Thread: 2 questions re RAID
No responses to my earlier post, I'm assuming because OS X experience is rather thin in this group ;-) So a couple of morespecific questions: 1) Is my impression correct that given a choice between Areca & Highpoint, it's a no-brainer to go with Areca? 2) I understand why RAID 5 is not generally recommended for good db performance. But if the database is not huge (10-20GB),and the server has enough RAM to keep most all of the db cached, and the RAID uses (battery-backed) write-backcache, is it sill really an issue? -- Scott Ribe scott_ribe@elevated-dev.com http://www.elevated-dev.com/ (303) 722-0567 voice
On Fri, Jun 17, 2011 at 11:02 AM, Scott Ribe <scott_ribe@elevated-dev.com> wrote: > No responses to my earlier post, I'm assuming because OS X experience is rather thin in this group ;-) So a couple of morespecific questions: > > 1) Is my impression correct that given a choice between Areca & Highpoint, it's a no-brainer to go with Areca? Generally, yes, but the model of the card is more important than the maker. I.e. an Areca 1880 or 1680 is a fantastic performer. But the older 1120 series aren't gonna set the world on fire or anything. Pluses for the Arecas I've used: Out Of Band monitoring. Heck, I've updated the firmware on them from 1000 miles away. fast in RAID-10. Lots of HW controllers (I'm looking at you, LSI) perform poorly with layered RAID. They all use the same simple standard battery backed unit, unlike some manufacturers that glue them onto the DIMM so you have to buy a new memory module to replace your BBU (again, I'm looking at you LSI) Great UI via the web and / or the BIOS. Again, some other RAID setup utils are not so nice (and again, I'm looking at you, LSI) > 2) I understand why RAID 5 is not generally recommended for good db performance. But if the database is not huge (10-20GB),and the server has enough RAM to keep most all of the db cached, and the RAID uses (battery-backed) write-backcache, is it sill really an issue? The problem with RAID-5 is crappy write performance. Being big or small won't change that. Plus if the db is small why use RAID-5?
Thanks much for the specific info on Areca RAID cards. Very helpful. On Jun 17, 2011, at 11:20 AM, Scott Marlowe wrote: > The problem with RAID-5 is crappy write performance. Being big or > small won't change that. Plus if the db is small why use RAID-5? It's small enough that there's some other things going on at the same small server with 4 disk bays ;-) My thinking was thatwrite-back cache might mitigate the poor write performance enough to not be noticed. This db doesn't generally get bigbatch updates anyway, it's mostly a constant stream of small updates coming in and I have a hard time imagining 256MBof cache filling up very often. (I have at least a fuzzy understanding of how WAL segments affect the write load.) RAID-1 & RAID-10 are not ruled out, I'm just exploring options. And I'm not actually wanting to use RAID 5; it's RAID 6 thatI'm considering... -- Scott Ribe scott_ribe@elevated-dev.com http://www.elevated-dev.com/ (303) 722-0567 voice
On Jun 17, 2011, at 11:20 AM, Scott Marlowe wrote: > Generally, yes, but the model of the card is more important than the > maker. I.e. an Areca 1880 or 1680 is a fantastic performer. But the > older 1120 series aren't gonna set the world on fire or anything. And, in further digging, I discover that ATTO ExpressSAS is an option for me. Anyone got comments on these? (I notice thatthey use ultracapacitor/flash to protect cache...) -- Scott Ribe scott_ribe@elevated-dev.com http://www.elevated-dev.com/ (303) 722-0567 voice
On Fri, Jun 17, 2011 at 11:35 AM, Scott Ribe <scott_ribe@elevated-dev.com> wrote: > It's small enough that there's some other things going on at the same small server with 4 disk bays ;-) My thinking wasthat write-back cache might mitigate the poor write performance enough to not be noticed. This db doesn't generally getbig batch updates anyway, it's mostly a constant stream of small updates coming in and I have a hard time imagining 256MBof cache filling up very often. (I have at least a fuzzy understanding of how WAL segments affect the write load.) We run our internal dev server on RAID-6 and it works well enough. Again, like your usage case, it doesn't get beat up too hard, so RAID-6 works fine. I prefer RAID-6 because it doesn't degrade as bad as RAID-5 when a single drive fails, and of course it's still fully redundant with a single drive failure.
