Register a SA Forums Account here!
JOINING THE SA FORUMS WILL REMOVE THIS BIG AD, THE ANNOYING UNDERLINED ADS, AND STUPID INTERSTITIAL ADS!!!

You can: log in, read the tech support FAQ, or request your lost password. This dumb message (and those ads) will appear on every screen until you register! Get rid of this crap by registering your own SA Forums Account and joining roughly 150,000 Goons, for the one-time price of $9.95! We charge money because it costs us money per month for bills, and since we don't believe in showing ads to our users, we try to make the money back through forum registrations.
 
  • Post
  • Reply
Kaddish
Feb 7, 2002



Just did an SVC upgrade including node swap to new hardware this weekend. Went from ----wait for it----- 4.3(!!!!!!) to 7.2. Overall it went very smoothly but I'd be lying if I didn't feel a little anxiety at times.

Adbot
ADBOT LOVES YOU

Kaddish
Feb 7, 2002



qutius posted:

My god you'll be so happy with the UI improvements.

I have various v7000s and other SVCs running 6.3+ so I was well aware how poo poo the GUI was. Now that this upgrade is out of the way I get to implement a Flash 820 and update firmware on a DS8100.

Kaddish
Feb 7, 2002



mattisacomputer posted:

Just in time for 7.3 to come out!

Yeah it just came out like a week ago. No chance I'm upgrading to it in the near future though. It adds support for those new "fat" nodes though, they look boss.

Kaddish
Feb 7, 2002



mattisacomputer posted:

What makes you say that? Oh...

Hahaha, sorry dude. Yes, never, ever, ever update to the latest IBM code until it cooks for a few months. Are you using real time compression?

In other news, I just configured my first Flash 820 and presented to the SVC. Next I get to set up a Vdisk mirror and see how this sumbitch performs with Oracle.

Kaddish
Feb 7, 2002



I'm strictly IBM with a few re-branded things here and there (NSeries, etc) and all the continual horror stories regarding EMC make me very glad of this fact.

Kaddish
Feb 7, 2002



You guys complaining about IBM support, have any of you experienced Premium Support with an Account Advocate? It's pretty boss.

Kaddish
Feb 7, 2002



NippleFloss posted:

Can he add contra-rotating cabling to your V7000?

No sir. Are there any SAS systems that have that type of cabling? I've only seen it with fiber.

Kaddish
Feb 7, 2002



Oh. Well then. That's pretty dumb.

Kaddish
Feb 7, 2002



I'm seriously considering consolidating our entire VMware environment (about 45TB) to a Pure FA-420. I can get 60TB usable (assuming 5:1 compression) for about 240k. Anyone have any first hand experience? It seems like a solid product and perfect for VMware.

Kaddish
Feb 7, 2002



NippleFloss posted:

Why would you assume 5:1 compression? You're going to trust their marketing that your data will fit because compression is magic? At least make them demo it and prove it.

That aside, 240k for 60TB (even granting that you get that much) is really high. You could get a similar level of performance for half the cost with a hybrid array. You don't need an AFA for a general purpose VI environment.

I also have some strong suspicions that Pure isn't long for this world, based on some conversations with their people at a conference last week. They're having a lot of trouble breaking into new accounts and moving out of their tier 0 niche because they are really expensive unless you believe in magical (like 12 to 1) compression ratios. If you do then they are merely kind of expensive.

What makes you think Pure is a good fit for your environment? What makes them perfect for VMware?

Because if we don't get 5:1 they will give us more flash to make up the difference. It seems well suited to VMware due to the RTC, dedupe, and free replication. I'm not sold on it but I do like the upfront licensing model. I could get 3 shelves of 900GB V7K plus SVC licensing for around 200k or spend another 40K and get a sexy new flash array.

Kaddish
Feb 7, 2002



Bitch Stewie posted:

Never looked at IBM for our refresh but how does 3 shelves of 900GB run out at $200K? I didn't think the 7000 was that expensive?

SVC licensing.

Kaddish
Feb 7, 2002



adorai posted:

Trip report: We were going to purchase a new netapp 8020 with a few shelves and complete pack for roughly $150k, including five years of maintenance. Instead I purchased a pair of Nimble CS300 series arrays for slightly more, and the performance is astounding. Not being a unified device is a big draw back, but I have some zfs appliances that can step in for cifs. Seriously, if you are buying storage right now, Nimble is pretty drat awesome.

