Re: [Maria-developers] [External] Re: Questions regarding closing partitions in MDEV-11084
Hi Holyfoot, Sergei and Jacob, Thank you for looking into this. Jacob there are no Jira ticket (that I am aware of) also I do see it as two different problems with the same source. 1) Spider engine does not handle opening/closing specific partitions (as I understand it currently relies on opening/closing all partitions at once in order). Currently this is a crashing bug! 2) MDEV-11084 makes explicit partition selection work like FLUSH TABLE, due to the closing partitions part. I think it would be nice if you could create tickets for those two issues (or at least for the first one). Regards Mattias On Thu, Apr 26, 2018 at 11:26 AM, Alexey Botchkov <holyfoot@mariadb.com> wrote: > No, there's no JIRA task for that problem. > So i guess you can create one :) > > Best regards. > HF > > On Thu, Apr 26, 2018 at 4:50 AM, Jacob Mathew <jacob.mathew@mariadb.com> > wrote: > >> Hi Mattias and Holyfoot, >> >> Is there a Jira bug for this problem? If not I can create a bug for it >> and assign it to Holyfoot. >> >> Thanks, >> Jacob >> >> Jacob B. Mathew >> Spider and Server Developer >> MariaDB Corporation >> +1 408 655 8999 (mobile) >> jacob.b.mathew (Skype) >> jacob.mathew@MariaDB.com >> >> >> On Tue, Apr 17, 2018 at 9:11 AM, Mattias Jonsson < >> mattias.jonsson@booking.com> wrote: >> >>> Hi Holyfoot, >>> >>> On Mon, Apr 16, 2018 at 3:38 PM, Alexey Botchkov <holyfoot@askmonty.org> >>> wrote: >>> > Hi, Mattias, guys! >>> > >>> > While investigating the crash, i'd like to discuss that >>> > >>> >> it seems to close partitions whenever it >>> >> is not used in a statement (i.e. require it to be reopened in the next >>> >> statement that would use another partition >>> > >>> > >>> > Yes, it does that, handling statements with the specified 'PARTITION' >>> > option. >>> > The patch supposed to solve the problem when there are too many >>> partitions >>> > opened, >>> > so i think it must close the unused partitions sometime. >>> > No, it doesn't have to happen that often. I planned to check the >>> > table_open_cache >>> > variable before the forced close. But decided not to do that initially >>> - as >>> > it simplified testing, >>> > and i thought if someone uses the PARTITION option, he would stick to >>> using >>> > this partition >>> > anyway. And i forgot about that issue. >>> >>> The reason for not closing partitions in this case is that it turns >>> the 'SELECT col FROM t PARTITION(p)' almost into a 'FLUSH TABLES t' >>> which is kind of unexpected. >>> I did a test mixing PK selects with and without 'PARTITION (p)' clause >>> and that shows it will close all but 1 partition and then (re)open all >>> but 1 partition. >>> >>> Think of the case when a server runs in production with a heavily >>> partitioned table serving simple PK queries and then someone runs a >>> query with explicit PARTITION selection, then it will introduce a >>> short stall. First for the query itself (closing all but 1 partitions) >>> and then for the next simple PK query using the same table (opening >>> all but 1 partition). >>> >>> As I read the bug report: the reporter wants to avoid opening all >>> partitions. Not that it keeps the partitions open in the table open >>> cache (which is an issue on the architectural level of partitioning >>> not really fitting into the open table cache). >>> >>> I attached a diff with the test and results (I also added handler >>> status variables to show my point). The diff is against >>> b4a2baffa82e5c07b96a1c752228560dcac1359b. >>> >>> Here is the part of the result file that shows what I mean with extra >>> comments prepended by MJ>: >>> CREATE TABLE t1 (a int PRIMARY KEY) >>> ENGINE = InnoDB >>> PARTITION BY HASH (a) PARTITIONS 1000; >>> INSERT INTO t1 VALUES (0), (1), (2), (3); >>> FLUSH STATUS; >>> >>> SELECT a FROM t1 PARTITION(p0) WHERE a = 0; >>> a >>> 0 >>> SHOW SESSION STATUS >>> WHERE VARIABLE_NAME LIKE 