toasters

toasters@lists.teaparty.net

13517 discussions

RE: Performance Indicators
by Glenn Walker 15 Sep '06

15 Sep '06

I believe that this is more along the lines of QoS (though not true QoS and not guaranteed). What Jochen is looking for is a bit difficult: you can't really guarantee performance on the head due to workload differences. Each of the machines Jochen listed is a very capable machine - but selling IOPS to a user is not easy. Each workload is different, each application would require a different amount of IOPS in order to facilitate functionality (hard to define these sometimes), you can't easily segregate workloads and adding workload from one user\system would definitely affect existing workloads. For example: CIFS typically requires about 5 IOPS to open a file as a minimum - if you add some newer functionality from 2k3\XP and MS Office, you could be talking about hundreds of IOPS to open a file (pulling file metadata). Perhaps the best thing to do is to talk to sales to get actual numbers, but simply list the capabilities of the filer and manage how the workload is spread among them. One last point: when you speak of performance, you mentioned IOPS - this is workload, not performance. Performance is more a measure of latency at a particular level of work: ie, how much latency for a given amount of IOPS. If I stated that the R200 was capable of about 12000 IOPS (CIFS) per second, that would be true - but I could also position it as being capable of higher if the latencies of the workload were to increase (still the same amount of work, but response time would decrease). You could speak to the user-base about performance in terms of latency and IOPS. Glenn -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Wischer, Sebastian Sent: Friday, September 15, 2006 5:18 AM To: Willeke, Jochen; NetApp Toasters List Subject: RE: Performance Indicators HI Jochen, have you think about FlexCache?? There you give volumes priorities. I know it is not exactly what you are looking for but perhaps with the help of FlexCache you can sell categories to your departments. Here is a small summary from a NetApp slideset regarding FlexCache: FlexShare(tm) is enabled when there are system bottlenecks CPU NVRAM Disk Control of priority by volume Five priority levels Sebastian -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Willeke, Jochen Sent: Friday, September 15, 2006 10:32 AM To: NetApp Toasters List Subject: Performance Indicators Hi Toasters, right now i am thinking about any kind of performance indicator for my fileservice. We do supply storage to other departments of our company. Of course one thing i can sell is space. But there is another indicator which is performance. Has anybody good ideas how to define the performance i can grant to a special service? I was thinking about I/O's per second but do not know what a good number would be. Has anybody ever had to work out those values and can give me something to think about? I am interesting in any experience you had with this topic... I/O's, data-throughput, responsetime....And what are good values which i can grant with netapp systems? We use R200, FAS960 and FAS3050. Thanks in advance Jochen

2 1

Need Advice for changing qtree from Unix to NTFS security style
by Stephen C. Losen 15 Sep '06

15 Sep '06

We have a filer that serves files in a qtree with both NFS and CIFS. The file security style is currently Unix. The qtree has about 7000 files using about 150G. CIFS authentication is done via a Windows DC and we have a Unix passwd NIS map for converting the Windows users to Unix users. There are over 100 folks using this share. Each one owns a directory (not a home directory -- just storage) and each one wants to be able to grant some of the 100 other users access to their folder. Plus they have an admin who needs to be able to manipulate access rights on everything. Obviously Unix security style isn't going to cut it. So we want to switch the qtree to NTFS security style. We switched a test qtree from Unix to NTFS style and discovered that no one (not even a Windows user with admin privs in the domain) has "modify" or "full control" rights on any file or folder that already exists. So it looks like we can't change the ACLs on anything that already exists. If a Windows user creates a new folder or file, then he has the ability to change the ACL. Has anyone else switched a qtree from Unix to NTFS style? Is there anything we can do either before or after we switch the security style to make it possible to set ACLs on files that already exist? I didn't try logging in to CIFS as a local filer user. Is that the secret? I suppose we could create a new empty qtree with NTFS style and copy the files from the old qtree to the new one. But I would really like to avoid that if possible. Steve Losen scl(a)virginia.edu phone: 434-924-0640 University of Virginia ITC Unix Support

2 2

Re: Need Advice for changing qtree from Unix to NTFS security style
by Stephen C. Losen 15 Sep '06

15 Sep '06

> > We switched a test qtree from Unix to NTFS style and discovered that no > one (not even a Windows user with admin privs in the domain) has "modify" > or "full control" rights on any file or folder that already exists. So > it looks like we can't change the ACLs on anything that already exists. > If a Windows user creates a new folder or file, then he has the ability > to change the ACL. > Ooops. Looks like I was wrong here. The owner of the file or folder does have full control. So I guess I can chown everything to me and then from my Windows account I can set all the ACLs. I'll check and see if the local filer "administrator" user can do it. Steve Losen scl(a)virginia.edu phone: 434-924-0640 University of Virginia ITC Unix Support

1 0

Performance Indicators
by Willeke, Jochen 15 Sep '06

15 Sep '06

Hi Toasters, right now i am thinking about any kind of performance indicator for my fileservice. We do supply storage to other departments of our company. Of course one thing i can sell is space. But there is another indicator which is performance. Has anybody good ideas how to define the performance i can grant to a special service? I was thinking about I/O's per second but do not know what a good number would be. Has anybody ever had to work out those values and can give me something to think about? I am interesting in any experience you had with this topic... I/O's, data-throughput, responsetime....And what are good values which i can grant with netapp systems? We use R200, FAS960 and FAS3050. Thanks in advance Jochen