Out of curiosity, do you have a FC infrastructure?

bull3964 posted:

We're demoing a Pure FA-420 right now. We just got our 10g networking up and running last week so we don't have a ton of stuff on it yet, but there's some mix of VMWare and just LUNs carved out to individual machines for various things.

On the two ESX volumes I have allocated, I'm seeing a 5.1:1 and 5.2:1 compression ratio. I have a physical SQL server attached and I'm seeing a 4.9:1 compression ratio for that data.

Most recent thing we added was a Mongodb node, and that's coming out to be 3.0:1 compression.

The compression ratio given does not take into account any deduplication you may be seeing (I think) and also ignores empty space. For example, that MSSQL volume I mentioned has a lot of free space in a few of the DB files. So, the 4.9:1 is only taking into account used space within a file. The empty space in the MDF files doesn't even register. If you compared the size taken on disk in the OS to the amount of storage taken on the array, it would be closer to 12:1.

There's about 200gb of data shared via deduplication right now.

We plan on putting more stuff on it over the next few weeks to run it through the paces. We don't have any 10g host adapters yet, so we've been limited by 2x1g iSCSI right now as far as performance goes.

If you can just drop it right into your environment, it's an easy thing to demo. They'll crate up and take the whole thing back after a 45 day trial if you don't want it.

We're going to be speaking with some existing midsize health systems tomorrow about their Pure implementations. My problem is my storage needs are becoming increasingly difficult to manage. I have an aging DS8100 that now needs more space, volume mirrors from the DS to a flash820 that needs more space, a DS4800 I need to replace soon, a v7000 that's close to maxed out and not under SVC currently. I have a 30 host VMware environment with data stores from both the standalone v7000 and an SVC with v7000 backend connected to a totally different fabric. It's getting really loving hard to manage.

Not to mention three separate n-series, two unified v7000's, DS4500, another DS4800 and SVC/v7000 at our DR site. All of this connected to a fabric of out of support 2005-b5k switches that aren't meshed properly so I have ISL issues constantly, etc etc etc


Ok, I'm done bitching

poo poo, I forgot about the 2 new v5000's ready to be configured for a new Tivoli implementation.

Kaddish fucked around with this message at 03:59 on Oct 29, 2014

Kaddish
Feb 7, 2002



Just found out our only area IBM CE was let go as part of the big "restructuring". IBM is hosed.

Kaddish
Feb 7, 2002



IBM is keeping Storwize everything as far as I know. Unless you're referring to the x86 parts since they're basically xSeries but IBM will still be supporting that.

Kaddish
Feb 7, 2002



We will begin looking to replace our DS8100 soon-ish. We (probably) won't be going with IBM.

What's infuriating is the worthless Client Executive in this area is still there. These guys make like $120k on average and in this area that's a lot of loving money. I know for a fact the CE guys get paid dick for the crazy hours they have to pull. The Client Executive provides the customer with no value and is mainly just a thorn in our side. Seriously, this guys main existence, as far as I can see, is sitting in the middle and taking a piece of the customer pie.

Kaddish fucked around with this message at 20:32 on Jan 30, 2015

Kaddish
Feb 7, 2002



We're going through a switch upgrade. We currently have old IBM 2005-B5K's (rebranded Brocade) and we're going upgrade to Directors. Our hardware rep has scheduled a call with Brocade and Hitachi - evidently Hitachi also sells rebranded Brocades and we can save some money. Any use Hitachi Directors? How is support?

Kaddish
Feb 7, 2002



Zephirus posted:

They're not even re-branded afaik. They're just brocades - at least that is the case for everything brocade i've bought through HDS. Support is OK, but with HDS YMMV depending on your area. APAC is not great, UK/EMEA is ok, US/CAN i've had good/bad experiences but not enough to judge.

You get access to everything brocade software wise, through the HDS portal - there is a partner link to brocade.com, so you don't lose out there.

Awesome, I didn't realize they weren't even re-branded. I popped over to the brocade website and that does seem to be the case. So support will be my primary concern still. Thanks for the info.

Kaddish
Feb 7, 2002



So, I've had about a week to play with a Pure FA-420. I'm using it for VMware. It took me a bit to wrap my head around the relationship between thin provisioning to the hosts and the de-dupe/compression on the backend. Let me just say this thing is pure magic awesome. VMware think it has 24TB in 6 datastores. Of that 24TB it says its actively using about 10TB. On the backend it's actually using 3TB on disk. Two thumbs up from me!

