Re: [OT] best tape backup system?
On Tue, 22 Feb 2005, Louis-David Mitterrand wrote: Is there a better solution out there? We using an 59P6744 200/400GB LTO Full-High Tape Drive with great resuslt. It made a full 180G backup within 100 minutes. We are using the tob script for full and differential backups. Bye, -=Lajbi= LAJBER Zoltan Szent Istvan Egyetem, Informatika Hivatal engineer: a mechanism for converting caffeine into designs. - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Can someone explain this?
Why is it reporting Raid Devices: 8 and actuall Devices: 5 ? Why Failed Devices: -1 (hehe do i have a credit of 1 failed device) ;) Is there anyway to get the figures stright again? mdadm -D /dev/md0 /dev/md0: Version : 00.90.01 Creation Time : Wed Jan 26 18:24:22 2005 Raid Level : raid5 Array Size : 1393991424 (1329.41 GiB 1427.45 GB) Device Size : 199141632 (189.92 GiB 203.92 GB) Raid Devices : 8 Total Devices : 5 Preferred Minor : 0 Persistence : Superblock is persistent Update Time : Mon Feb 21 20:52:24 2005 State : dirty, degraded, recovering Active Devices : 7 Working Devices : 7 Failed Devices : -1 Spare Devices : 1 Layout : left-symmetric Chunk Size : 64K Rebuild Status : 82% complete UUID : 129eed1b:e38f8cbf:dcb31e59:e6532935 Events : 0.10797242 Number Major Minor RaidDevice State 0 2210 active sync /dev/ide/host0/bus1/target0/lun0/part1 1 22 651 active sync /dev/ide/host0/bus1/target1/lun0/part1 2 3312 active sync /dev/ide/host2/bus0/target0/lun0/part1 3 33 653 active sync /dev/ide/host2/bus0/target1/lun0/part1 4 3414 active sync /dev/ide/host2/bus1/target0/lun0/part1 5 005 faulty removed 6 5716 active sync /dev/ide/host4/bus1/target0/lun0/part1 7 57 657 active sync /dev/ide/host4/bus1/target1/lun0/part1 8 5618 spare rebuilding /dev/ide/host4/bus0/target0/lun0/part1 - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
89.95$ = Microsoft Windows XP Pro + Microsoft Office XP Pro
http://54471.aflnkbjgn.com/?VErurGVzuZwObpV52007 - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: *terrible* direct-write performance with raid5
Michael Tokarev [EMAIL PROTECTED] wrote: When debugging some other problem, I noticied that direct-io (O_DIRECT) write speed on a software raid5 And normal write speed (over 10 times the size of ram)? is terrible slow. Here's a small table just to show the idea (not numbers by itself as they vary from system to system but how they relate to each other). I measured plain single-drive performance (sdX below), performance of a raid5 array composed from 5 sdX drives, and ext3 filesystem (the file on the filesystem was pre-created And ext2? You will be enormously hampered by using a journalling file system, especially with journal on the same system as the one you are testing! At least put the journal elsewhere - and preferably leave it off. during tests). Speed measurements performed with 8Kbyte buffer aka write(fd, buf, 8192*1024), units a Mb/sec. write read sdX 44.9 45.5 md1.7* 31.3 fs on md0.7* 26.3 fs on sdX 44.7 45.3 Absolute winner is a filesystem on top of a raid5 array: I'm afraid there are too many influences to say much from it overall. The legitimate (i.e. controlled) experiment there is between sdX and md (over sdx), with o_direct both times. For reference I personally would like to see the speed withut o_direct on those two. And the size/ram of the transfer - you want to run over ten times size of ram when you run without o_direct. Then I would like to see a similar comparison made over hdX instead of sdX. You can forget the fs-based tests for the moment, in other words. You already have plenty there to explain in the sdX/md comparison. And to explain it I would like to see sdX replaced with hdX. A time-wise graph of the instantaneous speed to disk would probably also be instructive, but I guess you can't get that! I would guess that you are seeing the results of one read and write to two disks happening in sequence and not happening with any great urgency. Are the writes sent to each of the mirror targets from raid without going through VMS too? I'd suspect that - surely the requests are just queued as normal by raid5 via the block device system. I don't think the o_direct taint persists on the requests - surely it only exists on the file/inode used for access. Suppose the mirrored requests are NOT done directly - then I guess we are seeing an interaction with the VMS, where priority inversion causes the high-priority requests to the md device to wait on the fulfilment of low priority requests to the sdX devices below them. The sdX devices requests may not ever get treated until the buffers in question age sufficiently, or until the kernel finds time for them. When is that? Well, the kernel won't let your process run .. hmm. I'd suspect the raid code should be deliberately signalling the kernel to run the request_fn of the mirror devices more often. Comments anyone? ;) Random guesses above. Purely without data, of course. Peter - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
RE: [OT] best tape backup system?
