Enterprise Storage Megathread: Why is my NAS a SAN?

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > Enterprise Storage Megathread: Why is my NAS a SAN?

«‹›207 »

YOLOsubmarine: Oct 19, 2004; When asked which Pokemon he evolved into, Kamara pauses.

"Motherfucking, what's that big dragon shit? That orange motherfucker. Charizard."

Wicaeed posted:

Probably more for the DBA thread, but correct me if I'm wrong, don't MSSQL HA Availability Groups use Cluster Shared Volumes for storage? Or can they use attached local storage as well?

The presentation I've been given says we are going to be using our SAN for write-intensive storage in a 2 node availability group, and then using another 2 nodes availability group for the read-intensive operations. Either way we are potentially wasting 50% of our resources since:

A) We have 4 nodes to use, and all 4 nodes have the same storage on them
B) Only two of those nodes (the read-intensive workloads) will be used in an availability group
C) Two nodes will have all of their local storage going to waste

AlwaysOn Availability groups can use local storage (just like Exchange 2010 and 2013) on each node. Shared storage is only required if you're running WFC in front of your SQL instance. This is only for 2012, however, replication in 2008 is handled differently.

You can even do things like have your primary DBs on SAN and secondary on local, or mix and match any way you like. Replication can be configured as synchronous or asynchronous.

# ? Jul 15, 2014 07:44

Adbot: ADBOT LOVES YOU

# ? May 9, 2024 22:17

Nebulis01: Dec 30, 2003; Technical Support Ninny

Wicaeed posted:

Probably more for the DBA thread, but correct me if I'm wrong, don't MSSQL HA Availability Groups use Cluster Shared Volumes for storage? Or can they use attached local storage as well?

The presentation I've been given says we are going to be using our SAN for write-intensive storage in a 2 node availability group, and then using another 2 nodes availability group for the read-intensive operations. Either way we are potentially wasting 50% of our resources since:

A) We have 4 nodes to use, and all 4 nodes have the same storage on them
B) Only two of those nodes (the read-intensive workloads) will be used in an availability group
C) Two nodes will have all of their local storage going to waste

CSVs are only supported in SQL2014 and above, also Availability groups are only an option if you're running Enterprise Edition of MSSQL. Standard only supports using a Failover Clustered Instance based on Windows Server Failover Clustering which requires shared storage.

It seems silly to that you're willing tos pend the $$ licensing enterprise but not drop $22k on the Nimble/Equallogic box.

# ? Jul 15, 2014 21:31

Wicaeed: Feb 8, 2005

Nebulis01 posted:

CSVs are only supported in SQL2014 and above, also Availability groups are only an option if you're running Enterprise Edition of MSSQL. Standard only supports using a Failover Clustered Instance based on Windows Server Failover Clustering which requires shared storage.

It seems silly to that you're willing tos pend the $$ licensing enterprise but not drop $22k on the Nimble/Equallogic box.

Our parent company literally showers us with license keys for anything Windows we want.

How I wish they would give us their Enterprise VMware licenses

# ? Jul 15, 2014 21:36

Vulture Culture: Jul 14, 2003; I was never enjoying it. I only eat it for the nutrients.

Wicaeed posted:

Our parent company literally showers us with license keys for anything Windows we want.

How I wish they would give us their Enterprise VMware licenses

Can you draw a picture of them literally showering you with Windows keys because I'm literally having a hard time visualizing this

# ? Jul 16, 2014 00:05

Maneki Neko: Oct 27, 2000

Misogynist posted:

Can you draw a picture of them literally showering you with Windows keys because I'm literally having a hard time visualizing this

I'm not willing to spend any real time on it, but I'm assuming it's something like this:

# ? Jul 16, 2014 00:15

Wicaeed: Feb 8, 2005

Maneki Neko posted:

I'm not willing to spend any real time on it, but I'm assuming it's something like this:

It's pretty much like this

# ? Jul 16, 2014 00:18

goobernoodles: May 28, 2011; Wayne Leonard Kirby.