Kaddish
Feb 7, 2002



I'm not a big Netapp guy but isn't BackupExec NDMP capable?

Kaddish
Feb 7, 2002



TSM for VM seems to work well but that's ruled right out due to both of your requirements.

Kaddish
Feb 7, 2002



Anyone have a chance to play with an IBM Flash 900 yet? I'm replacing a DS8100 later this year and have been weighing some options. With pretty much any system these days reliability is pretty great. I'm not seeing the point in something like DS8870 considering the cost. Right now I have our main EMR Oracle DB mirrored between a Flash 820 and DS8100 via SVC.

Kaddish
Feb 7, 2002



paperchaseguy posted:

I've worked with the Flashsystem 900, some of my customers are getting 900s and v9000s.

The 840 is a good bit nicer than the 820 all around, modular, nicer GUI, easier upgrades. The 900 is not drastically different than the 840, primarily it has more capacity.

Can you virtualize storage on a v9000 just like an SVC/v7000? It looks like they literally stuffed an SVC pair in there. If yes this may not be a bad option.....

Kaddish
Feb 7, 2002



IBM regular support sucks but Premium/Account Advocate is great.

Kaddish
Feb 7, 2002



Last I knew they still had some good NSeries/Data Ontap people in Raleigh, NC but they could be laid off now for all I know.

Kaddish
Feb 7, 2002



Anybody have experience with SolidFire flash arrays? How well does the scale-out and volume QoS features work?

Kaddish
Feb 7, 2002



If you're using block level de-dupe and compression at the storage level, it doesn't matter if your vmdk is thin provisioned or thick. A block is a block and it's either being used on the array or it isn't. At least this is true on Pure, which is the only de-dupe/ compression/thin provisioning I use.

As mentioned above, make sure you utilize SCSI UNMAP periodically, especially if you have aggressive DRS. This doesn't run automatically. You will need to run it against a datastore from any host in the cluster.

Kaddish fucked around with this message at 14:46 on Aug 17, 2015

Kaddish
Feb 7, 2002



If you really want to see actual storage usage from one interface Nimble probably has a plugin for Vcenter. Pure does.

Kaddish
Feb 7, 2002



I worked on a quote last year for a complete pie in the sky scenario to replace all of our IBM v7000's with new storage last year. I went with IBM again because I love SVC. (not that I needed to, I guess, but I do like Storwize)

My FlashSystem 5100 for prod and 5030 for DR were delivered, out of the blue, a couple of weeks ago.

I'm going all DRAID bitches!

Kaddish fucked around with this message at 21:17 on Mar 3, 2021

Kaddish
Feb 7, 2002



Langolas posted:

SRM is pretty lovely

I've seen Vmware sit and point fingers at storage vendors on it when it was a re-scan issue from Vmware taking longer then expected and timeout values weren't tuned right from SRM

If Pure has a plan to address it, I would follow what they say before trusting anything Vmware says


It's because SRM is dependent on storage vendor API usage and both VMware AND Pure are working with barely functioning code.

Edit - I love Pure storage, they are solid dudes with a solid product.

Kaddish fucked around with this message at 21:18 on Mar 3, 2021

Kaddish
Feb 7, 2002



lol internet. posted:

Hmm, I seem too of started to lose a LUN on a cluster that's been up for like a year.

This is the error message. Any idea?

Cluster Shared Volume 'Volume1' has entered a paused state because of 'STATUS_LOGON_FAILURE(c000006d)'. All I/O will temporarily be queued until a path to the volume is reestablished.

Did you resolve this? It sounds like a fabric/hardware layer issue.

If this is FC, and I assume it is, you might want to look at C3 discards on your target/initiator ports. A slow drain device could cause this, which is also the bane of my existence.

Kaddish fucked around with this message at 21:36 on Mar 3, 2021

Kaddish
Feb 7, 2002



Pikehead posted:

More News:

I've kept out of this for a while but got to talking to the people who deal with this. Apparently my organisation uses very descriptive names for the luns at the Pure level and they went over some magical 40 character or so limit. This meant that every time SRM got used a full rescan was required and it took hours.

Options are:

1. Fix it somehow
2. Wait for tags

Tags are now apparently a thing in the latest SRA so once everyone lines everything up and tests it first and then put it into production, SRM will finally work the way it's supposed to.