2 2

the Toasters list is now restricting posts!
by Kevin Davis 12 Sep '06

12 Sep '06

Folks, Just an FYI to all: we've had to take the measure of restricting posting to the 'toasters' list, meaning that the list will only accept posts from an address which is subscribed. Sorry for taking this measure, but the nature of email today has left us no alternative (not one that will still allow us to get our jobs done, that is!); it's a way to keep the spam to the list to a minimum. And for what it's worth, the spam to the list is almost nil now. (knock wood and cross your fingers). Hopefully, this measure along with the four (yep...four!) separate facilities we have for filtering the gooey pink stuff will be sufficient to make the list a friendlier place again. For most, this will have no effect. But some may be subscribed via list addresses which will be expanded on the destination domain, and for these folks the ability to reply to or start new threads will be foiled, unless the "From:" address matches the address which is subscribed to the list. In other words: Admin1 is one of 4 admins at XYZ Corp. who are all part of the alias 'toasters(a)xyzcorp.com'. The admins at XYZ subscribe 'toasters(a)xyzcorp.com' to 'toasters(a)mathworks.com'. An email to the 'toasters' list goes to XYZ and is expanded to all 4 admins. Admin1 then they won't be able to reply or create new posts because 'admin1(a)xyzcorp.com' is not subscribed to the 'toasters' list. So...sorry for any inconvenience to individuals subscribed in this manner, but you'll have to subscribe individually in order to post to the list. -- *------------------------------------------*-----------------------* | Kevin Davis (UNIX Sysadmin) | Natick, Massachusetts | | 508.647.7660 | 01760-2098 | | mailto:kevin.davis@mathworks.com *-----------------------* | http://www.mathworks.com | | *------------------------------------------*-----------------------*

2 2

RE: Snap status and wafl scan
by Glenn Walker 08 Sep '06

08 Sep '06

You are correct - the reclamation is a new feature - introduced around 7.0 I believe. Unsure what version of ONTAP the issue was being seen on, but there were some bugs related to that feature-set that were fixed in later versions - some bugs were related to its efficiency, while some were related to the timing (when the freed space was showing back up). Glenn -----Original Message----- From: Chris Thompson [mailto:cet1@cus.cam.ac.uk] Sent: Friday, September 08, 2006 12:07 PM To: toasters(a)mathworks.com Cc: Glenn Walker Subject: Re: Snap status and wafl scan ggwalker(a)mindspring.com (Glenn Walker) writes: > Ok - those are other things than what I previously mentioned. > > Active Bitmap rearrangement is normal - I believe it has been around > since 6.5 (pretty sure it wasn't a flexvol thing, that was 'deswizzler' > scan). It's to keep memory 'pretty', if memory serves. > > The container block reclamation is a scanner responsible for ferreting > out blocks that have been freed and marking them as usable to the system > (delayed free scanner\delayed freer working here). This too is normal. Just in case it wasn't clear, I did not mean to suggest that these things were not "normal", or that they had any implications for filing system integrity. It was the performance aspect of the "container block reclamation" scan that originally brought it to my attention: specifically that after our dumps had finished (and the associated snapshots thus deleted) disc I/O remained high and cache age very low for some time. I don't think ONTAP used to behave like that in 6.5. By contrast the "active bitmap rearrangement" scan (which is always running: it just cycles around the block numbers indefinitely) has no bad effect on performance that I have ever noticed. -- Chris Thompson Email: cet1(a)cam.ac.uk

1 0

RE: jumbo frames support
by Fox, Adam 08 Sep '06

08 Sep '06

Yes, Jumbo frames are supported via the ifconfig command. The command allows you to set the mtusize. Although it's probably worth opening a support case and have them run a perfstat on your system while you are experiencing slow performance. If it's really as slow as you say, it's probably not the mtusize. It might be, but there are probably other things you can adjust in your environment first. But NGS has the resources to work with you to figure those out. Just a thought. -- Adam Fox adamfox(a)netapp.com -----Original Message----- From: kunle windapo [mailto:kwindapo@yahoo.com] Sent: Friday, September 08, 2006 11:26 AM To: toasters(a)mathworks.com Subject: jumbo frames support hello, i have a FAS 270 with volumes nfs mounted on a Solaris 9 sparc server. there is a direct connection from the solaris server to ethernet interface on the FAS 270 running at 1Gb/s. problem is i experience very slow performance on the system, (io related) and i'm troubleshooting. i want to know if NetApp supports Jumbo frames on the Network interface. thank you __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

1 0

jumbo frames support
by kunle windapo 08 Sep '06

08 Sep '06

hello, i have a FAS 270 with volumes nfs mounted on a Solaris 9 sparc server. there is a direct connection from the solaris server to ethernet interface on the FAS 270 running at 1Gb/s. problem is i experience very slow performance on the system, (io related) and i'm troubleshooting. i want to know if NetApp supports Jumbo frames on the Network interface. thank you __________________________________________________ Do You Yahoo!? Tired of spam? Yahoo! Mail has the best spam protection around http://mail.yahoo.com