Orioles Magician.

# ? Jul 16, 2014 00:26

pixaal: Jan 8, 2004; All ice cream is now for all beings, no matter how many legs.

Maneki Neko posted:

I'm not willing to spend any real time on it, but I'm assuming it's something like this:

Are these actual keys?

# ? Jul 16, 2014 03:03

in a well actually: Jan 26, 2011; dude, you gotta end it on the rhyme

pixaal posted:

Are these actual keys?

Shameful. Forums user FCKGW would like a word with you.

# ? Jul 16, 2014 03:07

Vulture Culture: Jul 14, 2003; I was never enjoying it. I only eat it for the nutrients.

pixaal posted:

Are these actual keys?

http://en.wikipedia.org/wiki/Volume_license_key#Notable_keys

# ? Jul 16, 2014 04:36

Docjowles: Apr 9, 2009

Maneki Neko posted:

I'm not willing to spend any real time on it, but I'm assuming it's something like this:

When Misogynist posted this was 100% exactly what I pictured in my head :golfclap:

# ? Jul 16, 2014 04:37

Nukelear v.2: Jun 25, 2004; My optional title text

Misogynist posted:

In SQL Server 2008, a common topology was something like this:
code:
              ,----------------.          ,----------------.
              | Shared storage |          | Shared storage |
,--------.    `----------------'          `----------------'    ,--------.
| pri-01 |---\        |                            |        /---| mir-01 |
`--------'    \   ,------.                      ,------.   /    `--------'
               |--| MSCS |----SQL Mirroring---->| MSCS |--|
,--------.    /   `------'                      `------'   \    ,--------.
| pri-02 |---/                                              \---| mir-02 |
`--------'                                                      `--------'
In 2012, there's a few different patterns you can use. MS details them here:

http://blogs.msdn.com/b/sqlcat/archive/2013/11/20/sql-server-2012-alwayson-high-availability-and-disaster-recovery-design-patterns.aspx

You never want all four nodes having the same storage. At most, you would have two HA clusters that each shares the same volumes between the cluster members, and use log shipping to your DR site.

Will join the chorus, use SQL for this, much more flexibility.

We run the above model (using EQL, no syncrep) in 2012 still. Originally my plan was to use the AlwaysOn feature set to make the mirror side readable, but it turns out that you can't lash together two Windows HA SQL clusters at different locations into one cluster for AO.

Sadly plain mirroring means the second site isn't readable unless you want to snapshot it. But it's DR so not a big deal.

If he is doing this to have a read/write pair and don't need the ultra availability of a 4 node/2 san model, and just want 2 nodes/2 san then AlwaysOn would work out. If he want less waste he could use one of the more traditional replication technologies (i.e. transaction) to push from a 2 node HA write cluster to a farm of read nodes with local storage (or their own shelves)

# ? Jul 16, 2014 17:14

skipdogg: Nov 29, 2004; Resident SRT-4 Expert

Hopefully someone can help me out here... storage is not my strong suit.

I have a NDMP backup of a folder from our old NetApp filer.. Taken in Jan 2012 We moved to EMC shortly after this backup was taken. The data is 5 years old (hasn't changed since 2009), but now needs to be restored for legal reasons.

It seems NDMP to different devices doesn't work. (NetApp NDMP backup restored to EMC is a no go). The backup software is BackupExec 12.5.

What's the best way to get this data restored? Maybe fire up a virtual netapp appliance? It's only 20GB of data or so.

# ? Jul 16, 2014 23:53

Wicaeed: Feb 8, 2005

Nukelear v.2 posted:

Will join the chorus, use SQL for this, much more flexibility.

We run the above model (using EQL, no syncrep) in 2012 still. Originally my plan was to use the AlwaysOn feature set to make the mirror side readable, but it turns out that you can't lash together two Windows HA SQL clusters at different locations into one cluster for AO.