Um. Why do you have 40+ Pure LUNS is the pertinent question? I can't fathom a reason for this.

Like, you're trying to fix something that seems to be fundamentally broken, and fixing the fundamentals will help with whatever you're trying to do going forward.

Edit - I see, it's the LUN naming convention that is the problem, which gives me a headache even thinking about.

Pure LUNS are just a storage bucket, and unless you need specific compression/dedupe statistics per Pure LUN, they don't mean anything. And even if you need those stats, just create a 'test' LUN for compression/dedupe potential.

Or, there's always vvols (lol)

Kaddish fucked around with this message at 00:59 on Mar 4, 2021

Kaddish
Feb 7, 2002



Pile Of Garbage posted:

I miss working with IBM storage, also doing FC fabrics. Don't get to work on that kinda stuff these days...

Edit: those IBM FC switches/MPRs that were just re-branded Brocade devices, so nice to work with!

Ha, yep I still have some 8GB IBM switches at our DR location.

Edit - The new Storwize code has the option of setting up a storage array, including pools, mdisks, etc for SVC best practices, I tempted to try it. I'll need to recable/rezone my DR SVC cluster first to get NPIV working, though.

Kaddish fucked around with this message at 13:39 on Mar 4, 2021

Kaddish
Feb 7, 2002



ihafarm posted:

After 10+ years running a couple generations of EMC storage(Celerra/VNXg1) I opted to go with Nimble for my most recent refresh. Still migrating over, but have been converting some non-user facing systems to vvol, what gotchas should I be worried about? I already knew that veeam doesnít support direct backup from vvol data stores(at least w Nimble), but thatís not a concern in my environment. Seems like any enterprise storage advancement is met with skepticism, but vvols have been supported since what, 2015?

I actually haven't used vvols yet because it seemed to me the limitations and implementation was ....severely lacking. Vvol support wasn't even added to Storwize until later in the 7.x code stream and converting over seemed like a pain. Maybe the APIs/implementation/etc has gotten better over the years but it seemed like they attempted to solve a problem I didn't need solving. In general, I reduce complexity in the environment as much as possible.

Kaddish fucked around with this message at 16:36 on Mar 5, 2021

Kaddish
Feb 7, 2002



I used the new-to-me Storwize feature to set up a new Flashsystem 5100 -volumes, pools, mapping, etc, for use behind an SVC and it worked surprisingly well. I had no issues with any of it.

Kaddish
Feb 7, 2002



Thanks Ants posted:

Storwize Unified will do NFS, but probably can't tick the 'quickly' box unless someone has demo kit ready to ship.

IBM has pulled marketing for Unified and seems to be exiting that sector of the file business. I have three Unified arrays Iíll be looking to replace soonish.

Kaddish
Feb 7, 2002



Kyndryl is the worst name Iíve ever heard for a spin-off company. Like, way beyond Quikster levels of bad.

Kaddish
Feb 7, 2002



They didnít mention what industry their customer is in. I assume the OP is a managed service provider of some sort. There is a reason we use proven technologies on solid frameworks for critical production environments.

Kaddish
Feb 7, 2002



I feel like storage is still the Wild West of IT and poo poo like proxmox, ceph etc is cool and fun until it isnít.

Kaddish
Feb 7, 2002



devmd01 posted:

There were other reasons, but many years ago I straight up told a company why I refused their job offer without even seeing numbers - they were using a used CX4 without any kind of support contract and were buying parts off of eBay. gently caress. No.

I still support a cx4 at one of our small sites. We have a support contract though, obviously.

Edit - and itís scheduled to be replaced but Iím only one person, poo poo.

Kaddish fucked around with this message at 20:21 on Apr 15, 2021

Adbot
ADBOT LOVES YOU

Kaddish
Feb 7, 2002



Pile Of Garbage posted:



Edit: I was quoting Vulture Culture's post because it raised a good point in that while yeah you can defer liability and responsibility to the vendor depending on the problem you can still end up investing an extreme amount of time in helping them to reach a resolution.

This is definitely true with regards to investing time reaching a resolution but at the end of the day, in the corporate world, placing the responsibility on the vendor to provide a positive outcome is key to long-term job stability. I've identified several issues over the years, the biggest one off the top of my head was a serious SVC metromirror issue that led to data corruption. IBM released a version revision and I had a feather in my cap in the eyes of management.

  • 1
  • 2
  • 3
  • 4
  • 5
  • Post
  • Reply