You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to ClickHouse
Hi there,
I'm sending INSERTS in bulk but even with this I ran into limitations. We have a lot of data points and soon as we bulk more than 200 inserts into one call we exceed some limitations though at only 200 inserts per bulk we do 10-20 bulk inserts per second which is bad to say the least. Can we increase the bulk size without causing any issues? I'm not sure if this is some sort of http limit or the max_insert_block_size setting. Any help would be much appreciated.
Denis Zhuravlev
unread,
Jul 15, 2018, 11:33:58 AM7/15/18
Reply to author
Sign in to reply to author
Forward
Sign in to forward
Delete
You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to ClickHouse
Not a bulk:
insert into t values (?,?);
insert into t values (?,?);
Bulk:
insert into t values (?,?), (?,?), (?,?), (?,?), (?,?), (?,?), (?,?), (?,?), .... (?,?)*100000 times ;
Or accumulate input data in TSV/CSV file and execute [ cat accumulate.tsv|clickhouse-client -q "insert into t format TSV" ] once a minute.
Jack Gao
unread,
Jul 25, 2018, 3:38:06 AM7/25/18
Reply to author
Sign in to reply to author
Forward
Sign in to forward
Delete
You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to ClickHouse
you should increase the bulk size to 10000 or more.
ClickHouse works better with bulk insert.
Cause ClickHouse use background merge, which means each write is a part, the parts will be merged to a bigger one. If each write is too small, the merge progress will not catch the insert speed, and will throw an exception 'too many parts'
The exception is controlled by the following parameters:
the merge progress is controlled by 'background_pool_size', which is 16 default, it's fine for most situation.
在 2018年7月15日星期日 UTC+8下午10:03:21,__写道:
Yijian Huang
unread,
Jul 27, 2018, 1:29:44 AM7/27/18
Reply to author
Sign in to reply to author
Forward
Sign in to forward
Delete
You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to ClickHouse
What it is "records inserted per second" can we usually achieve for a single ClickHouse server ?
__
unread,
Jul 29, 2018, 10:06:51 AM7/29/18
Reply to author
Sign in to reply to author
Forward
Sign in to forward
Delete
You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to ClickHouse
I'm already doing these bulks. Yeah I could probably use a different method where I would not have the limitation of 200 per bulk. Though one main issue still remains which is that I need real time data. Meaning a delay of +30 seconds is a no go.
Denis Zhuravlev
unread,
Jul 30, 2018, 10:56:06 PM7/30/18
Reply to author
Sign in to reply to author
Forward
Sign in to forward
Delete
You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to ClickHouse
>more than 200 inserts into one call we exceed some limitations
What is 200 limit you talking about? CH able to insert millions rows by one insert.
On Sunday, 15 July 2018 11:03:21 UTC-3, __ wrote:
__
unread,
Jul 31, 2018, 7:10:53 AM7/31/18
Reply to author
Sign in to reply to author
Forward
Sign in to forward
Delete
You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to ClickHouse
I'm using https://github.com/nikepan/clickhouse-bulk which is inserting the data with http requests, I think there is some limit with this http method. Can this be increased?
__
unread,
Jul 31, 2018, 7:11:58 AM7/31/18
Reply to author
Sign in to reply to author
Forward
Sign in to forward
Delete
You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message
to ClickHouse
As soon as I set the bulk to more than 200 the data is cutoff and the entire query fails due to incompletion. (the inserts are quite long)
Denis Zhuravlev
unread,
Jul 31, 2018, 12:51:52 PM7/31/18
Reply to author
Sign in to reply to author
Forward
Sign in to forward
Delete
You do not have permission to delete messages in this group
Copy link
Report message
Show original message
Either email addresses are anonymous for this group or you need the view member email addresses permission to view the original message