In my case, I am sure I had the termination correct. I normally don't set the SCSI card to auto termination, since I don't trust it! But I did try every valid permutation. On non-Linux systems, I have never had problems mixing disks with tapes or anything else. And since I swapped terminators, cables and cards, the only hardware that could be at fault would be the tape or disk drives. But, they are still working to this day, just on separate SCSI cards. If I recall, I once had a Unix system where someone created about 128 Gig of sparse files. I backed up the files to a DDS-3 tape (12-24 Gig). The backup took more than 24 hours. That is better than 10 to 1. :) So, yes, giving the correct data, 2 to 1 can be done, but not in the real world. IMO. Guy -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Gordon Henderson Sent: Tuesday, February 22, 2005 12:40 PM To: Guy Cc: linux-raid@vger.kernel.org Subject: RE: [OT] best tape backup system? On Tue, 22 Feb 2005, Guy wrote: I have NOT been able to share a SCSI cable/card with disks and a tape drive. Tried for days. I would get disk errors, or timeouts. I corrected the problem by putting the tape drive on a dedicated SCSI bus/card. I don't recall the details of the failures, since it has been well over a year now. But I do know I only had problems while using the tape drive and the disks at the same time. Like, during a backup. The disks were happy while the tape drive was idle. I also swapped parts, it had no effect. I assumed Linux does not like to share SCSI. Linux is fine with multiple devices on SCSI - scanners, drives, tapes. I haven't had problems with these combinations. The usual culprit is termination. Either not enough or too much! You need termination at both ends of the bus - one end is usually (but not always) the SCSI card, and all cards I've used in the past few years have had the ability to turn termination on or off (You turn it off if the card is not at one end of the bus) Some tape drives I've used also have the ability to turn termination on, or not - sometimes you set the drive number to one range for termination, another range for not (eg. 0-7 no termination, 8-15 active termination). It's a real PITA to get right at times. I have not had problems with 2 or more tape drives on the same SCSI bus/card. I think I had 3 at one time. Amanda... The last time I checked, incremental backups require a different tape each time. You will need a lot of tapes. If a full backup fits on a single tape, just do a full backup each night. I need many tapes to do a full backup, but a single tape can hold many daily incremental backups. I deemed Amanda to wasteful for me. I use home made scripts and cpio and do a full backup about once per month, and do nightly incremental backups to 1 tape. Once the daily tape is full I do another full backup. Works very well for me. But, restoring is a pain, but that is very rare for me, only once in 1+ years, so far. My scripts seek the tape as required, so I can eject the tape if needed, as long as I put it back in time for the nightly backup. Yup. One tape per backup run with Amanda. Way back when tape capacity exceeded disk capacity, I used Amanda to backup many machines to one tape every night - that was its original strength, the ability to backup many machines over the LAN to one tape drive. These days with disk capacities generally exceeding tapes, it's not like that anymore (although I still have some smaller servers backed up to a non-local tape drive for reasons of economy) If anyone gets 1.3 to 1 compression, that is real good IMO. The last time I checked, I got about 1.1 to 1. What a marketing scam Compression is always a weirdism - I'm guessing your data is music, video or pictures - stuff thats already compressed which won't compress twice. Typical officy type data would be stuff that generally compresses well - text documents, spreadsheets, program source code ( compiled binaries) etc. Amanda can compress data and in some cases better than the tape drives (using gzip, etc), but it's slow and makes restoring more interesting... I think Sony are claiming 2.3:1 with their new tapes - well, they'll get that with text-files, but nothing near that with anything else! Gordon - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: *terrible* direct-write performance with raid5
