ZEO clients eating up / filesystem
Hi, We have a ZEO setup with a Plone based intranet, and for some reason it seems that Zope is filling up space in the root filesystem. I have to restart Zope twice a day to clear it up (restarting Zope clears up the space). lsof does not list anything open in the root filesystem and swapping is done in the swap partition. Has anyone observed the same and found a solution? System details: - FreeBSD 4.10-RELEASE - Pyhton 2.3.4 - Zope 2.7.3 - SOFTWARE_HOME /usr/local/zope/lib/python - ZOPE_HOME /usr/local/zope - INSTANCE_HOME /usr/local/WWW/intranet/zeo_client1 (2 CPU FreeBSD machines running two ZEO instances) Filesystem Size Used Avail Capacity Mounted on /dev/da0s1a 1032M 207M 742M 22% / /dev/da0s1e 4.1G 519M 3.3G 14% /usr /dev/da0s1g 133G 1.3G 121G 1% /usr/local /dev/da0s1f 4.1G 53M 3.7G 1% /var procfs 4.1K 4.1K 0B 100% /proc This is the zope startup script in /usr/local/etc/rc.d #!/bin/sh # Start or stop zope # $FreeBSD: ports/www/zope/files/zope.sh,v 1.4 2004/11/16 00:02:25 pav Exp $ # PROVIDE: zope # REQUIRE: DAEMON # BEFORE: LOGIN # KEYWORD: FreeBSD shutdown # prefix=/usr/local # Define these zope_* variables in one of these files: # /etc/rc.conf # /etc/rc.conf.local # /etc/rc.conf.d/zope # # DO NOT CHANGE THESE DEFAULT VALUES HERE # zope_enable=${zope_enable:-"NO"} # Enable gkrellmd zope_instances=${zope_instances:-""} # List of instancehome dirs . /usr/local/etc/rc.subr name="zope" rcvar=`set_rcvar` load_rc_config $name if checkyesno zope_enable; then case "$1" in start) echo "Starting Zope" ;; stop) echo "Stopping Zope" ;; restart) echo "Restarting Zope" ;; *) echo "Unknown action \"$1\"" ;; esac last_zope_instance=`echo $zope_instances | /usr/bin/awk {'print $NF'}` for instance in $zope_instances do if [ -r ${instance}/etc/${name}.conf -a -x ${instance}/bin/zopectl ]; then echo -n " Instance ${instance} -> " ${instance}/bin/zopectl $1 if [ $instance != $last_zope_instance ]; then echo "sleeping 40 seconds between instances to prevent read conflict errors during startup" sleep 40 fi fi done fi -- __________________________________________________ "Nothing is as subjective as reality" Reinoud van Leeuwen reinoud.v@n.leeuwen.net http://www.xs4all.nl/~reinoud __________________________________________________
Perhaps your FS grows because ZEO write the cache files to /tmp (which is on your root filesystem)? -aj --On 30. September 2005 14:06:16 +0200 Reinoud van Leeuwen <reinoud.v@n.leeuwen.net> wrote:
Hi,
We have a ZEO setup with a Plone based intranet, and for some reason it seems that Zope is filling up space in the root filesystem. I have to restart Zope twice a day to clear it up (restarting Zope clears up the space). lsof does not list anything open in the root filesystem and swapping is done in the swap partition.
Has anyone observed the same and found a solution?