'HANDLER_%' AND VALUE > 0; >>> Variable_name Value >>> MJ> Here it closes all but one partitions. >>> Handler_close 999 >>> >>> SELECT a FROM t1 WHERE a = 0; >>> a >>> 0 >>> SHOW SESSION STATUS >>> WHERE VARIABLE_NAME LIKE 'HANDLER_%' AND VALUE > 0; >>> Variable_name Value >>> Handler_close 1000 >>> Handler_external_lock 2006 >>> MJ> and here it re-opens those 'all but one' partitions. >>> Handler_open 999 >>> >>> >>> Regards >>> Mattias >>> > >>> > >>> > Best regards. >>> > HF >>> > >>> > >>> > >>> > >>> > 13.04.2018 19:07, Mattias Jonsson wrote: >>> >> >>> >> Hi MariaDB Devs, >>> >> >>> >> I tried to evaluate spider engine and found an issue where it crashes, >>> >> most likely due to MDEV-11084 (Stacktrace and reproducible test case >>> >> attached). >>> >> >>> >> That also leads me to wonder about the performance for partitioned >>> >> tables after MDEV-11084, when it seems to close partitions whenever it >>> >> is not used in a statement (i.e. require it to be reopened in the next >>> >> statement that would use another partition, effectively not using the >>> >> open table cache). >>> >> >>> >> I cannot see anything mentioned in the final commit message hinting on >>> >> that it closes partitions not used in the current query, but in the >>> >> previous patches it was mentioned without any reason. >>> >> >>> >> Why does it close the already opened partitions? >>> >> https://github.com/MariaDB/server/blob/10.3/sql/ha_partition.cc#L8365 >>> >> >>> >> I would not mind opening the partitions only when they are to be used >>> >> (even though there are engines that need to be tested more), but >>> >> closing them makes no sense to me performance wise. Also notice that >>> >> the partitions first will be put back into the open table cache and >>> >> then on the next query the non-used partitions will be closed and the >>> >> needed ones be (re)-opened. >>> >> >>> >> Regards >>> >> Mattias Jonsson >>> > >>> > >>> > >>> > >>> >> >> > -- Mattias Jonsson IT Team Lead - Development Booking.com B.V. Herengracht 597 Amsterdam 1017 CE Netherlands Direct +31207125646 [image: Booking.com] <http://www.booking.com/> The world's #1 accommodation site 43 languages, 198+ offices worldwide, 120,000+ global destinations, 1,550,000+ room nights booked every day No booking fees, best price always guaranteed Subsidiary of Booking Holdings Inc. (NASDAQ: BKNG)
Hi Mattias, There is an instance of Spider for each partition in the Spider node, so each instance of Spider handles exactly one partition/shard. So please elaborate on what you mean by Spider opening/closing specific partitions. Note that if we push the partitioning logic into each storage engine, then there could be one instance of Spider in the Spider node that handles all partitions. Thanks, Jacob Jacob B. Mathew Spider and Server Developer MariaDB Corporation +1 408 655 8999 (mobile) jacob.b.mathew (Skype) jacob.mathew@MariaDB.com On Thu, Apr 26, 2018 at 3:59 AM, Mattias Jonsson < mattias.jonsson@booking.com> wrote:
Hi Holyfoot, Sergei and Jacob,
Thank you for looking into this.
Jacob there are no Jira ticket (that I am aware of) also I do see it as two different problems with the same source. 1) Spider engine does not handle opening/closing specific partitions (as I understand it currently relies on opening/closing all partitions at once in order). Currently this is a crashing bug! 2) MDEV-11084 makes explicit partition selection work like FLUSH TABLE, due to the closing partitions part.
I think it would be nice if you could create tickets for those two issues (or at least for the first one).
Regards Mattias
On Thu, Apr 26, 2018 at 11:26 AM, Alexey Botchkov <holyfoot@mariadb.com> wrote:
No, there's no JIRA task for that problem. So i guess you can create one :)
Best regards. HF
On Thu, Apr 26, 2018 at 4:50 AM, Jacob Mathew <jacob.mathew@mariadb.com> wrote:
Hi Mattias and Holyfoot,
Is there a Jira bug for this problem? If not I can create a bug for it and assign it to Holyfoot.