Peter T. Breuer wrote: Michael Tokarev [EMAIL PROTECTED] wrote: When debugging some other problem, I noticied that direct-io (O_DIRECT) write speed on a software raid5 And normal write speed (over 10 times the size of ram)? There's no such term as normal write speed in this context in my dictionary, because there are just too many factors influencing the speed of non-direct I/O operations (I/O scheduler aka elevator is the main factor I guess). More, when going over the buffer cache, cache trashing is plays significant role for the whole system (eg, when just copying a large amount of data with cp, system becomes quite unresponsible die to cache trashing while the whole stuff it is copying should not be cached in the first place, for this task anyway). Also, I don't think linux elevator is optimized for this task (accessing large amounts of data). I come across this issue when debugging very slow database (oracle10 in this case) which tries to do direct I/O where possible because it knows better when/how to cache data. If I turn on vfs/block cache here, system becomes much slower (under normal conditions anyway, not counting this md slowness) (ok ok, i know it isn't a good idea to place datbase files on raid5... or wasn't some time ago when raid5 checksumming was the bottleneck anyway... but that's a different story). More to the point seems to be the same direct-io but in larger chunks - eg 1Mb or more instead of 8Kb buffer. And this indeed makes alot of difference, the numbers looks much more nice. is terrible slow. Here's a small table just to show the idea (not numbers by itself as they vary from system to system but how they relate to each other). I measured plain single-drive performance (sdX below), performance of a raid5 array composed from 5 sdX drives, and ext3 filesystem (the file on the filesystem was pre-created And ext2? You will be enormously hampered by using a journalling file system, especially with journal on the same system as the one you are testing! At least put the journal elsewhere - and preferably leave it off. This whole issue has exactly nothing to do with journal. I don't mount the fs with data=journal option, and the file I'm writing to is preallocated first (i create the file of a given size when measure re-writing speed). In this case, data never touches ext3 journal. during tests). Speed measurements performed with 8Kbyte buffer aka write(fd, buf, 8192*1024), units a Mb/sec. [] Absolute winner is a filesystem on top of a raid5 array: I'm afraid there are too many influences to say much from it overall. The legitimate (i.e. controlled) experiment there is between sdX and md (over sdx), with o_direct both times. For reference I personally would like to see the speed withut o_direct on those two. And the size/ram of the transfer - you want to run over ten times size of ram when you run without o_direct. I/O speed without O_DIRECT is very close to 44 Mb/sec for sdX (it's the spid of the drives it seems), and md performs at about 80 Mb/sec. That numbers are very close to the case with O_DIRECT and large block size (eg 1Mb). There's much more to the block size really. I just used 8Kb block because I have real problem with performance of our database server (we're trying oracle10 and it performs very badly, and now i don't understand whenever the machine has always been like that (the numbers above) but we never noticied with different usage pattern of previous oracle releases, or something else changed... Then I would like to see a similar comparison made over hdX instead of sdX. Sorry no IDE drives here, and i don't see the point in trying them anyway. You can forget the fs-based tests for the moment, in other words. You already have plenty there to explain in the sdX/md comparison. And to explain it I would like to see sdX replaced with hdX. A time-wise graph of the instantaneous speed to disk would probably also be instructive, but I guess you can't get that! I would guess that you are seeing the results of one read and write to two disks happening in sequence and not happening with any great urgency. Are the writes sent to each of the mirror targets from raid Hmm point. without going through VMS too? I'd suspect that - surely the requests are just queued as normal by raid5 via the block device system. I don't think the o_direct taint persists on the requests - surely it only exists on the file/inode used for access. Well, O_DIRECT performs very-very similar to O_SYNC here (both cases -- with and without a filesystem involved) in terms of speed. I don't care much now whenever it relly performs direct I/O (from userspace buffer directly to controller), esp. since it can't work exactly this way with raid5 implemented in software (checksums must be written too). I just don't want to see unnecessary cache trashing and do want to know about I/O errors immediately. Suppose the mirrored requests are NOT done directly - then I guess we are seeing an interaction with the
Re: [OT] best tape backup system?
