Chopper3's questions

We have a lot of datastore clusters, most if not all are supposed to have sDRS schedules applied to stop them moving VMs at busy periods.

Obviously we could trawl through them manually to check if they're how they should be but I like to do things like this via PowerCLI if possible.

I've tried what I can and can't seem to find a command-string to dump this information - does anyone know if it's possible please?

As far as I'm aware in Windows Server 2012 R2 you can configure Storage Spaces to have two tiers and it'll migrate data between them automatically. These tiers are called 'SSD' and 'HDD' and are meant to utilise those disk types.

My question is can you 'lie' to it about these and assign a disk pool of 15krpm HDD's to the 'SSD' tier and 7.2krpm HDD's to the 'HDD' tier - or does it somehow check this? I can't test this in our labs right now as the hardware hasn't arrived just yet sorry.

Since vSphere 5.x we can now request a combined vMotion and Storage vMotion at the same time via the Web Client - but can it be done via the CLI (not PowerCLI)?

I've read the documentation for 5.1 CLI and googled it pretty hard but I can't find any reference. Also if I do find an answer I thought it should have a place for it be easier to find.

We bought some software from a small'ish company, it's a Windows 32-bit video content workflow manager, there's been some customisation by them.

We've been working fine for over a year running this code in a VMWare ESXi 4.1u2 VM on W2K3EE-32-bit (that's what they support running it on).

Then they updated their code a month or so back and we started seeing one of the vCPUs periodically pegging at 100%, the second vCPU is fairly idle, say 5-7% - so we just assumed that the code's badly threaded and contacted them about it.

They've now come back to us saying that their code doesn't work in a VM, they've known about this requirement for 18 months or so, and that they want us to V2P it. They say they only see this problem when ran inside VMs. I've a call with their senior programmer scheduled in a few hours to discuss.

Now luckily we have a few physicals that we can do this on, bit time-consuming but do'able.

My question however is that given this VM doesn't touch any hardware directly, is on a very modern host and actually has very low requirements (2 x vCPU, 4GB, 20GB boot vdisk, 100GB data vdisk, single vNIC and nothing else) what could possibly be the issue with running it in a VM, if there is one?

Obviously I'm strongly pursuing this with them but I just wondered if anyone else has found a regular application, that somehow misbehaves inside a VM but not on a physical.

I know it sounds odd but I need a slower or cached filesystem.

I have a lot of firewalls that are syslog'ing their data to a pair of Linux VMs which write these files to their 'local' (actually FC SAN attached) ext3-formatted disks and also forward the messages to our Splunk servers.

The problem is that the syslog server is writing these syslog messages as hundreds, sometimes thousands, of tiny ~4k writes per second back to our FC SAN - which can handle this workload right now but our FW traffic's going to be growing by at least a factor of 5000% (really) in coming months and that'll be a pain for the SAN, I want to fix the root cause before it's a problem.

So I need some help figuring out a way of getting these writes cached or held-off in some way from the 'physical' disks so that the VMs fire off larger, but less frequent, writes - there's no way of avoiding these writes but there's no need for it to do so many tiny ones.

I've looked at the various ext3 options, setting noatime and nodiratime but that's not made much of a dent in the problem. Obviously I'm investigating other file systems but thought I'd throw this out in case others have the same problem in the future.

Oh and I can't just forward these messages to Splunk, our firewall team insist they're in their original format for diag purposes.

Bit of an open-ended question here sorry, hence why it's a CW, it's to help inform some toolset writers that I work with what direction to go with their new product, hope you don't mind.

If you could create your own management tool for your VMWare hosts, and optionally vCenter/s, are there any stats/reports/information that you'd like to get that you either can't using current tools or that are hard to get at or combine?

For instance I'm unaware of any built in tools to let me know how many IOPS per LUN or array any given cluster, data center or VC is using, I had to buy vCenter Operations to get at that easily. What else would be good info/reports to get at that are hard to get now?

We're lucky, every server we have has multiple NICs/HBAs/CNAs connected to multiple switches and this approach has kept our platform up on numerous occasions. That said we ran into a problem last week that I'm not sure how to fix.

We had a switch that was carrying a good chunk of our traffic crash (the details aren't important but it was a Cisco 6509, it had a hard CPU crash and didn't come back up automatically). Unfortunately it left its line cards working (i.e. L1 & L2 up) but lost all of its uplinks. The servers connected were the following;

• Windows Server 2003 32-bit EE SP2 with Veritas Storage Foundation
• Oracle Enterprise Linux 5.3 64-bit
• VMWare ESXi 4.0
• NetApp 3040 running OnTap 7.3.2

All of these machines failed to detect the crashed switch and kept sending traffic its way rather than detecting the failure and moving their traffic to the another switch.

I need help looking at my options for better multipathing, this can't be the first time this has happened - there must be other ways of doing this (polling the HSRP interfaces for instance) - can you help?

