Re: Some observations on very slow pg_restore operations

[MichaelDBA <MichaelDBA@xxxxxxxxxxx>]

Hi Ogden,

You didn't mention any details about your postgresql.conf settings.  Why don't you set them optimally for your loads and try again and see if there is any difference.  Make sure you do a DB restart since some of these parameters require it.

======================================
parameter		before	after
----------------	------	-------
shared_buffers		Reduce this value to about 25% of total memory
temp_buffers		Decrease this value to 8MB since we are not using temporary tables or doing intermediate sorts
work_mem		Reduce significantly (1MB) since we are not doing memory sorts or hashes per SQL
maintenance_work_mem	Increase signficantly for DDL bulk loading, restore operations
fsync    		off (so that time is not being spent waiting for stuff to be written to disk). Note: you may not be able to recover your database after a crash when set to off.
checkpoint_segments	Increase this significantly for DML bulk loading, restore operations
max_wal_size            Increase significantly like you would to checkpoint_segments
min_wal_size            Increase significantly like you would to checkpoint_segments
checkpoint_timeout      Increase to at least 30min
archive_mode 		off
autovacuum 		off
synchronous_commit	off
wal_level		minimal
max_wal_senders         0
full_page_writes        off during DML bulk loading, restore operations
wal_buffers             16MB during DML bulk loading, restore operations 

Regards,
Michael Vitale


Ogden Brash wrote on 10/3/2019 4:30 PM:

I recently performed a pg_dump (data-only) of a relatively large database where we store intermediate results of calculations. It is approximately 3 TB on disk and has about 20 billion rows.

We do the dump/restore about once a month and as the dataset has grown, the restores have gotten very slow. So, this time I decided to do it a different way and have some observations that puzzle me.

Background:

The data is extremely simple. The rows consist only of numbers and are all fixed length. There are no foreign keys, constraints, null values, or default values. There are no strings or arrays. There are 66 tables and the number of rows in each table forms a gaussian distribution; so there are 3 tables which have about 3 billion rows each and the rest of the tables have significantly fewer rows. 

I used the directory format when doing the pg_dump. The compressed data of the dump is 550 GB.

I am using: (PostgreSQL) 11.5 (Ubuntu 11.5-1.pgdg18.04+1)

The machine that I attempted to do a pg_restore to is a dedicated server just for one instance of posgresql. It has 32 GB of memory and is running Ubuntu 18.04 (headless). It physical hardware, not virtualized. Nothing else runs on the machine and the postgresql.conf settings have been tuned (to the best of my postgresql abilities which are suspect). While the operating system is installed on an SSD, there is one extra large, fast HDD that is dedicated to the posgresql server. It has been in use for this particular purpose for a while and has not had performance issues. (Just with pg_restore)

Autovacuum is off and all indexes have been deleted before the restore is started. There is nothing in the db except for the empty data tables.

Restoring over the net:

In the past we have always restored in a way where the dumped data is read over a gigabit connection while being restored to the local drive. But, the last time we did it it took 2 days and I was looking for something faster. So, I decided to copy the dumped directory to the local drive and restore from the dump locally. I knew that because the machine only had one drive that would fit the data, there would be some I/O contention, but I hoped that it might not be as bad as reading over the network.

The pg_restore went unbearably slowly... after many hours it had written less than 20GB to the database, so I started tracking it with iostat to see what was going on. The following is iostat output every 60 seconds. I tracked it for several hours and this is representative of what was happening consistently.

avg-cpu:  %user   %nice %system %iowait  %steal   %idle

           0.39    0.00    0.40   43.10    0.00   56.11

Device             tps    kB_read/s    kB_wrtn/s    kB_read    kB_wrtn

loop0             0.00         0.00         0.00          0          0

loop1             0.00         0.00         0.00          0          0

loop2             0.00         0.00         0.00          0          0

sda             263.33       132.87      2990.93       7972     179456

sdb               0.17         0.00         0.73          0         44

avg-cpu:  %user   %nice %system %iowait  %steal   %idle

           0.34    0.00    0.41   44.43    0.00   54.82

Device             tps    kB_read/s    kB_wrtn/s    kB_read    kB_wrtn

loop0             0.00         0.00         0.00          0          0

loop1             0.00         0.00         0.00          0          0

loop2             0.00         0.00         0.00          0          0

sda             262.95       140.47      2983.00       8428     178980

sdb               0.08         0.00         0.40          0         24

While I was tracking this I started experimenting with the IO scheduler to see if it had a noticable impact. I had been using cfq (ubuntu 18.04 default). Changing to deadline did not have a noticable difference. Changing to noop made things much slower. I went back to cfq. I also experimented with turning fsync off; that did speed things up a bit but not enough for me to leave it off.