Guy wrote: I have NOT been able to share a SCSI cable/card with disks and a tape drive. Tried for days. I would get disk errors, or timeouts. I corrected the problem by putting the tape drive on a dedicated SCSI bus/card. I don't recall the details of the failures, since it has been well over a year now. But I do know I only had problems while using the tape drive and the disks at the same time. Like, during a backup. The disks were happy while the tape drive was idle. I also swapped parts, it had no effect. I assumed Linux does not like to share SCSI. There may be various reasons for this, and i think all of them are related to hardware only, esp. to the SCSI bus. Eg, some modern disk drives don't understand SE mode anymore, so can't be plugged into an scsi bus with at least one SE device. On the other hand, it was qutie common for esp. old tapes to NOT understand LVD mode. Obviously you can't mix that sort of devices on the same bus. I guess the bus where you drives are is in LVD mode while tapes works in SE mode... BTW, I've seen quite modern tape drive with an old scsi connector (I don't even remember anymore how it's called -- the one which looks pretty like an IDE connector but a bit wider and with more (80?) pins; with max speed of 20Mb/sec aka 10MHz or so) inside the case and with an adaptor for current 68pin scsi connector standard. Obviously it works in SE mode only (that bus worked in that mode only). Don't remember which vendor that drive was... /mjt - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [OT] best tape backup system?
On Tue, 22 Feb 2005, Louis-David Mitterrand wrote: I am considering getting a Sony SAIT 3 with 500G/1TB tapes, which seems like a nice solution for backuping a whole server on a single tape. Has anyone used that hardware and can comment on its performance, linux-compatibility or otherwise? Is there a better solution out there? Better depends on what you want/need/can afford. Last time I was tape shopping, I thought this would be a good compromise on the need/can afford: Exabyte VXA-2 Packetloader 1x10 Native tape capacity is 800gb. The only downside is, no magazine...it stores the tapes in an internal carosel accessed from the front, one position at a time. For a bit more $, they have magazine based tape library systems with VXA-2 drives. Anyone used these? I'd still like one. -- Jon Lewis | I route Senior Network Engineer | therefore you are Atlantic Net| _ http://www.lewis.org/~jlewis/pgp for PGP public key_ - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [OT] best tape backup system?
On Tue, 22 Feb 2005, Jon Lewis wrote: On Tue, 22 Feb 2005, Louis-David Mitterrand wrote: I am considering getting a Sony SAIT 3 with 500G/1TB tapes, which seems like a nice solution for backuping a whole server on a single tape. Has anyone used that hardware and can comment on its performance, linux-compatibility or otherwise? Is there a better solution out there? Better depends on what you want/need/can afford. Last time I was tape shopping, I thought this would be a good compromise on the need/can afford: Exabyte VXA-2 Packetloader 1x10 Native tape capacity is 800gb. The only downside is, no magazine...it stores the tapes in an internal carosel accessed from the front, one position at a time. For a bit more $, they have magazine based tape library systems with VXA-2 drives. for 1TB of storage ... i'd put the data on 4 disks ( raided ) and take the disks and put in nice bubble wrap and nice cushion :-) don't pop the bubbles and yup.. i want the fast restore of 1TB of data too which to me is more important than the $50 or $100 tape costs (plus the tape drive) vs $600 set of 1TB disk c ya alvin i keep wondering why people pay $150K for 1TB brandname tape subsystems .. :-) - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [OT] best tape backup system?
On Tue, 22 Feb 2005, Alvin Oga wrote: Better depends on what you want/need/can afford. Last time I was tape shopping, I thought this would be a good compromise on the need/can afford: Exabyte VXA-2 Packetloader 1x10 Native tape capacity is 800gb. The only downside is, no magazine...it stores the tapes in an internal carosel accessed from the front, one position at a time. For a bit more $, they have magazine based tape library systems with VXA-2 drives. for 1TB of storage ... i'd put the data on 4 disks ( raided ) and take the disks and put in nice bubble wrap and nice cushion I should clarify, that's 80GB per tape...so 800GB native assumes you have 10 tapes in the unit. i keep wondering why people pay $150K for 1TB brandname tape subsystems .. I wouldn't pay that much...but I think the common wisdom is that tape is more durable/portable than disks. Once upon a time, it was cheaper than disks too...but that's no longer the case. It's part of why my plan to buy a bunch of Exabyte stuff got shot down and instead we bought P4's with 1TB SATA-RAID5 arrays to use as backup servers. -- Jon Lewis | I route Senior Network Engineer | therefore you are Atlantic Net| _ http://www.lewis.org/~jlewis/pgp for PGP public key_ - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [OT] best tape backup system?
