Thread: PERFORMANCE IMPROVEMENT by mapping WAL FILES
Hi all, By mapping the WAL files by each backend in to its address space using "mmap" system call , there will be bigimprovements in performance from the following point of view: 1. Each backend directly writes in to the address space which is obtained by maping the WAL files. this saves the write system call at the end of every transaction which transfres 8k of data from user space to kernel. 2. since everytransaction does not modify all the 8k content of WAL page , so by issuing the fsync , the kernel only transfers only the kernel pages which are modified , which is 4k for linux so fsync time is saved by twice. Any comments ?. Regards jana
> Hi all, > By mapping the WAL files by each backend in to its address > space using "mmap" system call , there will be big > improvements in performance from the following point of view: > 1. Each backend directly writes in to the address > space which is obtained by maping the WAL files. > this saves the write system call at the end of > every transaction which transfres 8k of > data from user space to kernel. > 2. since every transaction does not modify all the 8k > content of WAL page , so by issuing the > fsync , the kernel only transfers only the > kernel pages which are modified , which is 4k for > linux so fsync time is saved by twice. > Any comments ?. This is interesting. We are concerned about using mmap() for all I/O because we could eat up quite a bit of address space for big tables, but WAL seems like an ideal use for mmap(). -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000+ If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania19026
Janardhana Reddy <jana-reddy@mediaring.com.sg> writes: > By mapping the WAL files by each backend in to its address > space using "mmap" system call , There are a lot of problems with trying to use mmap for Postgres. One is portability: not all platforms have mmap, so we'd still have to support the non-mmap case; and it's not at all clear that fsync/msync semantics are consistent across platforms, either. A bigger objection is that mmap'ing a file in one backend does not cause it to become available to other backends, thus the entire concept of shared buffers breaks down. If you think you can make it work, feel free to try it ... regards, tom lane
> > Hi all, > > By mapping the WAL files by each backend in to its address > > space using "mmap" system call , there will be big > > improvements in performance from the following point of view: > > 1. Each backend directly writes in to the address > > space which is obtained by maping the WAL files. > > this saves the write system call at the end of > > every transaction which transfres 8k of > > data from user space to kernel. > > 2. since every transaction does not modify all the 8k > > content of WAL page , so by issuing the > > fsync , the kernel only transfers only the > > kernel pages which are modified , which is 4k for > > linux so fsync time is saved by twice. > > Any comments ?. > > This is interesting. We are concerned about using mmap() for all I/O > because we could eat up quite a bit of address space for big tables, but > WAL seems like an ideal use for mmap(). OK, I have talked to Tom Lane about this on the phone and we have a few ideas. Historically, we have avoided mmap() because of portability problems, and because using mmap() to write to large tables could consume lots of address space with little benefit. However, I perhaps can see WAL as being a good use of mmap. First, there is the issue of using mmap(). For OS's that have the mmap() MAP_SHARED flag, different backends could mmap the same file and each see the changes. However, keep in mind we still have to fsync() WAL, so we need to use msync(). So, looking at the benefits of using mmap(), we have overhead of different backends having to mmap something that now sits quite easily in shared memory. Now, I can see mmap reducing the copy from user to kernel, but there are other ways to fix that. We could modify the write() routines to write() 8k on first WAL page write and later write only the modified part of the page to the kernel buffers. The old kernel buffer is probably still around so it is unlikely to require a read from the file system to read in the rest of the page. This reduces the write from 8k to something probably less than 4k which is better than we can do with mmap. I will add a TODO item to this effect. As far as reducing the write to disk from 8k to 4k, if we have to fsync/msync, we have to wait for the disk to spin to the proper location and at that point writing 4k or 8k doesn't seem like much of a win. In summary, I think it would be nice to reduce the 8k transfer from user to kernel on secondary page writes to only the modified part of the page. I am uncertain if mmap() or anything else will help the physical write to the disk. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000+ If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania19026