What puzzled me is that the OS was spending such a large percentage of time in iowait, yet there was so little IO going on.

So, I decided to go back to restoring over the net. While the slow pg_restore was still going on, and while I was still tracking iostat, I copied the 550 GB dumps to an nfs drive. The copy happened pretty much at full speed (limit being the gigabit ethernet) and interestingly, it did not slow down kb_wrtn and kb_wrtn/s numbers in iostat (which was the postgresql server continuing with the restore). To me that seemed to indicate that it was not really a disk I/O limitation.

Restoring over the net:

After copying the dump files to an NFS drive, I stopped the restore, truncated the tables and started exactly the same command, but this time taking its input from the nfs drive. I did not reboot the machine or restart the postgresql server. I tracked iostate every 60 seconds and this is what it looks like:

avg-cpu:  %user   %nice %system %iowait  %steal   %idle

           8.87    0.00    1.62   39.89    0.00   49.61

Device             tps    kB_read/s    kB_wrtn/s    kB_read    kB_wrtn

loop0             0.00         0.00         0.00          0          0

loop1             0.00         0.00         0.00          0          0

loop2             0.00         0.00         0.00          0          0

sda             252.77       527.87     37837.47      31672    2270248

sdb               0.22         0.00         1.00          0         60

avg-cpu:  %user   %nice %system %iowait  %steal   %idle

           8.57    0.00    2.21   35.26    0.00   53.97

Device             tps    kB_read/s    kB_wrtn/s    kB_read    kB_wrtn

loop0             0.00         0.00         0.00          0          0

loop1             0.00         0.00         0.00          0          0

loop2             0.00         0.00         0.00          0          0

sda             236.10       465.27     54312.00      27916    3258720

sdb               0.08         0.00         0.40          0         24

Notice that the database is writing approximately 15 times as fast (and I have verified that by tracking the size of the posgresql data directory over time) while the number of i/o transactions per second has actually dropped a little bit. It has now been running about 24 hours and has maintained that speed.

My interpretation

At first sight this seems to me as being symptomatic of the pg_restore process doing a huge number of very small input operations when reading from the dump. If the proportion of input to output operations is the same now as it was when trying to restore from the local drive, that implies that the vast majority of i/o operations were inputs and not outputs.

However, I am not sure that even that would cause such a slowdown because the compressed data files in the directory format dump correspond to the tables and so there are 3 very large files that it starts with. So all of these stats were gathered in the first 24 hours of the restore when it was just restoring the first 3 tables (I have verbose on, so I know). Because those files are gzipped, we know that they are being read sequentially and because the machine has lots of memory we know that the OS has allocated a lot of space to disk buffers and so even if postgresql was doing lots of small reads, bouncing around between the 3 files, it would not hit the disk that often.

Now that restore is happening 15 times faster when reading from an nfs drive, I looked at the nfsiostat output for a while and it does not show any indication of any untoward behavior:

192.168.17.146:/volume1/Backups mounted on /nas/Backups:

           ops/s       rpc bklog

          27.000           0.000

read:              ops/s            kB/s           kB/op         retrans    avg RTT (ms)    avg exe (ms)

                  27.000        3464.332         128.309        0 (0.0%)          13.500          13.511

write:             ops/s            kB/s           kB/op         retrans    avg RTT (ms)    avg exe (ms)

                   0.000           0.000           0.000        0 (0.0%)           0.000           0.000

192.168.17.146:/volume1/Backups mounted on /nas/Backups:

           ops/s       rpc bklog

          24.000           0.000

read:              ops/s            kB/s           kB/op         retrans    avg RTT (ms)    avg exe (ms)

                  24.000        3079.406         128.309        0 (0.0%)          28.492          28.504

write:             ops/s            kB/s           kB/op         retrans    avg RTT (ms)    avg exe (ms)

                   0.000           0.000           0.000        0 (0.0%)           0.000           0.000

The nubmer of operations per second (if those correspond to reads from postgresql, which I do not know for a fact) does not seem high at all. 

I actually do not have a great theory for what is going on but it might be more obvious to someone who knows the postgresql implementation well. I would love to hear any thoughts that would be helpful on how to get my restores even faster.