2 1

RE: Snap status and wafl scan
by Glenn Walker 08 Sep '06

08 Sep '06

Ok - those are other things than what I previously mentioned. Active Bitmap rearrangement is normal - I believe it has been around since 6.5 (pretty sure it wasn't a flexvol thing, that was 'deswizzler' scan). It's to keep memory 'pretty', if memory serves. The container block reclamation is a scanner responsible for ferreting out blocks that have been freed and marking them as usable to the system (delayed free scanner\delayed freer working here). This too is normal. Glenn -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Chris Thompson Sent: Thursday, September 07, 2006 4:01 PM To: toasters(a)mathworks.com Cc: ecoli82(a)msn.com; etraitel(a)gmail.com Subject: Re: Snap status and wafl scan Eyal Traitel writes: > Note that snap status is an advanced command so its output shouldn't > necessarily cause you concerns... Well, while we're talking about things wot mere mortals were not meant to wot of... ec0li <ecoli82(a)msn.com> wrote: > When I check snap status, I see next to recent snapshots : (xxxx/xxxx > remaining). > Also, It seems that this is in fact the wafl scan process "snap create > summary update" The effect worrying me is somewhat different. After a snapshot is deleted the above state lasts only seconds, but there is then a much more extended period where "snap status" looks normal but "wafl scan status" shows e.g. carina*> wafl scan status Volume CUS: Scan id Type of scan progress 1 active bitmap rearrangement fbn 3340 of 4461 w/ max_chain_len 3 925 container block reclamation block 382 of 4461 The "container block reclamation" scan lasts several minutes (and this is a volume & aggregate of only a few hundred GB), and while it's going on disc read activity is high and severely impacts the cache - at least, if you trust the "cache age" shown by sysstat, which I am not sure I do. This is with ONTAP 7.1.1P1 (but similar effects observed in earlier 7.x's). -- Chris Thompson Email: cet1(a)cam.ac.uk

2 1

NDMP backups, timeouts
by Scott T. Mikusko 08 Sep '06

08 Sep '06

I'm running into a recurring issue with backing up my NearStore and a long F760 filer. In the procoess of backing some large filesystems (qtrees), it appears that NDMP is basically timing out and fails. I'm running Veritas NetBackup 6.0 (finally upgrading from 4.5 recently). I think the issue has to do with NDMP starting the dump commands on the filer, which then does a three-way backup to another filer with local tape drives. The debug from ndmp logs shows for a manual full backup of a filer: Sep 05 21:05:57 EDT [ndmpd:58]: Log message: DUMP: creating "/vol/vol0/../snapshot_for_backup.370" snapshot. Sep 05 21:06:05 EDT [ndmpd:58]: Log message: DUMP: Using Partial Volume Dump with Exclude Lists Sep 05 21:07:31 EDT [ndmpd:58]: Log message: DUMP: Date of this level 1 dump: Tue Sep 5 21:05:57 2006. Sep 05 21:07:31 EDT [ndmpd:58]: Log message: DUMP: Date of last level 0 dump: Fri Aug 18 19:20:57 2006. Sep 05 21:07:31 EDT [ndmpd:58]: Log message: DUMP: Dumping /vol/vol0/files3.rt to NDMP connection Sep 05 21:07:31 EDT [ndmpd:58]: Log message: DUMP: mapping (Pass I)[regular files] It never completes the mapping of the files in Pass I. It just sits there. The filesystems I'm backing up are between 250-650GB with a LOT of small files (millions) and many subdirs. Tons of mail and html files. My guess is that the amount of files and dirs are getting large that NDMP can't map them fully and is timing out after 8 hours. It shouldn't take that long to map files for the size of data it's doing; I've seem mapping times be much less for larger sets of data. I know that the filesystem topology can cause NDMP backups to be slow, depending on file sizes, data layout, filer load, network etc, but this is happening on moderately busy filers or bone idle NearStores. Anyone run across lengthy DUMP times? Chewing through NOW and Veritas support site hasn't turned up anything obvious. Just curious if others have run into NDMP/dump issues like this. -Scott

4 3

Snap status and wafl scan
by ec0li 07 Sep '06

07 Sep '06

Hello, When I check snap status, I see next to recent snapshots : (xxxx/xxxx remaining). Also, It seems that this is in fact the wafl scan process "snap create summary update" Does anyone knows what is this process, its purpose, what happens if the system reboots during this,... I'm asking this because on one system with lots of snapshots(snapvault) this process seems to never finish. In fact, each time a new snap is transfered the counter restarts from the beginning but has not enough time to finish until a new snapshot is created. Consequence : in snap status I see more and more snapshots with (xxx/xxx remaingint) on the right. Thanks in advance for your answers! -- View this message in context: http://www.nabble.com/Snap-status-and-wafl-scan-tf2232129.html#a6187160 Sent from the Network Appliance - Toasters forum at Nabble.com.

