I guess the frustration which led to the idea that perhaps vendor techies should spend time at users doing the job of storage management can actually be traced back to frustration with the storage management tools which are currently available. At one point I thought it was just me but it turns out that it's not just my incompetence but obviously the general incompetence of the whole Enterprise Storage userbase.
Almost the most entertaining session at any of the user councils that I've sat on is the discussion of management tools. The vitriol that is aimed at the tools year on year is staggering and yet they just don't seem to get any better. I think that the product managers for Storage Management Tools are actually masochists, they enjoy the beating.
Last year we had a consultant on site trying to get our various ECC environments working; by the end of six months; we'd given up and he'd had enough. The very competent and likeable consultant was broken and needed to be taken back into the mothership but was worse is that our ECC environments are still borked. They work a little better and sometimes a little longer before they break and now it mostly emails out alerts; sometimes it appears to forget tho', normally when it is really important. So we don't trust it any more; now I keep around for a couple of things, StorageScope is quite useful for me and we have some processes which would be incredibly painful to change.
EMC tell us that we need to upgrade ECC to the latest version and indeed we do so that it understands and can 'manage' our newest arrays but the conversation I've had with my peers in the industry do not lead me to believe that things will get any better.
I've sat through Powerpoint after Powerpoint and ECC looks quite good as slideware and as long as you don't want to manage more than say two arrays; I suspect it is really good.
I think EMC should close their ECC development teams for six months and scatter the team members to the winds; let them sit in some big enterprise environments and try to keep ECC healthy, working and useful. After about six months, we'll return them broken but enlightened.
But hey, I know you won't do that…because it's all fixed in the next release and anyway, wouldn't want all that nice PS trying to get ECC fixed to dry up.
And IBM in the back; don't sit there so comfortably, you've taken EMC's lead on Storage Management and run with it to new extremes of awfulness. HP and HDS, perhaps a walk of shame for you guys as well.
At times, the whole fiasco reminds me of CA Unicenter TNG; the systems management tool which required more capacity than the environment it was supposed to manage.
I felt such a great affinity with this post! We should give storage management the modern day equivalence of medieval torture…
ECC is a super-slow Java pile of crap. If I could rip it out of our environment, I would gladly do just that. SANscreen does such a better job.
And that is why EMC must buy Sun…to make Java run faster and hence fix ECC!!
I attended three of the ECC user group meetings. I thought it was just me. However, just about every user there who I spoke with on the side had the same pains. Subsequent releases are better — but — the product does not live up to expectations at all.
EMC needs to delve into the architecture and fix what doesn’t work before adding features. I believe just about every time I needed to manage my Symm I was forced to delete restart and re-install the agent. For quick tasks this was still sometimes faster than symcli.
Too much to say to post here! Check out http://storagearchitect.blogspot.com/2008/12/srm-conundrum.html
Totally agree! A number of major vendors appear to treat essential mngt tools as profit streams and this is just unacceptable.
That said of course a ‘fair price’ (included as part of the ‘per GB’ 5 yr TCO of the storage) is acceptable if (and it’s a big ‘if’) the product works…
The promises made over ECC in the last 8 years are almost a comedy act when looked back upon – I realise punishing the present for sins of the past can be unfair, but the context is key in this area.
Like most things future business will go to those that realise “fit for purpose” and “path of least resistance” and the increasingly key deciding factors.
Onaro made (and is making) massive progress and inroads into the storage mngt business by having the right customer facing culture, attitude and delivering on customer requests – when did you last even see the ECC RFE list & update, let alone get an RFE resolved in under 3yrs??
ECC is a dead dog and frankly not worth kicking any more – I continue to ask myself “if ECC was truely free, would I still use/deploy it?” sadly the answer is increasingly “no, it’s easier to manage our kit in other ways”. Another related questions is “Do I have time to wait for the penny to drop @ EMC, and have them deliver something sensible (including stop playing with pointless toys and aspirations in DC mngt way out of their depth)?” – again I’m increasingly of the view that the answer is “no”…
What version of ECC is StorageBod using btw.
