About rkaramchedu1

rkaramchedu1 · ‎2009-06-02

A few points to check/adjust settings. Some of them are general in nature, but I thought I'd throw them out as I feel they are applicable Look into the rsize and wsize options on the nfs client and set them to match closely to the workload. may be 32k? Look into using the nfs client option actimeo to cache file attributes. Note: This is NOT identical to noac. The option "noac" will not only disable client side file attr caching but will also disable write cachine, which will tank your performance, not increase it. Increase the number of nfs client threads on the solaris box. By default, I believe it is 8. (It's an /etc/system setting, which requires reboot) Has the aggregate been sized appropriately with adequate number of drives? You may want to engage NetApp or your partner SE to ensure that it is the case. For those who are unaware, Nagle algorithm is applied to outgoing packets and is a method of "pooling" packets and sending them when appropriate, instead of sending packets with too small of a payload (lots of packet with not a lot of data). I believe ONTAP does implement Nagle algo for TCP small packet optimization. Look here: http://now.netapp.com/NOW/cgi-bin/bol?Type=Detail&Display=13188 If all the components in your data path support this, look into turning jumbo frames on Make changes ONE at a time and baseline the performance. Check the network (and check it again) - Nothing against my esteemed colleagues from the networking world, but settings like flow control, portfast etc all matter when one is trying to tune the system for performance. Related links: https://now.netapp.com/Knowledgebase/solutionarea.asp?id=ntapcs4694 https://now.netapp.com/Knowledgebase/solutionarea.asp?id=ntapcs2846 https://now.netapp.com/Knowledgebase/solutionarea.asp?id=kb36582 HTH - and I'd be interested in the progress! rajeev

rkaramchedu1 · ‎2009-05-29

With fastpath enabled, the filer will respond to the same MAC address and tries to send the packet back thru the same interface as it came in. This allows for loadbalancing on multiple trunk interfaces on the same subnet. Disabling this means that ONTAP has to look up its routing table to figure out where to stick the packet in and *might* introduce a delay. In other scenarios, if you have the filers connected to switches running HSRP, then it might conflict with the fastpath setting. I would start by checking to see if there are any errors reported on the filer (netstat -nr, netdiag commands may be useful to check settings) Another thing to check if you have multiple interfaces in the same subnet, then try disabling fastpath. A few URLs to look into: https://now.netapp.com/Knowledgebase/solutionarea.asp?id=ntapcs11976 https://now.netapp.com/Knowledgebase/solutionarea.asp?id=ntapcs2130 http://now.netapp.com/NOW/knowledge/docs/ontap/rel7311/html/ontap/nag/networking/task/t_oc_netw_routing_fastpath.html https://now.netapp.com/Knowledgebase/solutionarea.asp?id=kb7710 https://now.netapp.com/Knowledgebase/solutionarea.asp?id=ntapcs4322 https://now.netapp.com/Knowledgebase/solutionarea.asp?id=ntapcs11740

rkaramchedu1 · ‎2009-05-29

Greetings In a Netbackup (6.0) environment, is it possible to restore Celerra NDMP backups to NetApp NDMP? What information do we need to gather from the Celerra/NBU side to find an answer. thx p.s. This is actually a duplicate post from another location where no answers were received - so my humble apologies in advance. rk

rkaramchedu1 · ‎2009-05-28

We've used SecureCopy with success in those types of migrations. There is also a new tool from Microsoft called RichCopy but it's buggy when it comes to retaining ACLs. Downside of SecureCopy is that it is not free.

rkaramchedu1 · ‎2009-05-05

While technically possible, I do not think this is supported. To configure it however, you'd just have to configure appropriate igroups and maps your LUNS to them.

rkaramchedu1 · ‎2009-04-20

http://communities.netapp.com/message/8649#8649 This presentation has a slide that contains side-by-side comparison of both of the technologies.