3 2

RE: Snap status and wafl scan
by Glenn Walker 07 Sep '06

07 Sep '06

It may have something to do with the fact that SnapVault 'stitches' snapshots together on the back end - by that I mean that there is a coalescing function that puts the multitude of individually transferred snapshots together once they are all on the destination system and it makes it one single destination snapshot. Glenn -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of ec0li Sent: Thursday, September 07, 2006 5:26 AM To: toasters(a)mathworks.com Subject: Snap status and wafl scan Hello, When I check snap status, I see next to recent snapshots : (xxxx/xxxx remaining). Also, It seems that this is in fact the wafl scan process "snap create summary update" Does anyone knows what is this process, its purpose, what happens if the system reboots during this,... I'm asking this because on one system with lots of snapshots(snapvault) this process seems to never finish. In fact, each time a new snap is transfered the counter restarts from the beginning but has not enough time to finish until a new snapshot is created. Consequence : in snap status I see more and more snapshots with (xxx/xxx remaingint) on the right. Thanks in advance for your answers! -- View this message in context: http://www.nabble.com/Snap-status-and-wafl-scan-tf2232129.html#a6187160 Sent from the Network Appliance - Toasters forum at Nabble.com.

1 0

RE: NDMP backups, timeouts
by George, Andrew 07 Sep '06

07 Sep '06

Hi We noticed it after upgrading from onTap 6.3 to 6.5 The fix in that case was to turn of ndmpd.offset_map.enable. What happened in (I think it was in 6.4) was ndmpd had a slight functionality change so it does a lot more processing up front. Andrew -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Scott T. Mikusko Sent: Thursday, 7 September 2006 9:20 AM To: toasters(a)mathworks.com Subject: NDMP backups, timeouts I'm running into a recurring issue with backing up my NearStore and a long F760 filer. In the procoess of backing some large filesystems (qtrees), it appears that NDMP is basically timing out and fails. I'm running Veritas NetBackup 6.0 (finally upgrading from 4.5 recently). I think the issue has to do with NDMP starting the dump commands on the filer, which then does a three-way backup to another filer with local tape drives. The debug from ndmp logs shows for a manual full backup of a filer: Sep 05 21:05:57 EDT [ndmpd:58]: Log message: DUMP: creating "/vol/vol0/../snapshot_for_backup.370" snapshot. Sep 05 21:06:05 EDT [ndmpd:58]: Log message: DUMP: Using Partial Volume Dump with Exclude Lists Sep 05 21:07:31 EDT [ndmpd:58]: Log message: DUMP: Date of this level 1 dump: Tue Sep 5 21:05:57 2006. Sep 05 21:07:31 EDT [ndmpd:58]: Log message: DUMP: Date of last level 0 dump: Fri Aug 18 19:20:57 2006. Sep 05 21:07:31 EDT [ndmpd:58]: Log message: DUMP: Dumping /vol/vol0/files3.rt to NDMP connection Sep 05 21:07:31 EDT [ndmpd:58]: Log message: DUMP: mapping (Pass I)[regular files] It never completes the mapping of the files in Pass I. It just sits there. The filesystems I'm backing up are between 250-650GB with a LOT of small files (millions) and many subdirs. Tons of mail and html files. My guess is that the amount of files and dirs are getting large that NDMP can't map them fully and is timing out after 8 hours. It shouldn't take that long to map files for the size of data it's doing; I've seem mapping times be much less for larger sets of data. I know that the filesystem topology can cause NDMP backups to be slow, depending on file sizes, data layout, filer load, network etc, but this is happening on moderately busy filers or bone idle NearStores. Anyone run across lengthy DUMP times? Chewing through NOW and Veritas support site hasn't turned up anything obvious. Just curious if others have run into NDMP/dump issues like this. -Scott "This e-mail and any attachments to it (the "Communication") is, unless otherwise stated, confidential, may contain copyright material and is for the use only of the intended recipient. If you receive the Communication in error, please notify the sender immediately by return e-mail, delete the Communication and the return e-mail, and do not read, copy, retransmit or otherwise deal with it. Any views expressed in the Communication are those of the individual sender only, unless expressly stated to be those of Australia and New Zealand Banking Group Limited ABN 11 005 357 522, or any of its related entities including ANZ National Bank Limited (together "ANZ"). ANZ does not accept liability in connection with the integrity of or errors in the Communication, computer virus, data corruption, interference or delay arising from or in respect of the Communication."

1 0

Re: when to use minra ?
by Eyal Traitel 02 Sep '06

02 Sep '06

By the way - an interesting option could be to use FlexClone instead of breaking the mirror - you clone the mirrored volume, and then you change its minra setting. By the way - if you're not using very large aggregates on the R200 that might your ultimate cure - you wanna have many disks involved in the I/O. Eyal. On 9/1/06, Peter D. Gray <pdg(a)uow.edu.au> wrote: > > On Fri, Sep 01, 2006 at 08:47:01AM -0500, Callaway, Michael wrote: > > Quote: > > "My NDMP tape backups from the R200 are pretty slow. I average > > about 2 MBytes per sec. Note that backups of other volumes > > on the R200 work really well, so there is no problem > > on the R200 or the network." > > > > Your 2MB/sec sampling, when was that taken? Was it at the beginning, > > middle, end of the backup? You -should- be getting a really slow > > through-put rate at the start of a backup, since ndmp basically walks > > through and takes a look at each file before actually backing them up. > > When you have a large volume it will take a while to look at each file, > > it's worse with volumes that have a large number of small files. > > > > Good thought, but no. > > The backup is slow after the metadata dump is finished. > > Peak is slow as well (5 Meg or so). > > Regards, > pdg > > -- > > See mail headers for contact information. > > -- Yours, Eyal.

