS, i.e., it’s the base/foundation of any operating system. If we have a robust kernel, the chance of getting server panic, etc., is reduced to almost negligible levels and that’s true with Solaris and even with IBM AIX. The bug levels and vulnerabilities to hit such robust kernels are very low.

Patch/package management: Change is a part of IT, with the increased technologies day by day, new software is evolving every day. If the way to install, manage, upgrade, configure them is not easy, then instead of using the software for growth, techies will be killing their time fixing them. With Solaris, this system is very strong. Regular security fixes, vulnerability fixes, recommended patches for new kernel and for new features is in Oracle’s release management process, which is very beneficial for customers to stay updated and fix old bugs.

Visibility at the OS level (nothing hidden): In case of issues, the logging system for Solaris is outstanding. With logs, we can debug the issues to a higher degree by ourselves. In case any changes to kernels are required, Oracle is always there to fix them via patches/pkgs, which is again valued added for any customer.

Compatibility with third-party DB's and applications: The compatibility to install databases and application on top of Solaris is just amazing, we hardly see any issues during installation/upgrade, except third-party driver issues. So overall, it’s an amazing OS to work with.

It integrates with different applications with complete stability.

I think Oracle should also promote x86 architecture for Solaris, so that the same can be used in ESX and in cloud environments with an x86 variant. It would be a bonus for Oracle.

SPARC hardware is costly. Most businesses want to run their infrastructure environments - especially non-production environments - on x86 hardware, where customers can run heterogeneous OS platforms (Linux, Solaris and/or AIX). However, this is not possible with AIX at all (especially with Solaris). So, if Oracle improved x86 support in Solaris, it could promote Solaris x86 as having the same stability and reliability levels as Solaris SPARC servers. This would provide customers a reason to move their servers - which are currently migrating to Linux servers – to Solaris x86. And I am sure this will boost Solaris even further. I am well aware that Oracle is promoting Oracle Linux for the same reasons, but I think the same can be done in a better way for the existing Solaris OS.

I have been using it since 2007 until now.

We always perform POCs first and try to do all possible testing in that phase. As such, we have not encountered any major deployment issues.

It actually depends on the application version, and compatibility also. Sometimes, after a kernel patch, we might encounter some issues, but that’s just because of poor planning and poor documentation.

For standalone systems, scalability’s always a challenge, but improved T-series and M-series have good options. However, Oracle still lags behind in this area, compared to ESX and IBM (LPARs), to a great extent.

Sun Microsystems customer service was superb!!! (My first love sunsolve.sun.com.) When Oracle initially took over, their customer service was just pathetic, but it slowly improved. It’s always hard to maintain the same standards and I can understand it’s tough in the initial phases. I would say, from my experience, there is still room for improvement in this area.

Technical support is good, as it’s divided into different levels. Sometimes, it takes time before L1 escalates to L2/L3 and that sometimes is frustrating (sometimes :-) ).

Before moving to engineering, I worked as admin / implementation team in a heterogeneous environment. So this solution completely depends upon cost and the client’s requirements.

Migration is always a challenging step, if you want everything to be the same as it is running on an existing environment. AIX to Solaris or Linux to Solaris or vice versa is always a job to be performed with extra precaution, as you are going to play with your data.

We implemented it ourselves and did the hardware replacements via a global vendor. It’s good.

ROI is good. I am sure Solaris has devoted everything to it. For an OS to survive a long time, Oracle has to maintain Solaris like a baby, as Sun Microsystems did until 2010.

Since taken over by Oracle, there has been an increase in the software cost (earlier patch/packages were free with SunSsolve), but that’s business and I think it’s okay.

Today, if you ask me for low-cost solutions, I have open-source Linux versions and for stable releases, we have RedHat and IBM AIX.

It’s a very good product to use. You are going to love this OS.

I still love Solaris; for me, it’s always been the best.

In, Oracle Solaris 11, the most valuable features to me are ZFS, Automated Installer, and Kernel Zones. In Solaris 10, they are ZFS, SMF, Live Upgrade, Zones, and JumpStart.

I would like to see improvements in user adaptability, the poor UI, and in the packaging of the applications. Solaris should look like Linux and the end user should not be afraid of using it. It's way different than linux.

It should have POSIX compatibility with Linux.

We have been used Oracle Solaris 11 for four years, and Solaris 10 for eight years.

There is limited deployment support outside of the documentation on the Oracle site, which is a problem when it comes to configuring, deploying and support for Solaris. Outside of docs.oracle.com, there’s very little knowledge base available, which is a big problem for Solaris.

For Oracle Solaris 11, I rate customer service and technical support a 7/10. For Solaris 10, 6/10.

Solaris 11 had its challenges, being a new OS & again with the limited knowledge outside of the subscription world and its adaptability rate.

Solaris 10 was not too complex to deploy.

If I had a chance, I would implement it on my own, as long as Oracle’s subscription is affordable.

It is NOT affordable compared to Linux. Oracle’s licensing policy is horrible.

Unfortunately, we are moving away from Oracle Solaris & completely into Linux now. We have replaced a whole lot of Solaris with RHEL. That's what’s happening more in IT. Oracle’s poor strategy to lift the Solaris OS and tight licensing policy is killing Solaris.

Easy-to-find self-help in Linux through different sources increases its user adaptability rate & popularity. If Solaris is even going to survive, they need to change the marketing & licensing strategy.

Solaris 11 OS is as good as Linux. In fact, some of their features are way more advanced than Linux. But you need to self-learn, get to the comfort level of using it and push others to do so, especially the platform consumers.