rkaramchedu1 · ‎2009-04-13

Josh is right on the money. The NetApp CIFS cannot serve as a DFS root. They can join a DFS root as "leaf nodes" however. You could use VFM for Global namespace purposes or even a Windows VM. Also, if the customer's environment has Macintosh (mac os x etc), then know that the CIFS client on mac os x does not yet work with DFS either..

rkaramchedu1 · ‎2009-03-07

I prefer to use LACP over static etherchannel cause the control protocol of LACP uses keep-alives to detect link loss. Just make sure your switch supports LACP. rajeev

rkaramchedu1 · ‎2008-10-26

Even better! The snap rename command can be used to rename snapshots. The snap sched uses the snapshot names to perform snapshot rotations. Note that this may not work in all cases. For example, a year end snapshot might be on a weekday and if you do not have nightly snapshots turned on, then you might want to take a special snapshot just for the year-end purposes.

rkaramchedu1 · ‎2008-10-25

Not a direct way of doing this using the netapp itself but using a linux/unix/windows admin host, one could setup a cron/scheduler, to take a snapshot. The same script that does the snapshot creation could also do the cleanup (snapshot rotation, if you will) A cron example might look like: 0 0 1 * * /path/to/script > /dev/null 2>&1 For the yearly snapshot, something like this might work: 0 0 31 12 * /path/to/script > /dev/null 2>&1 HTH

rkaramchedu1 · ‎2008-10-18

Q. How does ONTAP calculate the max number of files for a given volume size? A. By default, the volume will get 1 inode per every 32KB of disk space and the # of files that a volume can hold get arrived at using that number. The maximum number of inodes is limited to one inode per one block in the filesystem. (which is 1 inode per every 4KB). It is generally recommended to NOT go that low. Q. How many files can be added per volume above the default setting before you start to see an impact on performance? A. I do not know if there has been any study that was done to measure the file system performance w.r.t to the number of files. Also note that the performance of a volume is not only dependent upon the # of files the volume has but also on the structure of the files. For example, you'll get a lot better performance if you distribute your files across multiple directories at the same level, as opposed to having all of them in 1 directory. Q. How does having so many files affect performance when trying to do a backup? A. Very negatively. Your backup software will spend a LOT of time getting the file list and then actually doing the backup. Kick off a backup and examine the time difference between the time when it starts reading the file system and the time when it actually starts reading any bytes from the file system... that will tell you the story. Q. Is there a best practice for data set containing millions of small files? A. I would try to organize the files in such a way that you have files distrubuted among multiple directories. Also, try not to have too deep of a directory structure. Some other reading http://now.netapp.com/NOW/knowledge/docs/ontap/rel73/html/ontap/cmdref/man1/na_maxfiles.1.htm https://now.netapp.com/Knowledgebase/solutionarea.asp?id=ntapcs11850 https://now.netapp.com/Knowledgebase/solutionarea.asp?id=ntapcs5094 HTH

rkaramchedu1 · ‎2008-10-15

The more I read this thread, the more confusing it becomes. Let me stick to PRE-7.3 for a second. The table I saw above says optimal RAID group size for a 1TB SATA Drive is 17. First off, for SATA, the max RAID group size allowed is 16. So 17 is not even a workable option. Aggregate Size. From my readings, work and understanding, the total size of an aggregate was computed by adding together the raw size of all disks in the aggregate, including parity disks, and regardless of the amount of disk space available to be used in each data disk. What I am now confused about now is the use of the term "raw size". I believe it is the "physical capacity" might be the correct term. More specifically, it is the output from the "sysconfig -r" command from under the "Phys (MB/blks)" column. For e.g. for a 300GB FC Disk Drive, the Physical Size is 274845 MB = 268.40 GB. So for an aggregate limit of 16TB, the total # of drives = 16 * 1024 / 268.40 = 61 drives (parity and data together). Am I correct ?