Thanks, Jacob
Jacob B. Mathew Spider and Server Developer MariaDB Corporation +1 408 655 8999 (mobile) jacob.b.mathew (Skype) jacob.mathew@MariaDB.com
On Tue, Apr 17, 2018 at 9:11 AM, Mattias Jonsson < mattias.jonsson@booking.com> wrote:
Hi Holyfoot,
Hi, Mattias, guys!
While investigating the crash, i'd like to discuss that
it seems to close partitions whenever it is not used in a statement (i.e. require it to be reopened in the next statement that would use another partition
Yes, it does that, handling statements with the specified 'PARTITION' option. The patch supposed to solve the problem when there are too many
On Mon, Apr 16, 2018 at 3:38 PM, Alexey Botchkov <holyfoot@askmonty.org> wrote: partitions
opened, so i think it must close the unused partitions sometime. No, it doesn't have to happen that often. I planned to check the table_open_cache variable before the forced close. But decided not to do that initially - as it simplified testing, and i thought if someone uses the PARTITION option, he would stick to using this partition anyway. And i forgot about that issue.
The reason for not closing partitions in this case is that it turns the 'SELECT col FROM t PARTITION(p)' almost into a 'FLUSH TABLES t' which is kind of unexpected. I did a test mixing PK selects with and without 'PARTITION (p)' clause and that shows it will close all but 1 partition and then (re)open all but 1 partition.
Think of the case when a server runs in production with a heavily partitioned table serving simple PK queries and then someone runs a query with explicit PARTITION selection, then it will introduce a short stall. First for the query itself (closing all but 1 partitions) and then for the next simple PK query using the same table (opening all but 1 partition).
As I read the bug report: the reporter wants to avoid opening all partitions. Not that it keeps the partitions open in the table open cache (which is an issue on the architectural level of partitioning not really fitting into the open table cache).
I attached a diff with the test and results (I also added handler status variables to show my point). The diff is against b4a2baffa82e5c07b96a1c752228560dcac1359b.
Here is the part of the result file that shows what I mean with extra comments prepended by MJ>: CREATE TABLE t1 (a int PRIMARY KEY) ENGINE = InnoDB PARTITION BY HASH (a) PARTITIONS 1000; INSERT INTO t1 VALUES (0), (1), (2), (3); FLUSH STATUS;
SELECT a FROM t1 PARTITION(p0) WHERE a = 0; a 0 SHOW SESSION STATUS WHERE VARIABLE_NAME LIKE 'HANDLER_%' AND VALUE > 0; Variable_name Value MJ> Here it closes all but one partitions. Handler_close 999
SELECT a FROM t1 WHERE a = 0; a 0 SHOW SESSION STATUS WHERE VARIABLE_NAME LIKE 'HANDLER_%' AND VALUE > 0; Variable_name Value Handler_close 1000 Handler_external_lock 2006 MJ> and here it re-opens those 'all but one' partitions. Handler_open 999
Regards Mattias
Best regards. HF
13.04.2018 19:07, Mattias Jonsson wrote:
Hi MariaDB Devs,
I tried to evaluate spider engine and found an issue where it
crashes,
most likely due to MDEV-11084 (Stacktrace and reproducible test case attached).
That also leads me to wonder about the performance for partitioned tables after MDEV-11084, when it seems to close partitions whenever it is not used in a statement (i.e. require it to be reopened in the next statement that would use another partition, effectively not using the open table cache).
I cannot see anything mentioned in the final commit message hinting on that it closes partitions not used in the current query, but in the previous patches it was mentioned without any reason.
Why does it close the already opened partitions? https://github.com/MariaDB/server/blob/10.3/sql/ha_partition .cc#L8365
I would not mind opening the partitions only when they are to be used (even though there are engines that need to be tested more), but closing them makes no sense to me performance wise. Also notice that the partitions first will be put back into the open table cache and then on the next query the non-used partitions will be closed and the needed ones be (re)-opened.