Sadly plain mirroring means the second site isn't readable unless you want to snapshot it. But it's DR so not a big deal.

If he is doing this to have a read/write pair and don't need the ultra availability of a 4 node/2 san model, and just want 2 nodes/2 san then AlwaysOn would work out. If he want less waste he could use one of the more traditional replication technologies (i.e. transaction) to push from a 2 node HA write cluster to a farm of read nodes with local storage (or their own shelves)

I was finally able to convince our DBA that it would be to his benefit to have all of the Synchronous replication tech running in MSSQL as opposed to in our SAN simply because they can manage it & troubleshoot any problems.

Boss and I sat through an on-site Nimble demo today and he was super impressed. He sure likes his graphs

# ? Jul 17, 2014 01:59

Nukelear v.2: Jun 25, 2004; My optional title text

Wicaeed posted:

I was finally able to convince our DBA that it would be to his benefit to have all of the Synchronous replication tech running in MSSQL as opposed to in our SAN simply because they can manage it & troubleshoot any problems.

Boss and I sat through an on-site Nimble demo today and he was super impressed. He sure likes his graphs

Having fixed that bit, Eql is still is a very solid choice. We've run production SQL off PS6110XS's for a couple years now with no downtime. The tiering between SSD and HD works well and can be scaled out with more pure ssd boxes or big cheap archive boxes or more hybrids. The benefits of a single vendor certified stack are always nice. Bigger vendor who isn't likely to be bought up and disappear. Most importantly you have two vendors who can deliver what you want and you can price them against each other. Let them know you've got solutions from a couple vendors, be knowledgeable about their relative strengths so that when they try to dog each others products you know what is noise and what is real, then get ready for the discounts, so many discounts. I've never known Dell to lose on price. Either way, both are nice kit and you'll probably be happy with either.

Edit: Also sync rep sucks, async forever.

Nukelear v.2 fucked around with this message at 14:55 on Jul 17, 2014

# ? Jul 17, 2014 14:34

parid: Mar 18, 2004

skipdogg posted:

Hopefully someone can help me out here... storage is not my strong suit.

I have a NDMP backup of a folder from our old NetApp filer.. Taken in Jan 2012 We moved to EMC shortly after this backup was taken. The data is 5 years old (hasn't changed since 2009), but now needs to be restored for legal reasons.

It seems NDMP to different devices doesn't work. (NetApp NDMP backup restored to EMC is a no go). The backup software is BackupExec 12.5.

What's the best way to get this data restored? Maybe fire up a virtual netapp appliance? It's only 20GB of data or so.

No idea if this would work...

NetApp has these VM simulators. Maybe get a demo for one, do your restore, then copy the data off?

# ? Jul 17, 2014 16:07

skipdogg: Nov 29, 2004; Resident SRT-4 Expert

parid posted:

No idea if this would work...

NetApp has these VM simulators. Maybe get a demo for one, do your restore, then copy the data off?

This is the path I'm going to try, the simulators are a restricted download though, trying to work through that right now.

# ? Jul 17, 2014 17:02

Wicaeed: Feb 8, 2005

Nukelear v.2 posted:

Having fixed that bit, Eql is still is a very solid choice. We've run production SQL off PS6110XS's for a couple years now with no downtime. The tiering between SSD and HD works well and can be scaled out with more pure ssd boxes or big cheap archive boxes or more hybrids. The benefits of a single vendor certified stack are always nice. Bigger vendor who isn't likely to be bought up and disappear. Most importantly you have two vendors who can deliver what you want and you can price them against each other. Let them know you've got solutions from a couple vendors, be knowledgeable about their relative strengths so that when they try to dog each others products you know what is noise and what is real, then get ready for the discounts, so many discounts. I've never known Dell to lose on price. Either way, both are nice kit and you'll probably be happy with either.

Edit: Also sync rep sucks, async forever.