Not sure if it is important to many people, but tapes take a lot less electricity than online disks. --David Dougall On Tue, 22 Feb 2005, Jon Lewis wrote: On Tue, 22 Feb 2005, Alvin Oga wrote: Better depends on what you want/need/can afford. Last time I was tape shopping, I thought this would be a good compromise on the need/can afford: Exabyte VXA-2 Packetloader 1x10 Native tape capacity is 800gb. The only downside is, no magazine...it stores the tapes in an internal carosel accessed from the front, one position at a time. For a bit more $, they have magazine based tape library systems with VXA-2 drives. for 1TB of storage ... i'd put the data on 4 disks ( raided ) and take the disks and put in nice bubble wrap and nice cushion I should clarify, that's 80GB per tape...so 800GB native assumes you have 10 tapes in the unit. i keep wondering why people pay $150K for 1TB brandname tape subsystems .. I wouldn't pay that much...but I think the common wisdom is that tape is more durable/portable than disks. Once upon a time, it was cheaper than disks too...but that's no longer the case. It's part of why my plan to buy a bunch of Exabyte stuff got shot down and instead we bought P4's with 1TB SATA-RAID5 arrays to use as backup servers. -- Jon Lewis | I route Senior Network Engineer | therefore you are Atlantic Net| _ http://www.lewis.org/~jlewis/pgp for PGP public key_ - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [OT] best tape backup system?
hi ya On Tue, 22 Feb 2005, Jon Lewis wrote: I should clarify, that's 80GB per tape...so 800GB native assumes you have 10 tapes in the unit. yup...seen those puppies too ... too much headache for me i keep wondering why people pay $150K for 1TB brandname tape subsystems .. I wouldn't pay that much... the odd part is people do :-) just because its name branded enterprise blah blah but I think the common wisdom is that tape is more durable/portable than disks. Once upon a time, it was cheaper than disks too...but that's no longer the case. yup ... and disks have become more reliable tapes still ahve the same problem they always had ... the heads need to be cleaned, and if the drive unit goes in for repair, you have no alternative options - i usually get those calls, from panicky folks, which there is no solution other than buy 2 tape drives or buy disk-based backups in addition to tapes if they like tapes both tape and disk backups have its purposes and reasons It's part of why my plan to buy a bunch of Exabyte stuff got shot down and instead we bought P4's with 1TB SATA-RAID5 arrays to use as backup servers. i'd buy 2 disk based backups, because one has to backup the backup server :-) - disks is cheap, compared to loss of 1TB of corp data or 10TB of corp data is more along the lines of fun when things become interesting :-) -- david Not sure if it is important to many people, but tapes take a lot david less electricity than online disks. i doubt that it'd be a factor in buying tapes vs disks, but one never knows :-0 they should be more worried about magnets and phones next to the tape or tapes under the old fashion crt's - you'd be surprise what one finds, when one goes into server rooms which i guess is the fun part, fixing their problems c ya alvin - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Forcing a more random uuid (random seed bug)
Followup to: [EMAIL PROTECTED] By author:Niccolo Rigacci [EMAIL PROTECTED] In newsgroup: linux.dev.raid I get /dev/md5, /dev/md6, /dev/md7 and /dev/md8 all with the same UUID! It seems that there is a bug in mdadm: when generating the UUID for a volume, the random() function is called, but the random sequence is never initialized. The result is that every volume created with mdadm has an uuid of: 6b8b4567:327b23c6:643c9869:66334873 See also Debian bug 292784 at http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=292784 I fixed the problem adding the following patch to mdadm.c, but please bear in mind that I'm totally unaware of mdadm code and quite naive in C programming: Please don't use (s)random at all, except as a possible fallback to /dev/(u)random. -hpa - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: Forcing a more random uuid (random seed bug)
Followup to: [EMAIL PROTECTED] By author:[EMAIL PROTECTED] In newsgroup: linux.dev.raid +if ((my_fd = open(/dev/random, O_RDONLY)) != -1) { Please use /dev/urandom for such applications. /dev/random is the highest-quality generator, but will block if entropy isn't available. /dev/urandom provides the best available, immediately, which is what this application wants. Not 100% clear; the best would be to make it configurable. Either way you must not use read() in the way described. Short reads happen, even with /dev/urandom. Also, this will only produce 2^32 possible UUIDs, since that's the size of the seed. Meaning that after you've generated 2^16 of them, the chances are excellent that they're not UU any more. You might just want to get all 128 (minus epsilon) bits from /dev/urandom directly. You *do* want to get all bits from /dev/urandom directly. -hpa - To unsubscribe from this list: send the line unsubscribe linux-raid in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html