System details: - FreeBSD 4.10-RELEASE - Pyhton 2.3.4 - Zope 2.7.3 - SOFTWARE_HOME /usr/local/zope/lib/python - ZOPE_HOME /usr/local/zope - INSTANCE_HOME /usr/local/WWW/intranet/zeo_client1 (2 CPU FreeBSD machines running two ZEO instances)
Filesystem Size Used Avail Capacity Mounted on /dev/da0s1a 1032M 207M 742M 22% / /dev/da0s1e 4.1G 519M 3.3G 14% /usr /dev/da0s1g 133G 1.3G 121G 1% /usr/local /dev/da0s1f 4.1G 53M 3.7G 1% /var procfs 4.1K 4.1K 0B 100% /proc
This is the zope startup script in /usr/local/etc/rc.d
# !/bin/sh # Start or stop zope # $FreeBSD: ports/www/zope/files/zope.sh,v 1.4 2004/11/16 00:02:25 pav Exp # $ PROVIDE: zope # REQUIRE: DAEMON # BEFORE: LOGIN # KEYWORD: FreeBSD shutdown # prefix=/usr/local # Define these zope_* variables in one of these files: # /etc/rc.conf # /etc/rc.conf.local # /etc/rc.conf.d/zope # # DO NOT CHANGE THESE DEFAULT VALUES HERE # zope_enable=${zope_enable:-"NO"} # Enable gkrellmd zope_instances=${zope_instances:-""} # List of instancehome dirs . /usr/local/etc/rc.subr name="zope" rcvar=`set_rcvar` load_rc_config $name if checkyesno zope_enable; then
case "$1" in start) echo "Starting Zope" ;; stop) echo "Stopping Zope" ;; restart) echo "Restarting Zope" ;; *) echo "Unknown action \"$1\"" ;; esac
last_zope_instance=`echo $zope_instances | /usr/bin/awk {'print $NF'}` for instance in $zope_instances do if [ -r ${instance}/etc/${name}.conf -a -x ${instance}/bin/zopectl ]; then echo -n " Instance ${instance} -> " ${instance}/bin/zopectl $1 if [ $instance != $last_zope_instance ]; then echo "sleeping 40 seconds between instances to prevent read conflict errors during startup" sleep 40 fi fi done fi
-- __________________________________________________ "Nothing is as subjective as reality" Reinoud van Leeuwen reinoud.v@n.leeuwen.net http://www.xs4all.nl/~reinoud __________________________________________________ _______________________________________________ Zope maillist - Zope@zope.org http://mail.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope-dev )
On Fri, Sep 30, 2005 at 02:13:34PM +0200, Andreas Jung wrote:
Perhaps your FS grows because ZEO write the cache files to /tmp (which is on your root filesystem)?
No I checked that. Nothing in /tmp (and that would have been listed by lsof) -- __________________________________________________ "Nothing is as subjective as reality" Reinoud van Leeuwen reinoud.v@n.leeuwen.net http://www.xs4all.nl/~reinoud __________________________________________________
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Reinoud van Leeuwen wrote:
On Fri, Sep 30, 2005 at 02:13:34PM +0200, Andreas Jung wrote:
Perhaps your FS grows because ZEO write the cache files to /tmp (which is on your root filesystem)?
No I checked that. Nothing in /tmp (and that would have been listed by lsof)
Unless you make them "persistent", ZEO cache files are unlinked from their directory immediately after creation. Try looking at /proc/<zope's PID>/fd, and see if there aren't a couple of otherwise inaccessible, and *large*, files there. Tres. - -- =================================================================== Tres Seaver +1 202-558-7113 tseaver@palladion.com Palladion Software "Excellence by Design" http://palladion.com -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.5 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFDPTeh+gerLs4ltQ4RAmJyAJ9pk/jx4f5URs21XllHe4e0MZ3VpQCgtUjo 0KnkUVQmLU7rG9rM+BQDOOg= =T8HA -----END PGP SIGNATURE-----
On Fri, Sep 30, 2005 at 09:03:29AM -0400, Tres Seaver wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
Reinoud van Leeuwen wrote:
On Fri, Sep 30, 2005 at 02:13:34PM +0200, Andreas Jung wrote:
Perhaps your FS grows because ZEO write the cache files to /tmp (which is on your root filesystem)?
No I checked that. Nothing in /tmp (and that would have been listed by lsof)
Unless you make them "persistent", ZEO cache files are unlinked from their directory immediately after creation. Try looking at /proc/<zope's PID>/fd, and see if there aren't a couple of otherwise inaccessible, and *large*, files there.
Thanks. Setting "zeo-client-name" in the main section of zope.conf and "client" in the <zodb_db main> section seems to have solved this. -- __________________________________________________ "Nothing is as subjective as reality" Reinoud van Leeuwen reinoud.v@n.leeuwen.net http://www.xs4all.nl/~reinoud __________________________________________________
participants (3)
-
Andreas Jung -
Reinoud van Leeuwen -
Tres Seaver