We'll see how it plays out. This is kind of the first time I've had quotes in hand from both vendors, and they are both quoting solutions that will work for us.

Equallogic is offering a lot more usable storage, and coming in well under our budget as we already have some Equallogic tech onsite, thus we don't need to buy as much equipment. However, the available IOPS of their array is only about 3k for the shelf we are buying.

Nimble is offering something we haven't had before: Incredibly fast storage for a decent price. The original plan was to run just our critical environments on Nimble, but with the IOPS we can get from their equipment, we can potentially host some of our less critical DBs off of our SAN.

# ? Jul 17, 2014 20:32

YOLOsubmarine: Oct 19, 2004; When asked which Pokemon he evolved into, Kamara pauses.

"Motherfucking, what's that big dragon shit? That orange motherfucker. Charizard."

skipdogg posted:

This is the path I'm going to try, the simulators are a restricted download though, trying to work through that right now.

You need a registered support account (I do not know why, we really ought to make it freely available) to download the simulator. I can grab it and stick it on dropbox or something if you can't find another way to get it.

# ? Jul 17, 2014 21:01

skipdogg: Nov 29, 2004; Resident SRT-4 Expert

NippleFloss posted:

You need a registered support account (I do not know why, we really ought to make it freely available) to download the simulator. I can grab it and stick it on dropbox or something if you can't find another way to get it.

I registered on the support site and submitted a ticket for increased access (currently guest). I could probably find the Serial Numbers of our old NetApp 3020's if I had to.

I've been looking at using Solaris to do a ufsrestore as well, but that has it's own set of challenges.

We may even have to try to find a service to do this...

If you could hook me up... I'd be very grateful.

# ? Jul 17, 2014 21:45

YOLOsubmarine: Oct 19, 2004; When asked which Pokemon he evolved into, Kamara pauses.

"Motherfucking, what's that big dragon shit? That orange motherfucker. Charizard."

skipdogg posted:

I registered on the support site and submitted a ticket for increased access (currently guest). I could probably find the Serial Numbers of our old NetApp 3020's if I had to.

I've been looking at using Solaris to do a ufsrestore as well, but that has it's own set of challenges.

We may even have to try to find a service to do this...

If you could hook me up... I'd be very grateful.

If you have a solaris system that you can zone a drive from your library to then that's probably the easiest route. This article covers the necessary steps:

http://thiers.net/2011/06/09/use-ufsrestore-to-restore-multi-tape-ndmp-backup/

I'll grab a copy of the ONTAP vsim and upload it somewhere when I get home in a bit.

# ? Jul 18, 2014 00:28

skipdogg: Nov 29, 2004; Resident SRT-4 Expert

They just upgraded my support level and I downloaded the files so no worries man. Thanks for your help.

NippleFloss posted:

If you have a solaris system that you can zone a drive from your library to then that's probably the easiest route. This article covers the necessary steps:

http://thiers.net/2011/06/09/use-ufsrestore-to-restore-multi-tape-ndmp-backup/

I'll grab a copy of the ONTAP vsim and upload it somewhere when I get home in a bit.

Thanks for this link, I didn't find this one via Google. My main problem is the only place I can spin up a Solaris box is on our VMware cluster, and I'm not sure if I can pass the tape drive through the physical host to the VM.

# ? Jul 18, 2014 01:38

YOLOsubmarine: Oct 19, 2004; When asked which Pokemon he evolved into, Kamara pauses.

"Motherfucking, what's that big dragon shit? That orange motherfucker. Charizard."

skipdogg posted:

They just upgraded my support level and I downloaded the files so no worries man. Thanks for your help.

Thanks for this link, I didn't find this one via Google. My main problem is the only place I can spin up a Solaris box is on our VMware cluster, and I'm not sure if I can pass the tape drive through the physical host to the VM.

Cool, glad you were able to get it downloaded.

You might be able to use SRIOV or DirectPathIO to make the drive available to the VM, depending on what type of drive/library it is and your ESX version.