1 0

RE: when to use minra ?
by Glenn Walker 02 Sep '06

02 Sep '06

Actually - I don't understand why minra would not be possibly enabled on destination but not source. I understand that these are volume-level options, but it would make more sense to me to have an option to override that particular setting on a destination volume due to issues just as you've described here. Smells like a great RFE to me :) Glenn -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Peter D. Gray Sent: Thursday, August 31, 2006 7:53 PM To: toasters(a)mathworks.com Subject: Re: when to use minra ? Thanks for the replies. Some background on this question. I have 2 filers used for email storage (NFS, attached to 2 mirapoint mail stores). They have lots of small files. I am mirroring these 2 filers to an R200 for backup, but I also take tape copies via NDMP every 30 days or so. The volumes contain about 350GB each of actual data, excluding snapshots, and 5 million files. My NDMP tape backups from the R200 are pretty slow. I average about 2 MBytes per sec. Note that backups of other volumes on te R200 work really well, so there is no problem on the R200 or the network. I saw the comment that turning off minra can have a big impact on backup performance so I thought I would try it. But discovered that the setting would have to be changed on the mirror source which is not so good (maybe). It would be nice if that setting could be changed on the mirror target but not the source, but I can see why that not possible. So, I wondered if I could justify setting minra on on the source volumes but would need some way to predict the impact. It sounds like turing on minra is not usually a good idea, so I guess I am stuck with slow NDMP backups. Just out of interest, does anybody understand why minra has an impact on backup speed. I assume the read ahead does not go past end of file but that is the only explanation I can think of as to why it would have an impact? Regards, pdg -- See mail headers for contact information.

2 1

Re: when to use minra ?
by Peter D. Gray 02 Sep '06

02 Sep '06

Thanks for the replies. Some background on this question. I have 2 filers used for email storage (NFS, attached to 2 mirapoint mail stores). They have lots of small files. I am mirroring these 2 filers to an R200 for backup, but I also take tape copies via NDMP every 30 days or so. The volumes contain about 350GB each of actual data, excluding snapshots, and 5 million files. My NDMP tape backups from the R200 are pretty slow. I average about 2 MBytes per sec. Note that backups of other volumes on te R200 work really well, so there is no problem on the R200 or the network. I saw the comment that turning off minra can have a big impact on backup performance so I thought I would try it. But discovered that the setting would have to be changed on the mirror source which is not so good (maybe). It would be nice if that setting could be changed on the mirror target but not the source, but I can see why that not possible. So, I wondered if I could justify setting minra on on the source volumes but would need some way to predict the impact. It sounds like turing on minra is not usually a good idea, so I guess I am stuck with slow NDMP backups. Just out of interest, does anybody understand why minra has an impact on backup speed. I assume the read ahead does not go past end of file but that is the only explanation I can think of as to why it would have an impact? Regards, pdg -- See mail headers for contact information.