rkaramchedu1 · ‎2008-09-29

A while ago, I was involved in some tests relating to SnapMirror and documented portions of it here http://rajeev.name/blog/?s=snapmirror (external link). However, I am wondering if folks have done similar work and are able to measure the SnapMirror rates that resulted from the exercise/effort. Without going into any environment specifics, a one liner like "VSM with De-dupe over a T1 link, we observed a rate of XX kbps" might be a useful stat to gather.

rkaramchedu1 · ‎2008-09-15

First off, here's the link: http://now.netapp.com/NOW/knowledge/docs/ontap/rel73/html/ontap/onlinebk/protecting/task/t_oc_prot_sm-migrate-data-btwn-vols.html It can be used to migrate a volume to a different volume. Both the volumes need to be in a SM relationship tho. As the documentation points out, it will do all the leg work of moving the NFS connections to the destination volume, restricting the src volume etc. Imagine you have a volume built on FC drives and you want to archive it off to a SATA volume within the same controller. Once you set up the SM relationship, the last cutover can be done using this option.. While this migrate is happening, the NFS clients will see a small hiccup and will continue to resume... HTH rajeev

rkaramchedu1 · ‎2008-09-11

This appears to be a Apple-NTAP specific issue. Because I got this setup to work with a Windows system. In Mac OS X case, the LDAP request is never made. The communication breakdown occurs (looks like) between mac os x and NTAP. (I tested this with the new version of simulator 7.3 and still the same result) It would be wonderful if some of the CIFS folks can chime in here..

rkaramchedu1 · ‎2008-09-09

Well..with LDAP authentication, wcc does not put out any output since it is not joined into any domain. There is no windows domain to join. The qtree security style is mixed. (I even tried ntfs).

rkaramchedu1 · ‎2008-09-08

Hi Christopher I am using CIFS connection (Apple + K, cifs://<filername>/<share>) cheers - rajeev

rkaramchedu1 · ‎2008-09-08

Absolutely! Now - Having gotten that out of the way, wouldn't the cost of developing a wrapper (albeit, thin) to work with all the various MPxIO systems be very close or be only incrementally less than the cost of developing a MPxIO package itself ? With both the efforts, one has to keep up with the driver changes, certifications etc...A unified UI kit that provides NTAP MPxIO s/w and third-party s/w would be cool indeed. This may be slighly off-topic, but wanted to throw it out there that it could be funny, cool and downright kicka$$ if we can build PowerPath into the support matrix. Not only then NetApp can de-dup EMC/third party storage, migrations from EMC to NetApp will be a breeze. I do understand that this may not be possible cause powermt may have some proprietary code (CLARiiON and SYMM specific)..

rkaramchedu1 · ‎2008-09-08

I *think* this is a supposedly supported and possible configuration, however it does not work in my tests. Where To: Get a mac client to mount a SAMBA share from a NetApp Filer, which is using LDAP for user authentication Configuration Steps Setup a LDAP server with at least 1 posixAccount user object. -- DONE Setup the Simulator with LDAP using options.ldap settings and editing /etc/nsswitch.conf -- DONE Perform cifs setup and configure to use LDAP (#4 in the cifs setup) -- DONE Verify on the console that LDAP lookups can be performed (using the getXXbyYY getpwbyname_r <username>) command. -- DONE Verify CIFS authentication from a CIFS client -- NOT DONE. NO WORK. I am stuck at #5. Even with cifs trace logins on and ldap server logs revved up, when I attempt a CIFS authentication from my mac, nothing happens. No log entries in the LDAP server and no message on the filer console. Any thoughts ?