# ? Jul 18, 2014 02:04

Dilbert As FUCK: Sep 8, 2007; by Cowcaster; Pillbug

How many people here have oracle storage? Mind if I probe your brain for a bit?

# ? Jul 21, 2014 15:32

Yaos: Feb 22, 2003; She is a cat of significant gravy.

I've got a stupid question, if you've been in the VM thread it's equally as simple as the question I asked over there.

I took over a position where the previous person just memorized everything and left no documentation behind. Luckily I can pick the brain of out network admin so I'm not completely alone. We have an EMC VNX series SAN. How do I go about monitoring the health of the SAN?

# ? Jul 22, 2014 02:25

Dilbert As FUCK: Sep 8, 2007; by Cowcaster; Pillbug

The VNX has an email feature that can email you alerts, you can also login to the box at anytime and go to

What you can also do to check at anytime through unisphere is to select the array>system>monitoring and alerts

Reports can also pull out some stuff on demand, but monitoring and alerts is more what you want.

# ? Jul 22, 2014 03:16

Nitr0: Aug 17, 2005; IT'S FREE REAL ESTATE

Yaos posted:

I've got a stupid question, if you've been in the VM thread it's equally as simple as the question I asked over there.

I took over a position where the previous person just memorized everything and left no documentation behind. Luckily I can pick the brain of out network admin so I'm not completely alone. We have an EMC VNX series SAN. How do I go about monitoring the health of the SAN?

How old of the VNX? We have an old mofucker and we couldn't do anything because EMC wanted 30k for monitoring tools and the dumps we could do were encrypted and could only be read by EMC.

gently caress EMC btw.

# ? Jul 22, 2014 03:44

Docjowles: Apr 9, 2009

Nitr0 posted:

How old of the VNX? We have an old mofucker and we couldn't do anything because EMC wanted 30k for monitoring tools and the dumps we could do were encrypted and could only be read by EMC.

gently caress EMC btw.

Terrible pricing on proprietary devices is not limited to EMC. We recently looked into what it would take to move all of our app's traffic to SSL on our F5 load balancers. Right now we only run a couple very sensitive transactions over SSL like logins. The quote we got back for the licenses was well into 6 figures :lol:

The hardware could do it no problem, it was just a matter of backing dump trucks of money up to the VAR to unlock the functionality.

We opted not to go with that solution.

# ? Jul 22, 2014 04:29

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Dilbert As gently caress posted:

How many people here have oracle storage? Mind if I probe your brain for a bit?

I have two ZFS appliances. I'll answer what I can.

# ? Jul 22, 2014 04:58

Vulture Culture: Jul 14, 2003; I was never enjoying it. I only eat it for the nutrients.

One of the engineers on my old team actually ended up reverse-engineering IBM's internal performance data storage service on SONAS because the performance was so bad on the supported export method. :psyduck:

# ? Jul 22, 2014 05:56

Langolas: Feb 12, 2011; My mustache makes me sexy, not the hat

Nitr0 posted:

How old of the VNX? We have an old mofucker and we couldn't do anything because EMC wanted 30k for monitoring tools and the dumps we could do were encrypted and could only be read by EMC.

gently caress EMC btw.

If you aren't getting the monitoring software for free/rear end cheap on your quotes, you didn't play your cards right.

Also don't get me started on Cisco and Microsoft licensing holy gently caress that can be a nightmare

One trick I've used for customers is to have their hosts monitor disk util and items and if we see any performance issues, I can get EMC to give me a free performance review from their support guys. They get to the nitty gritty and find issues to fix or if I need to add some more spindles.

Second trick is to say we need to buy some more EMC equipment but we need help sizing a quote, can you do a performance review and let us know whats up?

# ? Jul 22, 2014 17:21

Sickening: Jul 16, 2007; Black summer was the best summer.

Langolas posted:

If you aren't getting the monitoring software for free/rear end cheap on your quotes, you didn't play your cards right.