5 6

RE: when to use minra ?
by Fair, Robert 01 Sep '06

01 Sep '06

In 7.0+ setting minra-on has less impact than earlier releases for many workloads, and many folks who used to use it automatically no longer need to. (Its not obsolete, just not useful for most workloads) Its worth trying minra-off to see if the results are acceptable on both source/destination with your specific workload. -- Rob -----Original Message----- From: Carl Howell [mailto:chowell@uwf.edu] Sent: Friday, September 01, 2006 12:21 PM To: Blake Golliher Cc: Brian Pascal; toasters(a)mathworks.com Subject: RE: when to use minra ? But, isn't the goal to test whether changing the minra setting effects the backup of the volume in question? I'm not suggesting breaking the mirror every time he needs to run a backup. If changing the minra setting is effective, then an RFE is certainly warranted. If not, he has some other problem. Wouldn't you agree? - Carl -----Original Message----- From: Blake Golliher [mailto:thelastman@gmail.com] Sent: Friday, September 01, 2006 10:57 AM To: Carl Howell Cc: Brian Pascal; toasters(a)mathworks.com Subject: Re: when to use minra ? I think the problem there is that you are breaking the snapmirror. I think we want to avoid that. -Blake On 9/1/06, Carl Howell <chowell(a)uwf.edu> wrote: > Vol options cannot change in snapmirror destinations? > > I'm running 7.0.4 on our R200 and I can do this with no problems. > > 1. snapmirror break destvol > 2. vol options destvol minra on > > Am I missing something? > > - Carl > > -----Original Message----- > From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] > On Behalf Of Brian Pascal > Sent: Friday, September 01, 2006 12:03 AM > To: Peter D. Gray > Cc: owner-toasters(a)mathworks.com; toasters(a)mathworks.com > Subject: RE: when to use minra ? > > > > > > Only option is to change the minra on the source since vol options > cannot change in snapmirror destinations. > If you really worried changing minra in a peak time, you can try > changing minra in a non peak time and let the change snapmirror to > destination volume in R200. Then try taking backup's off from that > volume. (just to > isolate) > > Seems like minra is not the only cause for your case. can you provide > more details on your appliances? Model's, disk size's, aggregates & > volume sizes in both source & destination. Also the details of your > tape library attached to R200. ( believe it's directly attached to > R200) > > > > Brian. > > > > > > > "Carl Howell" > > <chowell(a)uwf.edu> To: "Peter D. Gray" > <pdg(a)uow.edu.au>, <toasters(a)mathworks.com> > Sent by: cc: > > owner-toasters@ma Subject: RE: when to use > minra ? > thworks.com > > > > > > 09/01/2006 08:28 > > AM > > > > > > > > > Why not break the mirror change the minra setting and give it a go? You > should know pretty quickly if it makes any difference. Our backup config > is > the same as yours, and we have a couple of volumes with millions of > files that are also slow to backup. > > - Carl > ________________________________ > > From: owner-toasters(a)mathworks.com on behalf of Peter D. Gray > Sent: Thu 8/31/2006 6:53 PM > To: toasters(a)mathworks.com > Subject: Re: when to use minra ? > > > > Thanks for the replies. Some background on this question. > > I have 2 filers used for email storage (NFS, attached to 2 mirapoint > mail stores). They have lots of small files. > I am mirroring these 2 filers to an R200 for backup, but I also take > tape copies via NDMP every 30 days or so. The volumes contain about > 350GB each of actual data, excluding snapshots, and 5 million files. > > My NDMP tape backups from the R200 are pretty slow. I average about 2 > MBytes per sec. Note that backups of other volumes on te R200 work > really well, so there is no problem on the R200 or the network. > > I saw the comment that turning off minra can have a big impact on > backup performance so I thought I would try it. > But discovered that the setting would have to be changed on the mirror > source which is not so good (maybe). > > It would be nice if that setting could be changed on the mirror target > but not the source, but I can see why that not possible. > > So, I wondered if I could justify setting minra on on the source > volumes but would need some way to predict the impact. > > It sounds like turing on minra is not usually a good idea, so I guess > I am stuck with slow NDMP backups. > > Just out of interest, does anybody understand why minra has an impact > on backup speed. I assume the read ahead does not go past end of file > but that is the only explanation I can think of as to why it would > have an impact? > > Regards, > pdg > > -- > > See mail headers for contact information. > > > > > This mail has been scanned by InterScan-MSS/kbsl > > > > > This Mail Has Been Scanned For Virus By Scanmail For Lotus Notes > Enterprise/Kbsl > >

1 0

RE: when to use minra ?
by Carl Howell 01 Sep '06

01 Sep '06

But, isn't the goal to test whether changing the minra setting effects the backup of the volume in question? I'm not suggesting breaking the mirror every time he needs to run a backup. If changing the minra setting is effective, then an RFE is certainly warranted. If not, he has some other problem. Wouldn't you agree? - Carl -----Original Message----- From: Blake Golliher [mailto:thelastman@gmail.com] Sent: Friday, September 01, 2006 10:57 AM To: Carl Howell Cc: Brian Pascal; toasters(a)mathworks.com Subject: Re: when to use minra ? I think the problem there is that you are breaking the snapmirror. I think we want to avoid that. -Blake On 9/1/06, Carl Howell <chowell(a)uwf.edu> wrote: > Vol options cannot change in snapmirror destinations? > > I'm running 7.0.4 on our R200 and I can do this with no problems. > > 1. snapmirror break destvol > 2. vol options destvol minra on > > Am I missing something? > > - Carl > > -----Original Message----- > From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] > On Behalf Of Brian Pascal > Sent: Friday, September 01, 2006 12:03 AM > To: Peter D. Gray > Cc: owner-toasters(a)mathworks.com; toasters(a)mathworks.com > Subject: RE: when to use minra ? > > > > > > Only option is to change the minra on the source since vol options > cannot > change in snapmirror destinations. > If you really worried changing minra in a peak time, you can try > changing > minra in a non peak time and let the change snapmirror to destination > volume in R200. Then try taking backup's off from that volume. (just to > isolate) > > Seems like minra is not the only cause for your case. can you provide > more > details on your appliances? Model's, disk size's, aggregates & volume > sizes > in both source & destination. Also the details of your tape library > attached to R200. ( believe it's directly attached to R200) > > > > Brian. > > > > > > > "Carl Howell" > > <chowell(a)uwf.edu> To: "Peter D. Gray" > <pdg(a)uow.edu.au>, <toasters(a)mathworks.com> > Sent by: cc: > > owner-toasters@ma Subject: RE: when to use > minra ? > thworks.com > > > > > > 09/01/2006 08:28 > > AM > > > > > > > > > Why not break the mirror change the minra setting and give it a go? You > should know pretty quickly if it makes any difference. Our backup config > is > the same as yours, and we have a couple of volumes with millions of > files > that are also slow to backup. > > - Carl > ________________________________ > > From: owner-toasters(a)mathworks.com on behalf of Peter D. Gray > Sent: Thu 8/31/2006 6:53 PM > To: toasters(a)mathworks.com > Subject: Re: when to use minra ? > > > > Thanks for the replies. Some background on this question. > > I have 2 filers used for email storage (NFS, attached to 2 > mirapoint mail stores). They have lots of small files. > I am mirroring these 2 filers to an R200 for backup, but > I also take tape copies via NDMP every 30 days or so. The volumes > contain about 350GB each of actual data, excluding snapshots, > and 5 million files. > > My NDMP tape backups from the R200 are pretty slow. I average > about 2 MBytes per sec. Note that backups of other volumes > on te R200 work really well, so there is no problem > on the R200 or the network. > > I saw the comment that turning off minra can have a big > impact on backup performance so I thought I would try it. > But discovered that the setting would have to be changed on the > mirror source which is not so good (maybe). > > It would be nice if that setting could be changed on the > mirror target but not the source, but I can see why that not > possible. > > So, I wondered if I could justify setting minra on on the > source volumes but would need some way to predict the impact. > > It sounds like turing on minra is not usually a good idea, so I guess > I am stuck with slow NDMP backups. > > Just out of interest, does anybody understand why minra has an > impact on backup speed. I assume the read ahead does not go > past end of file but that is the only explanation I can > think of as to why it would have an impact? > > Regards, > pdg > > -- > > See mail headers for contact information. > > > > > This mail has been scanned by InterScan-MSS/kbsl > > > > > This Mail Has Been Scanned For Virus By Scanmail For Lotus Notes > Enterprise/Kbsl > >