Regards Mattias Jonsson
-- Mattias Jonsson IT Team Lead - Development
Booking.com B.V. Herengracht 597 Amsterdam 1017 CE Netherlands <https://maps.google.com/?q=Herengracht+597++Amsterdam++1017+CE+Netherlands&entry=gmail&source=g> Direct +31207125646 [image: Booking.com] <http://www.booking.com/> The world's #1 accommodation site 43 languages, 198+ offices worldwide, 120,000+ global destinations, 1,550,000+ room nights booked every day No booking fees, best price always guaranteed Subsidiary of Booking Holdings Inc. (NASDAQ: BKNG)
Hi Jacob, Sorry for the confusion between the partitioning engine and spider. To add to the confusion: There are also different 'share' objects, table share (meaning to share info between different instances of the same table) and also partition share (meaning to share information for different instances of the same partition). And they can also be on server level (TABLE_SHARE, ha_share) as well as handler level (ha_partition + ha_spider, Partition_share, SPIDER_SHARE, SPIDER_PARTITION_HANDLER_SHARE, SPIDER_PARTITION_SHARE). So what I mean is that the state of these shared objects may rely on the order of opening/closing partitions, especially the first opened partition (and probably relies on when closing the first opened partition there will be no more open partitions used). For pushing the partition logic into each storage engine, you can take a look into what was done in WL#4807/WL#6035 in MySQL where the partition engine was splitted into helper functions which was reused for both the partition engine as well as for the native partitioning in InnoDB. I think this would make a lot of sense to do for the spider engine as well. Regards Mattias On Thu, Apr 26, 2018 at 10:14 PM, Jacob Mathew <jacob.mathew@mariadb.com> wrote:
Hi Mattias,
There is an instance of Spider for each partition in the Spider node, so each instance of Spider handles exactly one partition/shard. So please elaborate on what you mean by Spider opening/closing specific partitions.
Note that if we push the partitioning logic into each storage engine, then there could be one instance of Spider in the Spider node that handles all partitions.
Thanks, Jacob
Jacob B. Mathew Spider and Server Developer MariaDB Corporation +1 408 655 8999 (mobile) jacob.b.mathew (Skype) jacob.mathew@MariaDB.com
On Thu, Apr 26, 2018 at 3:59 AM, Mattias Jonsson <mattias.jonsson@booking.com> wrote:
Hi Holyfoot, Sergei and Jacob,
Thank you for looking into this.
Jacob there are no Jira ticket (that I am aware of) also I do see it as two different problems with the same source. 1) Spider engine does not handle opening/closing specific partitions (as I understand it currently relies on opening/closing all partitions at once in order). Currently this is a crashing bug! 2) MDEV-11084 makes explicit partition selection work like FLUSH TABLE, due to the closing partitions part.
I think it would be nice if you could create tickets for those two issues (or at least for the first one).
Regards Mattias
On Thu, Apr 26, 2018 at 11:26 AM, Alexey Botchkov <holyfoot@mariadb.com> wrote:
No, there's no JIRA task for that problem. So i guess you can create one :)
Best regards. HF
On Thu, Apr 26, 2018 at 4:50 AM, Jacob Mathew <jacob.mathew@mariadb.com> wrote:
Hi Mattias and Holyfoot,
Is there a Jira bug for this problem? If not I can create a bug for it and assign it to Holyfoot.
Thanks, Jacob
Jacob B. Mathew Spider and Server Developer MariaDB Corporation +1 408 655 8999 (mobile) jacob.b.mathew (Skype) jacob.mathew@MariaDB.com
On Tue, Apr 17, 2018 at 9:11 AM, Mattias Jonsson <mattias.jonsson@booking.com> wrote:
Hi Holyfoot,
On Mon, Apr 16, 2018 at 3:38 PM, Alexey Botchkov <holyfoot@askmonty.org> wrote:
Hi, Mattias, guys!