Also don't get me started on Cisco and Microsoft licensing holy gently caress that can be a nightmare

One trick I've used for customers is to have their hosts monitor disk util and items and if we see any performance issues, I can get EMC to give me a free performance review from their support guys. They get to the nitty gritty and find issues to fix or if I need to add some more spindles.

Second trick is to say we need to buy some more EMC equipment but we need help sizing a quote, can you do a performance review and let us know whats up?

I haven't heard of anybody getting the storage monitoring tools for free or cheap unless they are already spending a god awful sum. Am I missing something here?

# ? Jul 22, 2014 17:24

Dilbert As FUCK: Sep 8, 2007; by Cowcaster; Pillbug

adorai posted:

I have two ZFS appliances. I'll answer what I can.

How many VM's are you hitting per appliance?
What kind of IOPS are you seeing?
Any major issues with your appliance as far a learning curves for people unfamiliar with ZFS, oracle's UI, etc?
What kind (if any) compression/dedupe ratios are you seeing?
How is the VMware intergration with things like cloning, backups, replication etc.
What storage protocol(s) are you using?

I'm largely familiar with ZFS so I am sure the thing pumps out a bunch of IOPS, has decent compression and is generally easy as cake; but it's nice to talk to someone with the actually oracle stuff.

Looking at the ZS3-2 right now, not sure what appliances you have.

Also if it's okay is there any ballpark numbers?

I really want to avoid ExtremIO as much as possible, not because it's rushed to market or anything...

Dilbert As FUCK fucked around with this message at 17:48 on Jul 22, 2014

# ? Jul 22, 2014 17:42

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Dilbert As gently caress posted:

How many VM's are you hitting per appliance?
What kind of IOPS are you seeing?
Any major issues with your appliance as far a learning curves for people unfamiliar with ZFS, oracle's UI, etc?
What kind (if any) compression/dedupe ratios are you seeing?
How is the VMware intergration with things like cloning, backups, replication etc.

I'm largely familiar with ZFS so I am sure the thing pumps out a bunch of IOPS, has decent compression and is generally easy as cake; but it's nice to talk to someone with the actually oracle stuff.

Looking at the ZS3-2 right now, not sure what appliances you have.

I have two 7320 HA pairs each with a single shelf of 900GB drives, two readzillas and two logzillas. Our primary use of the storage is VDI. We host around 500 VDI sessions, and in the early morning regularly hit 10k IOPS. Nearly all of our read requests are satisfied in arc, a very small percentage are satisfied by l2arc, and then by disk. At a normal time of day, let's say 10am, we are generally pushing 7000 read IOPS and 400 write IOPS on NFSv3. Unfortunately due to the way that ZFS handles writes, that's pretty close to the limit of write IO and once we hit it the system runs SIGNIFICANTLY slower for both reads and writes. We just purchased an additional shelf for each HA pair, which will theoretically double my write capacity. Of note, we chose raidz2 for our disk layout for a variety of reasons. If I could go back I might go with mirrored vdevs, but probably not.

The appliance itself is pretty easy to use, it's all web gui driven (you can get into a traditional solaris cli). The biggest oddity is the way replication works, especially regarding the difference between projects and shares. You can end up with a lot of snapshots. We picked up most of what we needed in a day or so.

We do not use deduplication, but we do use compression. For compression, the VMware stuff compresses to roughly 1.5x (so 2/3rds the original size) using lzjb and makes a significant difference in IO (improves it). Using gzip chewed up too much CPU and actually reduced the performance, despite better compression ratios.

Cloning replication and snapshots just take a bit to pick up on the specifics of, but it's all pretty basic stuff. Click a button to snap, click another one to clone. Replication makes it's own snaps.

I have not leveraged any tools to integrate with VMware, though they do have a connector for SRM.

We bought this stuff because it was so inexpensive for the performance we are getting. I am extremely happy with the purchase, and would absolutely do it again in the future. I do not think that for some applications it is appropriate, for instance write heavy databases, but we have a netapp for those. CIFS and read heavy VMware? It's an excellent value proposition.

