Hi all, I've run into a behaviour that I would like to understand and possibly disable. Two times per day the mserver daeomon starts trashing the disk furiously, I see that the size on disk of data directory goes up to 115 GB from 85 GB and then comes back again to the same initial size, during this process disk utilization goes trough the roof. Can anyone tell me what is monetdb doing? There are no import/export process running at this times, it seems that monetdb is doing some kind of data compaction... is there anyway to disable this? $ mserver5 --version MonetDB 5 server v11.15.7 "Feb2013-SP2" (64-bit, 64-bit oids) Thanks! -- Luis Neves
Hi Can you tell me which Linux version you are using. We have seen such behavior in some Fedora flavours and seems related to hugepagemap settings in the OS. regards, Martin On 17/01/14 19:13, Luis Neves wrote:
Hi all,
I've run into a behaviour that I would like to understand and possibly disable. Two times per day the mserver daeomon starts trashing the disk furiously, I see that the size on disk of data directory goes up to 115 GB from 85 GB and then comes back again to the same initial size, during this process disk utilization goes trough the roof. Can anyone tell me what is monetdb doing? There are no import/export process running at this times, it seems that monetdb is doing some kind of data compaction... is there anyway to disable this?
$ mserver5 --version MonetDB 5 server v11.15.7 "Feb2013-SP2" (64-bit, 64-bit oids)
Thanks!
Hello! On 01/17/2014 07:12 PM, Martin Kersten wrote:
Can you tell me which Linux version you are using. We have seen such behavior in some Fedora flavours and seems related to hugepagemap settings in the OS.
This a Debian 6 server. $ lsb_release -a No LSB modules are available. Distributor ID: Debian Description: Debian GNU/Linux 6.0.4 (squeeze) Release: 6.0.4 Codename: squeeze $ uname -a Linux golias-kpi01 2.6.32-5-amd64 #1 SMP Mon Jan 16 16:22:28 UTC 2012 x86_64 GNU/Linux Do you remember what is the problematic setting? This is what the server reports: # cat /proc/meminfo | grep Huge HugePages_Total: 0 HugePages_Free: 0 HugePages_Rsvd: 0 HugePages_Surp: 0 Hugepagesize: 2048 kB # cat /proc/sys/vm/nr_hugepages 0 # cat /sys/devices/system/node/node*/meminfo | grep Huge Node 0 HugePages_Total: 0 Node 0 HugePages_Free: 0 Node 0 HugePages_Surp: 0 Node 1 HugePages_Total: 0 Node 1 HugePages_Free: 0 Node 1 HugePages_Surp: 0 Thanks! -- Luis Neves
Hello! This issue is killing us... is there anyway to disable this behaviour? Will upgrading to the latest MonetDB version help? Thanks! On 01/17/2014 07:12 PM, Martin Kersten wrote:
Hi
Can you tell me which Linux version you are using. We have seen such behavior in some Fedora flavours and seems related to hugepagemap settings in the OS.
regards, Martin
On 17/01/14 19:13, Luis Neves wrote:
Hi all,
I've run into a behaviour that I would like to understand and possibly disable. Two times per day the mserver daeomon starts trashing the disk furiously, I see that the size on disk of data directory goes up to 115 GB from 85 GB and then comes back again to the same initial size, during this process disk utilization goes trough the roof. Can anyone tell me what is monetdb doing? There are no import/export process running at this times, it seems that monetdb is doing some kind of data compaction... is there anyway to disable this?
$ mserver5 --version MonetDB 5 server v11.15.7 "Feb2013-SP2" (64-bit, 64-bit oids)
Thanks!
_______________________________________________ users-list mailing list users-list@monetdb.org https://www.monetdb.org/mailman/listinfo/users-list
-- Luis Neves
participants (2)
-
Luis Neves -
Martin Kersten