While investigating the crash, i'd like to discuss that
> it seems to close partitions whenever it > is not used in a statement (i.e. require it to be reopened in the next > statement that would use another partition
Yes, it does that, handling statements with the specified 'PARTITION' option. The patch supposed to solve the problem when there are too many partitions opened, so i think it must close the unused partitions sometime. No, it doesn't have to happen that often. I planned to check the table_open_cache variable before the forced close. But decided not to do that initially - as it simplified testing, and i thought if someone uses the PARTITION option, he would stick to using this partition anyway. And i forgot about that issue.
The reason for not closing partitions in this case is that it turns the 'SELECT col FROM t PARTITION(p)' almost into a 'FLUSH TABLES t' which is kind of unexpected. I did a test mixing PK selects with and without 'PARTITION (p)' clause and that shows it will close all but 1 partition and then (re)open all but 1 partition.
Think of the case when a server runs in production with a heavily partitioned table serving simple PK queries and then someone runs a query with explicit PARTITION selection, then it will introduce a short stall. First for the query itself (closing all but 1 partitions) and then for the next simple PK query using the same table (opening all but 1 partition).
As I read the bug report: the reporter wants to avoid opening all partitions. Not that it keeps the partitions open in the table open cache (which is an issue on the architectural level of partitioning not really fitting into the open table cache).
I attached a diff with the test and results (I also added handler status variables to show my point). The diff is against b4a2baffa82e5c07b96a1c752228560dcac1359b.
Here is the part of the result file that shows what I mean with extra comments prepended by MJ>: CREATE TABLE t1 (a int PRIMARY KEY) ENGINE = InnoDB PARTITION BY HASH (a) PARTITIONS 1000; INSERT INTO t1 VALUES (0), (1), (2), (3); FLUSH STATUS;
SELECT a FROM t1 PARTITION(p0) WHERE a = 0; a 0 SHOW SESSION STATUS WHERE VARIABLE_NAME LIKE 'HANDLER_%' AND VALUE > 0; Variable_name Value MJ> Here it closes all but one partitions. Handler_close 999
SELECT a FROM t1 WHERE a = 0; a 0 SHOW SESSION STATUS WHERE VARIABLE_NAME LIKE 'HANDLER_%' AND VALUE > 0; Variable_name Value Handler_close 1000 Handler_external_lock 2006 MJ> and here it re-opens those 'all but one' partitions. Handler_open 999
Regards Mattias
Best regards. HF
13.04.2018 19:07, Mattias Jonsson wrote: > > Hi MariaDB Devs, > > I tried to evaluate spider engine and found an issue where it crashes, > most likely due to MDEV-11084 (Stacktrace and reproducible test case > attached). > > That also leads me to wonder about the performance for partitioned > tables after MDEV-11084, when it seems to close partitions whenever it > is not used in a statement (i.e. require it to be reopened in the next > statement that would use another partition, effectively not using the > open table cache). > > I cannot see anything mentioned in the final commit message hinting on > that it closes partitions not used in the current query, but in the > previous patches it was mentioned without any reason. > > Why does it close the already opened partitions? > https://github.com/MariaDB/server/blob/10.3/sql/ha_partition.cc#L8365 > > I would not mind opening the partitions only when they are to be used > (even though there are engines that need to be tested more), but > closing them makes no sense to me performance wise. Also notice that > the partitions first will be put back into the open table cache and > then on the next query the non-used partitions will be closed and the > needed ones be (re)-opened. > > Regards > Mattias Jonsson
Hello, guys.
2) MDEV-11084 makes explicit partition selection work like FLUSH TABLE, due to the closing partitions part. After the discussion, I removed that part of the fix, so now the explicit statement won't force the partition close. So now the 'explicit_partition.test' passes.
Best regards. HF 26.04.2018 14:59, Mattias Jonsson wrote:
Hi Holyfoot, Sergei and Jacob,
Thank you for looking into this.
Jacob there are no Jira ticket (that I am aware of) also I do see it as two different problems with the same source. 1) Spider engine does not handle opening/closing specific partitions (as I understand it currently relies on opening/closing all partitions at once in order). Currently this is a crashing bug! 2) MDEV-11084 makes explicit partition selection work like FLUSH TABLE, due to the closing partitions part.