Thanks in advance.

I'm a server guy, specifically a blade guy and know my high-end gear well but I've got a new little challenge that I'm a bit lost on and you may be able to help.

I have a server spec that will do a specific job, I've looked at virtualising it and for dull reasons it's not going to work (not just yet anyway).

The machines have to have;

• dual core consumer processor with quite low power/heat (I'm thinking one of the newer i3's perhaps?)
• no more than 4GB of regular desktop memory
• two PCIe slots - one capable of taking a single-slot x16 GPU (not a 480 or similar, more along the lines of a 9800GT or 240 etc. - about 150W max), the other slot is for a custom low power DSP
• Built in regular dual-channel sound card
• a single GigE PXE/iSCSI-bootable NIC
• I don't need any USB, keyboard, mouse, sound I/O, SATA/PATA or DVD/hard-disk at all.
• They'll be running either XP or W7 (32-bit).

Now I need to get as many of these into a data centre as possible, I also want them as inexpensive as possible too (given the base specs), something I'm not normally bothered about therefore no expert in. I'll need several hundred or thousands of these machines, and have around an 8KW-per-rack limit (this could go up a little).

I normally use HP blades but even their BL2x220's work out very expensive and don't give me exactly what I need anyway. I looked at SGI/Rackable but they're all server-oriented with Xeons etc.

What are your thoughts? I know this isn't directly server related but it is for professional reasons. Thanks for your help.

I've ran into a problem and I'm looking for a new secure protocol/client/server that's faster over a 1Gb/s fibre link - let me tell you the story...

• I have a pair of redundant, diversely-routed, 1Gb/s links over a distance of around 250 miles or so (not dark fibre but a dedicated point to point link, not a mesh).
• At the 'client' end I have a HP DL380 G5 (2 x dual-core 2.66Ghz Xeon's, 4GB, Windows 2003EE 32-bit), at the 'server' end I have a HP BL460c G6 (2 x quad-core 2.53Ghz Xeons, 48GB, Oracle Linux 5.3 64-bit).
• I need to transfer around 500 x 2GB files per week from the client to the server machines per week - but the transfer NEEDS to be secure.
• Using both iPerf or regular FTP I can get ~80MB/s of transfer pretty consistently, which is great.
• Using WinSCP or Windows SFTP I can't seem to get more that ~3-4MB/s, at this point the server's CPU is >3% busy while CPU0 of the client goes to ~30% utilised. We've tried editing various TCP window sizes with little success.

• Both ends are connected to quite low-usage Cisco Cat6509's with Sup720's.

• I can replace the client machine with a newer machine and/or move it to Linux - but this will take time.

Clearly these single-threaded secure Windows clients are introducing too much latency doing their encryption.

So a few questions/thoughts;

• Are there any higher performing secure protocols or client software for Windows that I could try? I'm pretty protocol-gnostic so long as it'll work between Windows and Linux.
• Should I be using hardware to do the encryption, either in the client or the network parts? If so what would you recommend?
• I'm not convinced that just swapping the server would be that much faster, the CPU was only at 30% but then again that's higher than I'd have expected given the load - moving to Linux at the client end may be a better idea but would be quite disruptive.
• Am I missing a trick?

Thanks in advance.

We have a pair of new diversely-routed 1Gbps Ethernet links between locations about 200 miles apart. The 'client' is a new reasonably-powerful machine (HP DL380 G6, dual E56xx Xeons, 48GB DDR3, R1 pair of 300GB 10krpm SAS disks, W2K8R2-x64) and the 'server' is a decent enough machine too (HP BL460c G6, dual E55xx Xeons, 72GB, R1 pair of 146GB 10krpm SAS disks, dual-port Emulex 4Gbps FC HBA linked to dual Cisco MDS9509s then onto dedicated HP EVA 8400 with 128 x 450GB 15krpm FC disks, RHEL 5.3-x64).

Using SFTP from the client we're only seeing about 40Kbps of throughput using large (>2GB) files. We've performed server to 'other local server' tests and see around 500Mbps through the local switches (Cat 6509s), we're going to do the same on the client side but that's a day or so away.

What other testing methods would you use to prove to the link providers that the problem is theirs?

I have a smallish lab environment (16 x ESX4iU1 hosts and VC4U1) that I periodically want to backup.

Normally in production we snap to secondary SAN boxes then have disk-based VTL backups via NetBackup which eventually migrate to off-site removable disks but this seems like an overkill for my own kit.

I've spent a bit of time with vSphere's 'Data Recovery' appliance, it was easy enough to setup and I've not really ran into any issues with it but that doesn't mean I trust it fully.

Have you had any experiences with it, positive or negative that would help me decide whether to trust it or pay Symantec for more licences?

Thank you in advance.

When we setup FC LUNs for our MSSQL boxes we rarely have to present them with more than about 8 distinct LUNs of various kinds (Quorum, MSDTC, TempDB, Data, Logs, Backup and a few others).