rkaramchedu1 · ‎2008-08-10

With a simulator, I do not know how to test a "failed" nic without actually downing it.. (it is actually a simulator in a VM on a mac) I knew that it is not persistent across reboots. Now, I am drawing a blank, but if we configure a VIF (single or multimode) and there is a VIF failure, it would cause a cf event, correct, even without NFO enabled ? 'Cause I noticed that one can enable NFO on a VIF and it got me thinking as to its use cases. The active/active guide mentions it but it would help if they added some clarity to it. For e.g. ( say "Yes, if set up to do so using NFO" as opposed to just "Yes, if set up to do so" ) http://now.netapp.com/NOW/knowledge/docs/ontap/rel7251/html/ontap/cluster/failing_over/reference/r_oc_fo_failover-events.html

rkaramchedu1 · ‎2008-08-10

I have a cluster simulator running on which I am trying to test the Negotiated Failover (NFO) functionality for interface. As a background, NFO can be enabled on a physical interface and configured so that if an NFO-enabled interface fails on the partner, a CF event occurs (at least in theory) Here is the node1 configuration of the cluster: node1> options cf cf.giveback.auto.cifs.terminate.minutes 5 cf.giveback.auto.enable off cf.giveback.auto.terminate.bigjobs on cf.giveback.check.partner off cf.takeover.change_fsid on cf.takeover.detection.seconds 10 cf.takeover.on_disk_shelf_miscompare off cf.takeover.on_failure on cf.takeover.on_network_interface_failure on cf.takeover.on_network_interface_failure.policy any_nic (same value in local+partner recommended) cf.takeover.on_panic on cf.takeover.on_short_uptime on node1> ifconfig -a ns0: flags=848043<UP,BROADCAST,RUNNING,MULTICAST> mtu 1500 inet 192.168.97.133 netmask 0xffffff00 broadcast 192.168.97.255 partner inet 192.168.97.135 (not in use) ether 00:50:56:1b:03:f8 (Linux AF_PACKET socket) nfo enabled ns1: flags=808042<BROADCAST,RUNNING,MULTICAST> mtu 1500 ether 00:50:56:1c:03:f8 (Linux AF_PACKET socket) lo: flags=1948049<UP,LOOPBACK,RUNNING,MULTICAST,TCPCKSUM> mtu 4064 inet 127.0.0.1 netmask 0xff000000 broadcast 127.0.0.1 ether 00:00:00:00:00:00 (Shared memory) node1> cf status Cluster enabled, node2 is up. Negotiated failover enabled (network_interface). node1> And here's the node2 configuration of the cluster node2> cf status Cluster enabled, node1 is up. Negotiated failover enabled (network_interface). node2> options cf cf.giveback.auto.cifs.terminate.minutes 5 cf.giveback.auto.enable off cf.giveback.auto.terminate.bigjobs on cf.giveback.check.partner off cf.takeover.change_fsid on cf.takeover.detection.seconds 10 cf.takeover.on_disk_shelf_miscompare off cf.takeover.on_failure on cf.takeover.on_network_interface_failure on cf.takeover.on_network_interface_failure.policy any_nic (same value in local+partner recommended) cf.takeover.on_panic on cf.takeover.on_short_uptime on node2> ifconfig -a ns0: flags=808042<BROADCAST,RUNNING,MULTICAST> mtu 1500 inet 192.168.97.135 netmask 0xffffff00 broadcast 192.168.97.255 partner inet 192.168.97.133 (not in use) ether 00:50:56:0f:25:e3 (Linux AF_PACKET socket) nfo enabled ns1: flags=808042<BROADCAST,RUNNING,MULTICAST> mtu 1500 ether 00:50:56:10:25:e3 (Linux AF_PACKET socket) lo: flags=1948049<UP,LOOPBACK,RUNNING,MULTICAST,TCPCKSUM> mtu 4064 inet 127.0.0.1 netmask 0xff000000 broadcast 127.0.0.1 ether 00:00:00:00:00:00 (Shared memory) node2> cf status Cluster enabled, node1 is up. Negotiated failover enabled (network_interface). node2> However, when I down the ns0 interface on a node, nothing really happens.. node2> date; ifconfig ns0 down Sun Aug 10 14:43:53 GMT 2008 node2> ifconfig -a ns0: flags=808042<BROADCAST,RUNNING,MULTICAST> mtu 1500 inet 192.168.97.135 netmask 0xffffff00 broadcast 192.168.97.255 partner inet 192.168.97.133 (not in use) ether 00:50:56:0f:25:e3 (Linux AF_PACKET socket) nfo enabled ns1: flags=808042<BROADCAST,RUNNING,MULTICAST> mtu 1500 ether 00:50:56:10:25:e3 (Linux AF_PACKET socket) lo: flags=1948049<UP,LOOPBACK,RUNNING,MULTICAST,TCPCKSUM> mtu 4064 inet 127.0.0.1 netmask 0xff000000 broadcast 127.0.0.1 ether 00:00:00:00:00:00 (Shared memory) node2> ping 192.168.97.133 ping: wrote 192.168.97.133 64 chars, error=Network is down ping: wrote 192.168.97.133 64 chars, error=Network is down ping: wrote 192.168.97.133 64 chars, error=Network is down ping: wrote 192.168.97.133 64 chars, error=Network is down ping: wrote 192.168.97.133 64 chars, error=Network is down coffee break.......... node2> date Sun Aug 10 14:51:29 GMT 2008 node2> cf status Cluster enabled, node1 is up. Negotiated failover enabled (network_interface). node2> What am I missing ?