I think it would be nice if you could create tickets for those two issues (or at least for the first one).
Regards Mattias
On Thu, Apr 26, 2018 at 11:26 AM, Alexey Botchkov <holyfoot@mariadb.com <mailto:holyfoot@mariadb.com>> wrote:
No, there's no JIRA task for that problem. So i guess you can create one :)
Best regards. HF
On Thu, Apr 26, 2018 at 4:50 AM, Jacob Mathew <jacob.mathew@mariadb.com <mailto:jacob.mathew@mariadb.com>> wrote:
Hi Mattias and Holyfoot,
Is there a Jira bug for this problem? If not I can create a bug for it and assign it to Holyfoot.
Thanks, Jacob
Jacob B. Mathew Spider and Server Developer MariaDB Corporation +1 408 655 8999 (mobile) jacob.b.mathew (Skype) jacob.mathew@MariaDB.com
On Tue, Apr 17, 2018 at 9:11 AM, Mattias Jonsson <mattias.jonsson@booking.com <mailto:mattias.jonsson@booking.com>> wrote:
Hi Holyfoot,
On Mon, Apr 16, 2018 at 3:38 PM, Alexey Botchkov <holyfoot@askmonty.org <mailto:holyfoot@askmonty.org>> wrote: > Hi, Mattias, guys! > > While investigating the crash, i'd like to discuss that > >> it seems to close partitions whenever it >> is not used in a statement (i.e. require it to be reopened in the next >> statement that would use another partition > > > Yes, it does that, handling statements with the specified 'PARTITION' > option. > The patch supposed to solve the problem when there are too many partitions > opened, > so i think it must close the unused partitions sometime. > No, it doesn't have to happen that often. I planned to check the > table_open_cache > variable before the forced close. But decided not to do that initially - as > it simplified testing, > and i thought if someone uses the PARTITION option, he would stick to using > this partition > anyway. And i forgot about that issue.
The reason for not closing partitions in this case is that it turns the 'SELECT col FROM t PARTITION(p)' almost into a 'FLUSH TABLES t' which is kind of unexpected. I did a test mixing PK selects with and without 'PARTITION (p)' clause and that shows it will close all but 1 partition and then (re)open all but 1 partition.
Think of the case when a server runs in production with a heavily partitioned table serving simple PK queries and then someone runs a query with explicit PARTITION selection, then it will introduce a short stall. First for the query itself (closing all but 1 partitions) and then for the next simple PK query using the same table (opening all but 1 partition).
As I read the bug report: the reporter wants to avoid opening all partitions. Not that it keeps the partitions open in the table open cache (which is an issue on the architectural level of partitioning not really fitting into the open table cache).
I attached a diff with the test and results (I also added handler status variables to show my point). The diff is against b4a2baffa82e5c07b96a1c752228560dcac1359b.
Here is the part of the result file that shows what I mean with extra comments prepended by MJ>: CREATE TABLE t1 (a int PRIMARY KEY) ENGINE = InnoDB PARTITION BY HASH (a) PARTITIONS 1000; INSERT INTO t1 VALUES (0), (1), (2), (3); FLUSH STATUS;
SELECT a FROM t1 PARTITION(p0) WHERE a = 0; a 0 SHOW SESSION STATUS WHERE VARIABLE_NAME LIKE 'HANDLER_%' AND VALUE > 0; Variable_name Value MJ> Here it closes all but one partitions. Handler_close 999
SELECT a FROM t1 WHERE a = 0; a 0 SHOW SESSION STATUS WHERE VARIABLE_NAME LIKE 'HANDLER_%' AND VALUE > 0; Variable_name Value Handler_close 1000 Handler_external_lock 2006 MJ> and here it re-opens those 'all but one' partitions. Handler_open 999