1 0

RE: when to use minra ?
by Carl Howell 01 Sep '06

01 Sep '06

Vol options cannot change in snapmirror destinations? I'm running 7.0.4 on our R200 and I can do this with no problems. 1. snapmirror break destvol 2. vol options destvol minra on Am I missing something? - Carl -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Brian Pascal Sent: Friday, September 01, 2006 12:03 AM To: Peter D. Gray Cc: owner-toasters(a)mathworks.com; toasters(a)mathworks.com Subject: RE: when to use minra ? Only option is to change the minra on the source since vol options cannot change in snapmirror destinations. If you really worried changing minra in a peak time, you can try changing minra in a non peak time and let the change snapmirror to destination volume in R200. Then try taking backup's off from that volume. (just to isolate) Seems like minra is not the only cause for your case. can you provide more details on your appliances? Model's, disk size's, aggregates & volume sizes in both source & destination. Also the details of your tape library attached to R200. ( believe it's directly attached to R200) Brian. "Carl Howell" <chowell(a)uwf.edu> To: "Peter D. Gray" <pdg(a)uow.edu.au>, <toasters(a)mathworks.com> Sent by: cc: owner-toasters@ma Subject: RE: when to use minra ? thworks.com 09/01/2006 08:28 AM Why not break the mirror change the minra setting and give it a go? You should know pretty quickly if it makes any difference. Our backup config is the same as yours, and we have a couple of volumes with millions of files that are also slow to backup. - Carl ________________________________ From: owner-toasters(a)mathworks.com on behalf of Peter D. Gray Sent: Thu 8/31/2006 6:53 PM To: toasters(a)mathworks.com Subject: Re: when to use minra ? Thanks for the replies. Some background on this question. I have 2 filers used for email storage (NFS, attached to 2 mirapoint mail stores). They have lots of small files. I am mirroring these 2 filers to an R200 for backup, but I also take tape copies via NDMP every 30 days or so. The volumes contain about 350GB each of actual data, excluding snapshots, and 5 million files. My NDMP tape backups from the R200 are pretty slow. I average about 2 MBytes per sec. Note that backups of other volumes on te R200 work really well, so there is no problem on the R200 or the network. I saw the comment that turning off minra can have a big impact on backup performance so I thought I would try it. But discovered that the setting would have to be changed on the mirror source which is not so good (maybe). It would be nice if that setting could be changed on the mirror target but not the source, but I can see why that not possible. So, I wondered if I could justify setting minra on on the source volumes but would need some way to predict the impact. It sounds like turing on minra is not usually a good idea, so I guess I am stuck with slow NDMP backups. Just out of interest, does anybody understand why minra has an impact on backup speed. I assume the read ahead does not go past end of file but that is the only explanation I can think of as to why it would have an impact? Regards, pdg -- See mail headers for contact information. This mail has been scanned by InterScan-MSS/kbsl This Mail Has Been Scanned For Virus By Scanmail For Lotus Notes Enterprise/Kbsl