5.x; we should be at 6.x but unfortunately we broke the consultant who was going to get us to 5.x by first fixing all the problems we had with 5.x. Suffice to say, it’s still borked.
So in the new year, we’ll go to 6.x by….re-implementing the whole lot.
The only reason to go to 6.x is to get support for the DMX4s we have. And none of my peers in the industry are saying that 6.x is major revelation. Quite the opposite to be honest..
Hmmm. I managed 3PB’s of EMC storage through ECC across 3 different DataCenters, leveraging proxies with very little issue at all. We are a very large OutSourcer with over 7PB’s of managed storage. Storage Resource Management is not an EMC challenge but a challenge among the major storage vendors. I have deployed HPSE, what was APPIQ, Onaro/SanScreen, TPC, and HSSM. I encourage anyone to go through this process and say these competitors are rock solid. In fact you will find HPSE is very challenging to install and the reporting is very basic. Their performance metrics are not pulled from the back-end but rather statistics from the switch ports. Onaro, has an interesting approach (agentless) and I like the option of acquisition agents, which makes the pain of opening ports on the FW a pain of the past. Utilizing port 443 for communication makes deploying/communication easy. BUT… This tool has a LONG way to go. There are some cool features but just like anything AGENTLESS, you are limited with what you can capture. There were plenty of discrepancies we encountered and made some recommendations to Onaro, now owned by NetApp. I did like the risk analysis or what if option. This product has some promise but it is no where near “Enterprise”. It didn’t meet our needs but I expect them to get better. HSSM is no different than HPSE just re branded. Don’t get me started on TPC.. Updating agents (CIMOM), database, etc is a nightmare. The interface is nothing spectacular and not very intuitive. The best part about TPC is their free application called TPC Reporter. This by far across any vendor pulls together the all-in-one report. None of the vendors including EMC offer a comprehensive report like the output from TPC Reporter. This alone doesn’t make this tool awesome but by far wins the customize reporting prize. I think by far EMC has come the furthest with the SRM product including integration with IBM, HDS, HP, NetApp, and VMware.
I am not bias to EMC but I have deployed the other products and before you jump down the path that ECC is junk status, you need to educate yourself on the other products. All have potential but all are faced with the same challenges. Not every vendor wants to play in the sandbox nicely. I think SRM has come a long way and given the products on the market, I personally feel EMC has a better approach. Just my opinion…
Hi Martin,
I think this is a topic you should discuss in more detail. I am a member of a large SAN team, and everyone agreed with this post. The number of comments backs up this view.
We don’t care if an array has 1578 new features, if we cannot manage the thing, find out where storage is assigned to, what can be reclaimed, and what is it being used for then screw it.
Your blog seems to be read by some of the important people in the storage industry, fellows in ivory towers who think that having a bitchfest about their competitors flash drive implementations is the be all and end all. Meanwhile, the chaps who manage their products are going nuts trying to even assign a device to a host without having to patch agents, setup discovery policies that then cause performance issues. Eventually its just better to hit the command line.
SANScreen is by far the best product on the market for monitoring and planning changes, but it is not perfect. ECC is a disaster. We are letting our maintenance run out and will not renew. Patching hundreds of host with multiple agents (will one efficiently programmed master agent not provide all the functionality we need?), a slow interface, a RDF collection policy that can either be accurate but cause performance issues, or hopelessly out of date and not causing excessive syscalls.
And HDS can take a big walk of shame as well.
Frustrated!
I’m planning to do a follow-up post on this at some point in the near future. Actually, I had a very positive response from EMC and they really do want to learn now; it’s acknowledged that ECC is a beast and needs some serious rework; when I say rework, I mean the long promised and never delivered rewrite.
BTW, for day-to-day work, the SMC GUI front-end is a good tool…