# ? Jul 22, 2014 17:54

Langolas: Feb 12, 2011; My mustache makes me sexy, not the hat

Sickening posted:

I haven't heard of anybody getting the storage monitoring tools for free or cheap unless they are already spending a god awful sum. Am I missing something here?

Spend more money? I must have an odd region because I've never not seen it get discounted or given away.

# ? Jul 22, 2014 18:00

Dilbert As FUCK: Sep 8, 2007; by Cowcaster; Pillbug

adorai posted:

We bought this stuff because it was so inexpensive for the performance we are getting. I am extremely happy with the purchase, and would absolutely do it again in the future. I do not think that for some applications it is appropriate, for instance write heavy databases, but we have a netapp for those. CIFS and read heavy VMware? It's an excellent value proposition.

Yeah we have a similar number of VDI and expect more growth, as well as a few hefty sql environments. We are currently looking at all flash but I am leaning more towards a hybrid array due to how rushed some of the vendors are with their all flash arrays; as well as I don't like to build a SAN environment based of "what you should get from Dedupe", to me build the environment to capacity and IO without factoring in on dedupe and role.

I mean I get dedupe with windows VM's the more you add the more you save, but still.

# ? Jul 22, 2014 18:37

GrandMaster: Aug 15, 2004; laidback

Nitr0 posted:

How old of the VNX? We have an old mofucker and we couldn't do anything because EMC wanted 30k for monitoring tools and the dumps we could do were encrypted and could only be read by EMC.

thats only for performance stats (navi/unisphere analyzer) and yeah, it's stupidly expensive.
you can still send fault alerts via email without that, just configure under the system/monitoring section in unisphere.

# ? Jul 22, 2014 18:58

Yaos: Feb 22, 2003; She is a cat of significant gravy.

Dilbert As gently caress posted:

The VNX has an email feature that can email you alerts, you can also login to the box at anytime and go to

What you can also do to check at anytime through unisphere is to select the array>system>monitoring and alerts

Reports can also pull out some stuff on demand, but monitoring and alerts is more what you want.

Thanks for the help, I'm dumb and impatient.

Our network admin had the IP for it and we got right in and he even set me up to get alerts. I'm learning new stuff every day though.

# ? Jul 22, 2014 23:11

YOLOsubmarine: Oct 19, 2004; When asked which Pokemon he evolved into, Kamara pauses.

"Motherfucking, what's that big dragon shit? That orange motherfucker. Charizard."

Dilbert As gently caress posted:

We are currently looking at all flash but I am leaning more towards a hybrid array due to how rushed some of the vendors are with their all flash arrays

NetApp has an all flash version of both FAS and E-Series that are based on existing platforms and codelines and don't have the same first adopter worries. HP has something similar with 3PAR. There are options to get an AFA without being a guinea pig for an entirely new platform. Internal testing on the all flash FAS has actually shown some incredibly good performance despite not being a platform developed for flash from the ground up. It's definitely competitive with Pure, for instance.

# ? Jul 23, 2014 00:01

Adbot: ADBOT LOVES YOU

# ? May 9, 2024 22:17

adorai: Nov 2, 2002; 10/27/04 Never forget; Grimey Drawer

Dilbert As gently caress posted:

as well as a few hefty sql environments.

I think it would probably perform quite well in raid 10 for these workloads, but do not have empirical evidence for it. The thing that most people do not realize about ZFS is that it was not designed with speed as the primary or even secondary feature. It was designed for data integrity and scalability. Speed was bolted on later. It does a very good job of hiding behind immense amounts of relatively inexpensive cache to hide this fact.

# ? Jul 23, 2014 00:35

The Something Awful Forums > Discussion > Serious Hardware/Software Crap > Enterprise Storage Megathread: Why is my NAS a SAN?

«‹›207 »