I have just completed the functional testing the WAL using mmap , it is working fine, I have tested by commenting out the "CreateCheckPoint " functionality so that when i kill the postgres and restart it will redo all the records from the WAL log file which is updated using mmap. Just i need to clean code and to do some stress testing.By the end of thisweek i should able to complete the stress test and generate the patch file . As Tom Lane mentioned i see the problem in portability to all platforms, what i propose is to use mmap for only WAL for some platforms like linux,freebsd etc . For other platforms we canuse the existing method by slightly modifying thewrite() routine to write only the modified part of the page. Regards jana > > > OK, I have talked to Tom Lane about this on the phone and we have a few > ideas. > > Historically, we have avoided mmap() because of portability problems, > and because using mmap() to write to large tables could consume lots of > address space with little benefit. However, I perhaps can see WAL as > being a good use of mmap. > > First, there is the issue of using mmap(). For OS's that have the > mmap() MAP_SHARED flag, different backends could mmap the same file and > each see the changes. However, keep in mind we still have to fsync() > WAL, so we need to use msync(). > > So, looking at the benefits of using mmap(), we have overhead of > different backends having to mmap something that now sits quite easily > in shared memory. Now, I can see mmap reducing the copy from user to > kernel, but there are other ways to fix that. We could modify the > write() routines to write() 8k on first WAL page write and later write > only the modified part of the page to the kernel buffers. The old > kernel buffer is probably still around so it is unlikely to require a > read from the file system to read in the rest of the page. This reduces > the write from 8k to something probably less than 4k which is better > than we can do with mmap. > > I will add a TODO item to this effect. > > As far as reducing the write to disk from 8k to 4k, if we have to > fsync/msync, we have to wait for the disk to spin to the proper location > and at that point writing 4k or 8k doesn't seem like much of a win. > > In summary, I think it would be nice to reduce the 8k transfer from user > to kernel on secondary page writes to only the modified part of the > page. I am uncertain if mmap() or anything else will help the physical > write to the disk. > > -- > Bruce Momjian | http://candle.pha.pa.us > pgman@candle.pha.pa.us | (610) 853-3000 > + If your life is a hard drive, | 830 Blythe Avenue > + Christ can be your backup. | Drexel Hill, Pennsylvania 19026
Sounds good. Keep us posted. This will probably not make it into 7.2 but can be added to 7.3. We can perhaps conditionally use your code in place of what is there. I have also looked at reducing the write() size for WAL secondary writes. That will have to wait for 7.3 too because we are so near beta. > I have just completed the functional testing the WAL using mmap , it is > > working fine, I have tested by commenting out the "CreateCheckPoint " > functionality so that > when i kill the postgres and restart it will redo all the records from the > WAL log file which > is updated using mmap. > Just i need to clean code and to do some stress testing. > By the end of this week i should able to complete the stress test and > generate the patch file . > As Tom Lane mentioned i see the problem in portability to all platforms, > > what i propose is to use mmap for only WAL for some platforms like > linux,freebsd etc . For other platforms we can use the existing method by > slightly modifying the > write() routine to write only the modified part of the page. > > Regards > jana > > > > > > > OK, I have talked to Tom Lane about this on the phone and we have a few > > ideas. > > > > Historically, we have avoided mmap() because of portability problems, > > and because using mmap() to write to large tables could consume lots of > > address space with little benefit. However, I perhaps can see WAL as > > being a good use of mmap. > > > > First, there is the issue of using mmap(). For OS's that have the > > mmap() MAP_SHARED flag, different backends could mmap the same file and > > each see the changes. However, keep in mind we still have to fsync() > > WAL, so we need to use msync(). > > > > So, looking at the benefits of using mmap(), we have overhead of > > different backends having to mmap something that now sits quite