Regards Mattias > > > Best regards. > HF > > > > > 13.04.2018 19:07, Mattias Jonsson wrote: >> >> Hi MariaDB Devs, >> >> I tried to evaluate spider engine and found an issue where it crashes, >> most likely due to MDEV-11084 (Stacktrace and reproducible test case >> attached). >> >> That also leads me to wonder about the performance for partitioned >> tables after MDEV-11084, when it seems to close partitions whenever it >> is not used in a statement (i.e. require it to be reopened in the next >> statement that would use another partition, effectively not using the >> open table cache). >> >> I cannot see anything mentioned in the final commit message hinting on >> that it closes partitions not used in the current query, but in the >> previous patches it was mentioned without any reason. >> >> Why does it close the already opened partitions? >> https://github.com/MariaDB/server/blob/10.3/sql/ha_partition.cc#L8365 <https://github.com/MariaDB/server/blob/10.3/sql/ha_partition.cc#L8365> >> >> I would not mind opening the partitions only when they are to be used >> (even though there are engines that need to be tested more), but >> closing them makes no sense to me performance wise. Also notice that >> the partitions first will be put back into the open table cache and >> then on the next query the non-used partitions will be closed and the >> needed ones be (re)-opened. >> >> Regards >> Mattias Jonsson > > > >
-- Mattias Jonsson IT Team Lead - Development
Booking.com B.V. Herengracht 597 Amsterdam 1017 CE Netherlands Direct +31207125646 Booking.com <http://www.booking.com/> The world's #1 accommodation site 43 languages, 198+ offices worldwide, 120,000+ global destinations, 1,550,000+ room nights booked every day No booking fees, best price always guaranteed Subsidiary of Booking Holdings Inc. (NASDAQ: BKNG)
Hi Holyfoot, Thank you for the fix. Btw should the tests be included as well or did you add them outside the commit? Regards Mattias On Sat, Apr 28, 2018 at 1:50 PM, Alexey Botchkov <holyfoot@askmonty.org> wrote:
Hello, guys.
2) MDEV-11084 makes explicit partition selection work like FLUSH TABLE, due to the closing partitions part. After the discussion, I removed that part of the fix, so now the explicit statement won't force the partition close. So now the 'explicit_partition.test' passes.
Best regards. HF
26.04.2018 14:59, Mattias Jonsson wrote:
Hi Holyfoot, Sergei and Jacob,
Thank you for looking into this.
Jacob there are no Jira ticket (that I am aware of) also I do see it as two different problems with the same source. 1) Spider engine does not handle opening/closing specific partitions (as I understand it currently relies on opening/closing all partitions at once in order). Currently this is a crashing bug! 2) MDEV-11084 makes explicit partition selection work like FLUSH TABLE, due to the closing partitions part.
I think it would be nice if you could create tickets for those two issues (or at least for the first one).
Regards Mattias
On Thu, Apr 26, 2018 at 11:26 AM, Alexey Botchkov <holyfoot@mariadb.com> wrote:
No, there's no JIRA task for that problem. So i guess you can create one :)
Best regards. HF
On Thu, Apr 26, 2018 at 4:50 AM, Jacob Mathew <jacob.mathew@mariadb.com> wrote:
Hi Mattias and Holyfoot,
Is there a Jira bug for this problem? If not I can create a bug for it and assign it to Holyfoot.
Thanks, Jacob
Jacob B. Mathew Spider and Server Developer MariaDB Corporation +1 408 655 8999 (mobile) jacob.b.mathew (Skype) jacob.mathew@MariaDB.com
On Tue, Apr 17, 2018 at 9:11 AM, Mattias Jonsson < mattias.jonsson@booking.com> wrote:
Hi Holyfoot,
Hi, Mattias, guys!
While investigating the crash, i'd like to discuss that
it seems to close partitions whenever it is not used in a statement (i.e. require it to be reopened in the next statement that would use another partition
Yes, it does that, handling statements with the specified 'PARTITION' option. The patch supposed to solve the problem when there are too many
On Mon, Apr 16, 2018 at 3:38 PM, Alexey Botchkov <holyfoot@askmonty.org> wrote: partitions
opened, so i think it must close the unused partitions sometime. No, it doesn't have to happen that often. I planned to check the table_open_cache variable before the forced close. But decided not to do that initially - as it simplified testing, and i thought if someone uses the PARTITION option, he would stick to using this partition anyway. And i forgot about that issue.