Zones and resource allocation through capping and project is my favourite feature.

Lately, I have been using ZFS and I absolutely love it, but I didn't get much of a chance to explore it fully in a production environment.

This product performed wonderfully with my banking client, where I participated in implementation of virtualization through Solaris zones and then capping CPUs. We integrated zones with VCS clusters. It provided unmatchable stability, high availability, scalability and the best tunable performance.

We used it on M series, X series or the latest T series. It gave great reliable performance on all of the hardware.

I believe it's a great product and its latest versions are also really good. However, I believe Oracle is not utilising its full potential by restricting it best performance with Oracle hardware. Even though it can be run on SPARC, as well as Intel hardware, the problem lies with the way Oracle chooses to promote it. They are always saying that it performs best with Oracle hardware. They should understand current demand for open source and publish white papers for its performance on Intel hardware. And they must change their stratergy with Dell, HP and other blade server manufacturers and enable them to use Solaris and promote Solaris.

Also, they should promote Oracle Solaris with open source tools like HANA, Hadoop, Puppet, Chef, and Ansible. Meanwhile, they can continue to develop and promote their in-house competitive products as well.

To summarise, I feel the main issue lies with their promotion and sales strategies, and also their relations with competing hardware vendors and database/application vendors.

I have used Solaris for more than 8 years, almost all of my career, with all of my clients.

I have done many deployments, migrations and so on with Solaris or to Solaris, and I never faced a problem where I would have received a response from Oracle/Sun support that it was not possible. The product and its features work almost exactly as promised and the documentation available for the product.

Yes, I have seen bugs like zoneadmd hanging, or a zone getting stuck in a shutting down state, but they usually don't happen during deployments or planned activities.

Experience with Sun support was absolutely fantastic, but it deteriorated a little when Oracle support took over.

Most of the environments used Solaris, and we upgraded from Solaris 8/9 to Solaris 10.

In my experience, we mostly moved up from older hardware running Solaris 8 to Solaris 10 on new hardware. Complexities came in the form of an upgraded version of Veritas Cluster and volume manager or storage migration. Solaris itself didn't create any issues at OS level.

We mostly did implementation through in-house teams.

The most important thing is to have a sufficient downtime window and application or database support teams to be available to verify immediately.

I don't have much of an idea about pricing, but it should be decoupled from SPARC architecture.

Even though nowadays, I am using RedHat Linux, in my environment, I miss Solaris a lot.

Trust Solaris. It is still better than Linux in many ways.

Solaris' most valuable features to me are its stability, scalability, dependability, RAS, HA, I know there are loads more TLA’s that can be used, and of course it’s grown into all the new cloud features, also, to be ready for the next generation.

For Oracle, it’s Oracle; is there any better database? I’m biased. It can be used on the smallest device running a simple meta data store to the biggest, hard-hitting, critical system.

I just think the maturity of Solaris, the base core has been proven, and it is evident in these Enterprise level/required features. People don't look at Solaris and ask is it production ready, it is probably one of the first options written down when people need to look at a Unix OS for big critical solutions because of the core features. Other features thats always been there is of course Security also, and now being expanded with all the Cloud ready features.

I’m working less and less with the OS it seems. Where I used to think, "I’d love them to improve this," I’ve heard that's exactly what they’ve done. Even the newest Oracle Mini Cluster only has a visual interface for deployment and management.

The OS always needed and still does need more of a visual interface. Not to take anything away from the command line - I love that - but for basic mass user community acceptance, there is a large Windows, under-30 user base that doesn’t know how to think when they don’t have a mouse to do things with.

I've been using Solaris as an OS to host mostly Oracle primarily since 1998. I first ran it on a little Sun Classic 50. That would have been Oracle 7.3.2. Next machine up was a Sun Sparc E450. I called it the coffee table, as it was right next to my desk. I had 2 of them, one running an Oracle Database, the other Running Dynamo Application Server + Apache Web server. Eventually I lost my coffee tables to the server room when the project went live.

With Solaris and Oracle, never. If you’ve done your homework and you're prepared and know what you want to accomplish and how to get it done, then all goes perfectly, but then that's life: Prepare and things work; don’t and you have a hard time ahead trying to hammer it into shape/direction.

You DO need to know how to work with support, they do have all the information, and the means to figure out a problem, but like any support division, not always the easiest to engage with.

I use Linux a lot, and, well before Oracle, I used to work on Sybase.

I’ve never done anything straightforward. I tend to specialise in the complex deployments, putting them together in such a way that they actually become simple to manage with the minimal of skills.

I use to be with a platinum partner, then I worked for the vendor, and now I'm with a partner again, so I’ve done both sides of the fence. The most important part of any solution is KNOW your problem first, then look at the proposed solution and it’s components and the features, and identify which features are to be used to solve which part of the problem, and try and follow KISS. ;)

Consolidate where it makes sense. At times, simply trying to increase ROI can increase complexity, which pushes up operational complexity and associated costs/risks, which actually hurts cost of ownership and has a silent impact on the ROI, as it might not get adopted as eagerly as hoped.

Again, with any project, know the problem; know your available options. There might be multiple options in different products available from the same vendor. Decide which ones together will work best for you, and it will at times definitely not be the one that is being proposed by sales. Don’t be afraid to push the vendor to the edge. At times, the best options/solution might cost you some red eyes/sleepless nights, but they also long-term put you ahead of the curve of your competitors.