easily > > in shared memory. Now, I can see mmap reducing the copy from user to > > kernel, but there are other ways to fix that. We could modify the > > write() routines to write() 8k on first WAL page write and later write > > only the modified part of the page to the kernel buffers. The old > > kernel buffer is probably still around so it is unlikely to require a > > read from the file system to read in the rest of the page. This reduces > > the write from 8k to something probably less than 4k which is better > > than we can do with mmap. > > > > I will add a TODO item to this effect. > > > > As far as reducing the write to disk from 8k to 4k, if we have to > > fsync/msync, we have to wait for the disk to spin to the proper location > > and at that point writing 4k or 8k doesn't seem like much of a win. > > > > In summary, I think it would be nice to reduce the 8k transfer from user > > to kernel on secondary page writes to only the modified part of the > > page. I am uncertain if mmap() or anything else will help the physical > > write to the disk. > > > > -- > > Bruce Momjian | http://candle.pha.pa.us > > pgman@candle.pha.pa.us | (610) 853-3000 > > + If your life is a hard drive, | 830 Blythe Avenue > > + Christ can be your backup. | Drexel Hill, Pennsylvania 19026 > > ---------------------------(end of broadcast)--------------------------- > TIP 6: Have you searched our list archives? > > http://archives.postgresql.org > -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000+ If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania19026
I have added this to TODO.detail/mmap. > I have just completed the functional testing the WAL using mmap , it is > > working fine, I have tested by commenting out the "CreateCheckPoint " > functionality so that > when i kill the postgres and restart it will redo all the records from the > WAL log file which > is updated using mmap. > Just i need to clean code and to do some stress testing. > By the end of this week i should able to complete the stress test and > generate the patch file . > As Tom Lane mentioned i see the problem in portability to all platforms, > > what i propose is to use mmap for only WAL for some platforms like > linux,freebsd etc . For other platforms we can use the existing method by > slightly modifying the > write() routine to write only the modified part of the page. > > Regards > jana > > > > > > > OK, I have talked to Tom Lane about this on the phone and we have a few > > ideas. > > > > Historically, we have avoided mmap() because of portability problems, > > and because using mmap() to write to large tables could consume lots of > > address space with little benefit. However, I perhaps can see WAL as > > being a good use of mmap. > > > > First, there is the issue of using mmap(). For OS's that have the > > mmap() MAP_SHARED flag, different backends could mmap the same file and > > each see the changes. However, keep in mind we still have to fsync() > > WAL, so we need to use msync(). > > > > So, looking at the benefits of using mmap(), we have overhead of > > different backends having to mmap something that now sits quite easily > > in shared memory. Now, I can see mmap reducing the copy from user to > > kernel, but there are other ways to fix that. We could modify the > > write() routines to write() 8k on first WAL page write and later write > > only the modified part of the page to the kernel buffers. The old > > kernel buffer is probably still around so it is unlikely to require a > > read from the file system to read in the rest of the page. This reduces > > the write from 8k to something probably less than 4k which is better > > than we can do with mmap. > > > > I will add a TODO item to this effect. > > > > As far as reducing the write to disk from 8k to 4k, if we have to > > fsync/msync, we have to wait for the disk to spin to the proper location > > and at that point writing 4k or 8k doesn't seem like much of a win. > > > > In summary, I think it would be nice to reduce the 8k transfer from user > > to kernel on secondary page writes to only the modified part of the > > page. I am uncertain if mmap() or anything else will help the physical > > write to the disk. > > > > -- > > Bruce Momjian | http://candle.pha.pa.us > > pgman@candle.pha.pa.us | (610) 853-3000 > > + If your life is a hard drive, | 830 Blythe Avenue > > + Christ can be your backup. | Drexel Hill, Pennsylvania 19026 > > ---------------------------(end of broadcast)--------------------------- > TIP 6: Have you searched our list archives? > > http://archives.postgresql.org > -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 853-3000+ If your life is a hard drive, | 830 Blythe Avenue + Christ can be your backup. | Drexel Hill, Pennsylvania19026