The reason for not closing partitions in this case is that it turns the 'SELECT col FROM t PARTITION(p)' almost into a 'FLUSH TABLES t' which is kind of unexpected. I did a test mixing PK selects with and without 'PARTITION (p)' clause and that shows it will close all but 1 partition and then (re)open all but 1 partition.
Think of the case when a server runs in production with a heavily partitioned table serving simple PK queries and then someone runs a query with explicit PARTITION selection, then it will introduce a short stall. First for the query itself (closing all but 1 partitions) and then for the next simple PK query using the same table (opening all but 1 partition).
As I read the bug report: the reporter wants to avoid opening all partitions. Not that it keeps the partitions open in the table open cache (which is an issue on the architectural level of partitioning not really fitting into the open table cache).
I attached a diff with the test and results (I also added handler status variables to show my point). The diff is against b4a2baffa82e5c07b96a1c752228560dcac1359b.
Here is the part of the result file that shows what I mean with extra comments prepended by MJ>: CREATE TABLE t1 (a int PRIMARY KEY) ENGINE = InnoDB PARTITION BY HASH (a) PARTITIONS 1000; INSERT INTO t1 VALUES (0), (1), (2), (3); FLUSH STATUS;
SELECT a FROM t1 PARTITION(p0) WHERE a = 0; a 0 SHOW SESSION STATUS WHERE VARIABLE_NAME LIKE 'HANDLER_%' AND VALUE > 0; Variable_name Value MJ> Here it closes all but one partitions. Handler_close 999
SELECT a FROM t1 WHERE a = 0; a 0 SHOW SESSION STATUS WHERE VARIABLE_NAME LIKE 'HANDLER_%' AND VALUE > 0; Variable_name Value Handler_close 1000 Handler_external_lock 2006 MJ> and here it re-opens those 'all but one' partitions. Handler_open 999
Regards Mattias
Best regards. HF
13.04.2018 19:07, Mattias Jonsson wrote:
Hi MariaDB Devs,
I tried to evaluate spider engine and found an issue where it
crashes,
most likely due to MDEV-11084 (Stacktrace and reproducible test case attached).
That also leads me to wonder about the performance for partitioned tables after MDEV-11084, when it seems to close partitions whenever it is not used in a statement (i.e. require it to be reopened in the next statement that would use another partition, effectively not using the open table cache).
I cannot see anything mentioned in the final commit message hinting on that it closes partitions not used in the current query, but in the previous patches it was mentioned without any reason.
Why does it close the already opened partitions? https://github.com/MariaDB/server/blob/10.3/sql/ha_partition .cc#L8365
I would not mind opening the partitions only when they are to be used (even though there are engines that need to be tested more), but closing them makes no sense to me performance wise. Also notice that the partitions first will be put back into the open table cache and then on the next query the non-used partitions will be closed and the needed ones be (re)-opened.
Regards Mattias Jonsson
-- Mattias Jonsson IT Team Lead - Development
Booking.com B.V. Herengracht 597 Amsterdam 1017 CE Netherlands <https://maps.google.com/?q=Herengracht+597+Amsterdam+1017+CE+Netherlands&entry=gmail&source=g> Direct +31207125646 [image: Booking.com] <http://www.booking.com/> The world's #1 accommodation site 43 languages, 198+ offices worldwide, 120,000+ global destinations, 1,550,000+ room nights booked every day No booking fees, best price always guaranteed Subsidiary of Booking Holdings Inc. (NASDAQ: BKNG)
-- Mattias Jonsson IT Team Lead - Development Booking.com B.V. Herengracht 597 Amsterdam 1017 CE Netherlands Direct +31207125646 [image: Booking.com] <http://www.booking.com/> The world's #1 accommodation site 43 languages, 198+ offices worldwide, 120,000+ global destinations, 1,550,000+ room nights booked every day No booking fees, best price always guaranteed Subsidiary of Booking Holdings Inc. (NASDAQ: BKNG)
participants (3)
-
Alexey Botchkov -
Jacob Mathew -
Mattias Jonsson