We have a new Oracle DBA, and he's given me a list of LUNs he wants for his first new server - there's 38 of them! and this is for a really basic DB box, with only a single DB. They're all fairly small (100GB) LUNs and they clearly bolt together using ASM in an LVM-type way.

Is the the best way to do this, I'm really no Oracle expert but it seems overcomplex to me, what's your thoughts and experiences on this matter?

For one of our platforms I have about 20 MSSQL 2K3 machines that need to drop their backups onto a few NetApp boxes (which then copy the backups to other data centres and does long-term storage).

These NetApps are currently only used by our Linux boxes and are NFS-only, I can buy the SMB licences if needed but need to consider a stable/trustworthy Windows NFS client.

Do you trust the built-in Windows SUA NFS client enough? Is there a 'better' third-party NFS client you'd recommend? Or would you just use SMB to mount the backup volumes?

Thanks.

I want to build a bare-metal RHEL box connected to an FC-array that's presenting it a ~16TB LUN, I can format this LUN any way I like (might choose XFS) but I then want to share this volume out to a number of Apache web-serving RHEL VMs running on ESX 3.5/4 via NFS. Are there any limits to the size of an NFS mount in this range (or possibly slightly larger) that'll get in my way?

I'm a few weeks from being able to test this but will post the answer then if nobody knows right now.

Thanks.

I have a bunch of RHEL/OEL Apache VMs (on ESX 3.5) hardware load-balanced into a little http file farm. I need to expand the dataset they're serving from 1TB to 10TB and although I can use OCFS2 and LVM to make it actually work I'm a bit worried about running into any limitations that I'm currently aware of. In particular I'm worried about LUN sizes that LVM is happy to deal with and also the number of files OCFS2 will support in a directory and also per volume. I've looked around in the obvious places but does anyone have any practical experience of this area?

Thanks in advance.

I've got to buy a bunch of FC SAN switches in the next week or so, I have to, and want to, buy Cisco MDSs.

Servers are HP BL490c G6's in C7000 chassis with Virtual-Connect Flex-10 ethernet interconnects and VC FC interconnects (Emulex HBAs btw), all running ESX 3.5U4 (for now).

I think I've only really got two choices;

• MDS 9509's with dual-supervisors with a single 48-port 4Gb FC card
• MDS 9222i's with single supervisor and the built-in 18-FC-port/4-GigE-FCIP-port option

Both have the same functionality (I think, buying the enterprise licence btw), both have plenty enough performance and adequate ports for now and the next three years.

The 9222i's are about 55% the price of the 9509's - logic says get the 'i's but will I really miss the dual-supervisors? I've got lots of 9509's with dual-supervisors that I'm very happy with but I'm not sure I've every benefitted from the dual-sups in the past and they are nearly twice the price - but if I don't buy them and miss them I can't retrofit them later.

What are your thoughts?

There are a number of ways to deploy/install ESX; RDP/PXE, from CD, remote ISO boot etc.

What's your favourite method and why?

I need an end-to-end multi-client/multi-supplier secure storage system, I can build it using various bits and pieces but it feels overcomplex - am I missing a trick?

Specifically I need to offer multiple clients VPN-secured, secure file transfer into a secure filesystem (as in encrypted on the disk/over fibre) then served out, again via VPN-secured, secure file transfer to a range of other suppliers. No client files can 'mix' at any point and ALL points in the chain must be secure. Oh and it has to be highly resilient, so this must all be clustered/load-balanced. Expecting around 500GB/client/supplier per day, total storage is unlikely to go over 30TB.

My current thought is to use a combination of Checkpoint Virtual Firewalls, 2/3 VMware hosts with lots of NICs connected to the FW, each host connected via Brocade secure FC HBAs to a secure FC SAN box, with each client's VPN tunnel being truncated inside their own VM (which will be clustered using a front-end load balancer), the VMs will then SFTP over the tunnel from the client, drop the files over the secure FC into a dedicated disk-group/LUN/datastore/VMDK chain. Then I'd do the same the other way out to suppliers (except we do apply some DRM as we transmit).

I'd appreciate any suggestions of how to make this simpler, more secure or both.

Thank you.

I'm looking to buy some CNAs to trial with, probably the Emulex board. Just wondering if anyone's using these cards or any competing cards and if so have you have any problems with them yet? In particular with stability, timeouts or high host CPU utilisation.

Oh this will be a mixed test of W2K8, RHEL 5U3 and ESX 3.5U4 via a Cisco Nexus 7K/5K.

Thanks in advance.

We have lots of regular ESX 3.5 hosts but I'm about to take delivery of a bunch of Xeon 55xx based servers and think it's time to question whether to use 3i in the future. It would save me buying local boot disk pairs and most of the patches are to the COS rather than the kernel. We use vCenter so don't need SSH access. So my question is do I really need the COS? What do you need the COS for? Thanks.