On 06/17/2011 01:02 PM, Scott Ribe wrote: > 1) Is my impression correct that given a choice between Areca& Highpoint, it's a no-brainer to go with Areca? > I guess you could call Highpoint a RAID manufacturer, but I wouldn't do so. They've released so many terrible problems over the years that it's hard to take the fact that they may have something reasonable you can buy now (the 43XX cards I think?) seriously. > And, in further digging, I discover that gh is an option for me. Anyone got comments on these? (I notice that they useultracapacitor/flash to protect cache...) Atto is so Mac focused that you're not going to find much experience here, for the same reason you didn't get any response to your original question. Their cards are using the same Intel IO Processor (IOP) hardware as some known capable cards. For example, the ExpressSAS R348 is named that because it has an Intel 348 IOP. That's the same basic processor as on the medium sized Areca boards: http://www.areca.us/products/pcietosas1680series.htm So speed should be reasonable, presuming they didn't make any major errors in board design or firmware. The real thing you need to investigate is whether the write cache setup is done right, and whether monitoring is available in a way you can talk to. What you want is for the card to run in write-back mode normally, degrading to write-through when the battery stops working well. If you don't see that sort of thing clearly documented as available, you really don't want to consider their cards. > 2) I understand why RAID 5 is not generally recommended for good db performance. But if the database is not huge (10-20GB),and the server has enough RAM to keep most all of the db cached, and the RAID uses (battery-backed) write-backcache, is it sill really an issue? > You're basically asking "if I don't write to the database, does the fact that write performance on RAID5 is slow matter?" When asked that way, sure, it's fine. If after applying the write cache to help, your write throughput requirements don't ever exceed what a single disk can provide, than maybe RAID5 will be fine for you. Make sure you keep shared_buffers low though, because you're not going to be able to absorb a heavy checkpoint sync on RAID5. -- Greg Smith 2ndQuadrant US greg@2ndQuadrant.com Baltimore, MD PostgreSQL Training, Services, and 24x7 Support www.2ndQuadrant.us "PostgreSQL 9.0 High Performance": http://www.2ndQuadrant.com/books
On Jun 17, 2011, at 11:23 PM, Greg Smith wrote: > I guess you could call Highpoint a RAID manufacturer, but I wouldn't do so. They've released so many terrible problemsover the years that it's hard to take the fact that they may have something reasonable you can buy now (the 43XXcards I think?) seriously. Ah, I see. So they're on par with Apple's RAID controller instead of being the first step up. > Atto is so Mac focused that you're not going to find much experience here, for the same reason you didn't get any responseto your original question. Their cards are using the same Intel IO Processor (IOP) hardware as some known capablecards. For example, the ExpressSAS R348 is named that because it has an Intel 348 IOP. That's the same basic processoras on the medium sized Areca boards: http://www.areca.us/products/pcietosas1680series.htm So speed should be reasonable,presuming they didn't make any major errors in board design or firmware. Good info. Didn't know about their focus, because the last time I dealt with them was so many years ago they still had asignificant focus on Windows, or so it seemed to me at the time. Focus on Mac says nothing about the firmware on the card,but it should bode well for the driver. > The real thing you need to investigate is whether the write cache setup is done right, and whether monitoring is availablein a way you can talk to. What you want is for the card to run in write-back mode normally, degrading to write-throughwhen the battery stops working well. If you don't see that sort of thing clearly documented as available, youreally don't want to consider their cards. Well, right up front in their marketing materials they make a major point about cache protection, how important it is, howgood it is, using ultracapacitor+flash over batteries (on some of their controllers). So they have awareness & intent;competence and follow-through of course are not assured by marketing materials. (Also they talk about background scanningof drives for defects.) And it looks like they offer all of: GUI setup/monitoring that runs on OS X, command-linesetup/monitoring that runs on OS X, SNMP... > You're basically asking "if I don't write to the database, does the fact that write performance on RAID5 is slow matter?" When asked that way, sure, it's fine. If after applying the write cache to help, your write throughput requirementsdon't ever exceed what a single disk can provide, than maybe RAID5 will be fine for you. Make sure you keepshared_buffers low though, because you're not going to be able to absorb a heavy checkpoint sync on RAID5. Yes, basically I wanted to confirm that's what I was actually asking ;-) The only circumstance under which I could see overflowingthe card's write cache is during migrations. So my choice then really is better performance during rare migrationsvs being able to lose any 2 drives out of 4 (RAID6). Which is OK, since neither choice is really bad--having beenburned by bad disk runs before, I'll probably go for safety. (FYI this is not my only margin for failure. Two geographically-distributedWAL-streaming replicas with low-end RAID1 are the next line of defense. Followed by, god forbidI should ever have to use them, daily dumps.) Thanks for all the info. I guess about all I have remaining to do is sanity-check my beliefs about disk I/O. -- Scott Ribe scott_ribe@elevated-dev.com http://www.elevated-dev.com/ (303) 722-0567 voice
On Fri, Jun 17, 2011 at 1:20 PM, Scott Marlowe <scott.marlowe@gmail.com> wrote: > Pluses for the Arecas I've used: > Out Of Band monitoring. Heck, I've updated the firmware on them from > 1000 miles away. > fast in RAID-10. Lots of HW controllers (I'm looking at you, LSI) > perform poorly with layered RAID. > They all use the same simple standard battery backed unit, unlike some > manufacturers that glue them onto the DIMM so you have to buy a new > memory module to replace your BBU (again, I'm looking at you LSI) > Great UI via the web and / or the BIOS. Again, some other RAID setup > utils are not so nice (and again, I'm looking at you, LSI) > If I could repeat this 1000 times, I would.
On Fri, Jun 17, 2011 at 1:35 PM, Scott Ribe <scott_ribe@elevated-dev.com> wrote: > RAID-1 & RAID-10 are not ruled out, I'm just exploring options. And I'm not actually wanting to use RAID 5; it's RAID 6that I'm considering... You have 4 disk bays and you want RAID-6? How will that improve anything over RAID-10? You will have the same amount of available space, and the writes will be slower. Ok...there is *one* advantage: you can lose any two drives at the same time and still survive, with RAID-10 if you lose the wrong two drives you're hosed. That said, on one of my production DB's, I have a 16-bay raid enclosure and I run RAID-6 + hot spare. It even has an Areca controller made visible to the server as a single drive via fibre channel. Very sweet setup.
On Jun 21, 2011, at 7:49 AM, Vick Khera wrote: > Ok...there is *one* advantage: > you can lose any two drives at the same time and still survive, with > RAID-10 if you lose the wrong two drives you're hosed. Exactly. The performance advantage of RAID-10 over RAID-6 in this sever is, I think, not useful. -- Scott Ribe scott_ribe@elevated-dev.com http://www.elevated-dev.com/ (303) 722-0567 voice