rkaramchedu1 · ‎2008-08-05

It would be definitely helpful. Thinking various scenarios, the following pop up in my head Mmultiple ports with a single switch (multi-mode vif) Multiple ports with a cross-stack switch set (multi-mode vif) Multiple ports with two switches with HSRP (single mode vif) Multiple ports with two switches with HSRP (single mode vif consisting of multimode vifs) Along the way, if your document can also contain a better treatment of what NetApp does and does not support w.r.t 802.3ad and LACP protocols, it would be great. There is enough confusion in people's mind about EtherChannel vs 803.3ad static vs 802.3ad dynamic vs LACP.. And then there are limitations with 10Gbps TOEs... best

rkaramchedu1 · ‎2008-03-10

It would, if someone is just interested in de-dupe rates. However, having it run in a simulator would give a customer a real world feeling. Ability to use ASIS commands, run other complimentary ontap commands and see other snapXXX interoperability would definitely add a whole lot of value.

rkaramchedu1 · ‎2008-03-10

Having the a-sis/de-dupe capability in the ontap simulator would be a good thing. This would allow for customers to see the de-dupe ratios on their data.

rkaramchedu1 · ‎2008-03-09

Well Having supported the life sciences community for a few years, when I first heard of de-dupe, I wanted to see if we can use de-dupe algorithms to seek out the commonalities between multiple genomic data sets between organisms, species etc...However, since NetApp de-dupe works on a WAFL block, compares are restricted to 4KB. If that would be configurable, the resulting scenarios would be interesting at best.. that would be a good test for netapp admins currently supporting such data sets.

Re: nfs tuning for 1G/sec load

Re: ip.fastpath.enable

Celerra to NetApp Migration - NDMP Question

Re: Migrate Data to NetApp (CIFS)

Re: Simultaneously accessing LUNs via iSCSI and FCP (different servers)

Re: Anyone have a good 1 page Snapmirror Snapvault Solutions Comparison?

Re: Can CIFS shares be DFS roots?

Re: lacp vs multi-mode

Re: monthly snapshots?

Re: monthly snapshots?

Re: Max number of files

Re: Raid group size recommendation

Observed SnapMirror rates

Re: How snapmirror "migrate" works

Re: CIFS authentication with LDAP

Re: CIFS authentication with LDAP

Re: CIFS authentication with LDAP

Re: NetApp and Multi-path Management - Feedback Request

CIFS authentication with LDAP

Re: Interface NFO funkyness

Interface NFO funkyness

Re: Deciphering the switch config - How to configure single, multi and LACP vifs

Re: De-Dupe in ONTAP Simulator

De-Dupe in ONTAP Simulator

Re: Most unusual application of dedupe?