2 1

RE: Active Directory security on CIFS volumes
by Glenn Dekhayser 01 Sep '06

01 Sep '06

Check the Qtee security type in Volumes-->Qtrees. Make sure it's set to either Windows or Mixed. Glenn (the other one) ________________________________ From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Brent Ellis Sent: Friday, September 01, 2006 11:00 AM To: toasters(a)mathworks.com Subject: Active Directory security on CIFS volumes I recently reinitialized one of our filers and I thought I had everything configured until I started testing CIFS shares. When I create a folder and try to change the permissions, the security tab is not available in windows. The other filer in the cluster allows me to change the permissions on individual folders. I thought I had gone through all the cifs options but I can't seem to change that behavior. Any ideas? Brent Ellis Boston University interi(a)bu.edu

1 0

Active Directory security on CIFS volumes
by Brent Ellis 01 Sep '06

01 Sep '06

I recently reinitialized one of our filers and I thought I had everything configured until I started testing CIFS shares. When I create a folder and try to change the permissions, the security tab is not available in windows. The other filer in the cluster allows me to change the permissions on individual folders. I thought I had gone through all the cifs options but I can't seem to change that behavior. Any ideas? Brent Ellis Boston University interi(a)bu.edu

4 4

RE: when to use minra ?
by Callaway, Michael 01 Sep '06

01 Sep '06

Quote: "My NDMP tape backups from the R200 are pretty slow. I average about 2 MBytes per sec. Note that backups of other volumes on the R200 work really well, so there is no problem on the R200 or the network." Your 2MB/sec sampling, when was that taken? Was it at the beginning, middle, end of the backup? You -should- be getting a really slow through-put rate at the start of a backup, since ndmp basically walks through and takes a look at each file before actually backing them up. When you have a large volume it will take a while to look at each file, it's worse with volumes that have a large number of small files. We had an 8TB 560ish million file vol that would take forever. It seemed to just sit there practically idle, before it would take off and begin backing up. This was too impractical for such a long time for a single backup, so we changed up our future growth area, and now do a snapmirror each qtree to a single volume and then ndmp that out. It still takes time, but it allows us to get a clean backup of separate areas each within their own timeframe. Anyway, my point being, are you seeing slow movement in just the beginning, are you competing backups, and have you captured peak performance? Also, would it be possible to change the way you store your data, or back up in sections? I know it's not fixing the issue, it's just working around it, but it could either help in the meantime while you look at minra, or possibly become a smoother process for you. -- *------------------- Michael Callaway Server Administrator NTTA/IT 214-461-2033 -------------------- -----Original Message----- From: owner-toasters(a)mathworks.com [mailto:owner-toasters@mathworks.com] On Behalf Of Peter D. Gray Sent: Thursday, August 31, 2006 6:53 PM To: toasters(a)mathworks.com Subject: Re: when to use minra ? Thanks for the replies. Some background on this question. I have 2 filers used for email storage (NFS, attached to 2 mirapoint mail stores). They have lots of small files. I am mirroring these 2 filers to an R200 for backup, but I also take tape copies via NDMP every 30 days or so. The volumes contain about 350GB each of actual data, excluding snapshots, and 5 million files. My NDMP tape backups from the R200 are pretty slow. I average about 2 MBytes per sec. Note that backups of other volumes on te R200 work really well, so there is no problem on the R200 or the network. I saw the comment that turning off minra can have a big impact on backup performance so I thought I would try it. But discovered that the setting would have to be changed on the mirror source which is not so good (maybe). It would be nice if that setting could be changed on the mirror target but not the source, but I can see why that not possible. So, I wondered if I could justify setting minra on on the source volumes but would need some way to predict the impact. It sounds like turing on minra is not usually a good idea, so I guess I am stuck with slow NDMP backups. Just out of interest, does anybody understand why minra has an impact on backup speed. I assume the read ahead does not go past end of file but that is the only explanation I can think of as to why it would have an impact? Regards, pdg -- See mail headers for contact information.

1 0

get rid of a false imported aggregate
by Alexander Schalek 01 Sep '06

01 Sep '06

Hi all, Is there any way except reboot the storage appliance to get rid of a failed aggregate in sysconfig-r? During maintenance work I attached a shelve to the wrong filer b (in a 3020er cluster configuration). I corrected my mistake and now I have a failed aggregate on filer b. filer a has the aggregate imported and online, filer b can see the partner aggregate (disks) but still has the failed aggregate in sysconfig -r. thanks for your help. alex

2 1

Yahoo! News Story - Extreme Makeover: "Star Trek" Edition - Yahoo! News
by Peter D. Gray 01 Sep '06

01 Sep '06

Peter D. Gray (pdg(a)uow.edu.au) has sent you a news article. (Email address has not been verified.) ------------------------------------------------------------ Personal message: Extreme Makeover: "Star Trek" Edition - Yahoo! News http://news.yahoo.com/s/eo/20060901/en_tv_eo/19912 ============================================================ Yahoo! News http://news.yahoo.com/

1 0

Yahoo! News Photo - Yahoo! News Photo
by Peter D. Gray 01 Sep '06

01 Sep '06

Peter D. Gray (pdg(a)uow.edu.au) has sent you a news photo. (Email address has not been verified.) ------------------------------------------------------------ Personal message: Yahoo! News Photo http://news.yahoo.com/photo/060831/photos_fashionwire/574cee2ba0024aefcfce1… ============================================================ Yahoo! News http://news.yahoo.com/

1 0

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

toasters