Do you have reference to (document with) this recommendation? ... View more

If you can do it with “normal” Windows server, you most probably can do it with filer. Is it possible with Windows? I think it could be possible from technical point of view, but access rights management becomes nightmare, as you will have to use raw SIDs for one domain (no backward SID-to-name resolution). ... View more

Some hosts are accessing LUNs via non-optimal path (path to partner head). While it may seem to not have immediate impact, it means that - There are some performance implications because all requests go via NVRAM/cluster interconnect - There may be high-availability implications; may be this is the only remaining path so losing it means completely losing access to data The reasons vary from misconfigured host to faulty hardware; if you are not sure how to approach it open support case with NetApp, they will guide you. ... View more

I do not think it will work. You really need FC shelf. ... View more

TCP window size is amount of unacknowledged data host is willing to accept from partner. So it is set independently on both sides of connection. You should be careful with word “performance” because it is meaningless by itself. Before evaluating any setting for mythical “performance impact” you need to define your objectives (speed, amount of data, response time). TCP window size helps with bandwidth (i.e. – amount of data that can be transmitted in given time) on high latency links. It usually has little impact on LAN except when set to ridiculously small value. ... View more

FAS will accept all frame sizes up to maximum, but your clients may reject large frames from filer. In general, MTU must be the same on all devices along the path (including all switches in between). ... View more

Selective acknowledge (SACK) may help with retransmits. ... View more

It depends on how many disks you have. If this is just a couple of them, it is IMHO better to increase RG size and add them into existing RG. If you have many disks (let’s say, starting from 10 available) it may make sense to create new RG. In general it is recommended to have raid group of equal size as far as possible. It really depends on future plans, whether you plan to add more disks later. Both is possible online without interruption. ... View more

You can create aggregate with 19 1TB data disks which means 23 disks in total (2 RAID_DP RGs). I would not try to oversubscribe it, really … but it is interesting to know what happens ☺ ... View more

You do not need to backup anything., When head is replaced, all existing data, including system volume and configuration, is retained. ... View more

For the sake of correctness ☺ When client writes 2K, DOT must read whole 4K from disk to update partial content. If these 2K are not aligned at WAFL block boundary, DOT must read two whole 4K blocks, which potentially may be spread far away on disk. Of course, redo logs tend to be relatively small and are written sequentially and so have good chances of being cached anyway; if not, at least UI rate is probably lower than from other sources. ... View more

Yes, it is confusing. You use target-type to distinguish between volume or aggregate in snap-list, but this parameter is not described for any other API call. ... View more

Could you show the exact dd command you used as well as NFS mount options? ... View more

Both are correct ☺ While you can change onboard adapter mode, it does not necessarily mean that all modes are supported in all cases. (No, I do not know the reason why NetApp mandates external adapter). Notice that e.g. for V3210 onboard adapters are supported. You can set one port to initiator and another to target although this is not recommended. ... View more

It is not possible to answer without detail logs from CLARiiON and NetApp, so you should really open support case, probably both for NetApp and CLARiiON. But it is quite possible that failed disk blocked backend bus on one side of CLARiiON that made accessing this RG impossible and required failover to another SP. ... View more

You will never be able to max out system with single single-stream command like dd. You need multiple concurrent IO streams for this. ... View more

Is “disk remove_ownership” in any way different from “disk assign –s unowned”? The latter can be also used in standard admin mode. ... View more

The needed counters were added to Linux kernel in 2.6.17. If you have this version (or vendor kernel that backported it) you could use nfs-iostat (http://git.linux-nfs.org/?p=steved/nfs-utils.git;a=blob;f=tools/nfs-iostat/nfs-iostat.py;h=9626d42609b9485c7fda0c9ef69d698f9fa929fd;hb=HEAD) to obtain exactly the information you need. ... View more

I am not sure whether this qualifies as “have to”, but if you use dedupe, you have less physical blocks to access and they likely are accessed more frequently, which means they have better chances to be cached by PAM. OTOH, someone (may be, it was you – sorry, do not remember) reported that running A-SIS trashed PAM. If this is really true, and it happens with flexscale.lopri_blocks off, it looks more like a bug – for me A-SIS job definitely qualifies as low priority w.r.t. PAM caching. ... View more

Just map volume and use host command to move data. ... View more

Do you ask how to install them or how to order them? When ordering, you specify single active/active system with single license bundle; internally NetApp expands it to two license sets. When installing, each controller has own licenses, that could be looked up under serial number of respective controller. ... View more

I got curious and checked on MS premier support site – fix is available for x86. Not all MS hotfixes are provided for public download; sometimes you need to explicitly request them (they are free, just provided on request). Have you tried to contact MS for this fix? ... View more

Actually, default is 1. 2 are recommended for maintenance center to be functional, but it is not hard requirement. Of course, pretty GUIs could force you to something else … On NetApp itself it is controlled by option raid.min_spare_count ... View more