We replace "reserved_memory= 8192" with "reserved_memory = 1024" in /etc/lvm/lvm.conf used by "lvm2", this can save 7MB peak memory consumption, so we lower the possibility of kdump OOM.
For kdump, we don't have too many lvm targets, lvm2 locates in the RAM(rootfs), so don't need that much memory, as discussed with lvm people, they agree that we can use 1MB under kdump as long as there are not so many lvm targets invloved.
We modify /etc/lvm/lvm.conf when "99kdumpbase" install() is executed, because it is parsed after "90lvm" by dracut.
Signed-off-by: Xunlei Pang xlpang@redhat.com --- dracut-module-setup.sh | 5 +++++ 1 file changed, 5 insertions(+)
diff --git a/dracut-module-setup.sh b/dracut-module-setup.sh index 1f96bb8..674f3ee 100755 --- a/dracut-module-setup.sh +++ b/dracut-module-setup.sh @@ -743,4 +743,9 @@ install() { # target. Ideally all this should be pushed into dracut iscsi module # at some point of time. kdump_check_iscsi_targets + + # For the lvm type target under kdump, in /etc/lvm/lvm.conf we can safely replace + # "reserved_memory = 8192" with "reserved_memory = 1024" to lower memory pressure + sed -i -e 's/(^[[:space:]]*)reserved_memory[[:space:]]*=[[:space:]]*[[:digit:]]*/ \ + \1reserved_memory = 1024/' ${initdir}/etc/lvm/lvm.conf &>/dev/null }
Hi Xunlei,
On 03/29/17 at 02:38pm, Xunlei Pang wrote:
We replace "reserved_memory= 8192" with "reserved_memory = 1024" in /etc/lvm/lvm.conf used by "lvm2", this can save 7MB peak memory consumption, so we lower the possibility of kdump OOM.
For kdump, we don't have too many lvm targets, lvm2 locates in the RAM(rootfs), so don't need that much memory, as discussed with lvm people, they agree that we can use 1MB under kdump as long as there are not so many lvm targets invloved.
We modify /etc/lvm/lvm.conf when "99kdumpbase" install() is executed, because it is parsed after "90lvm" by dracut.
Signed-off-by: Xunlei Pang xlpang@redhat.com
dracut-module-setup.sh | 5 +++++ 1 file changed, 5 insertions(+)
diff --git a/dracut-module-setup.sh b/dracut-module-setup.sh index 1f96bb8..674f3ee 100755 --- a/dracut-module-setup.sh +++ b/dracut-module-setup.sh @@ -743,4 +743,9 @@ install() { # target. Ideally all this should be pushed into dracut iscsi module # at some point of time. kdump_check_iscsi_targets
- # For the lvm type target under kdump, in /etc/lvm/lvm.conf we can safely replace
- # "reserved_memory = 8192" with "reserved_memory = 1024" to lower memory pressure
The code is setting reserved_memory = 1024 unconditionally, not only for relacing 8192 with 1024, need update the commit to match the code. Also the line is too long as well.
- sed -i -e 's/(^[[:space:]]*)reserved_memory[[:space:]]*=[[:space:]]*[[:digit:]]*/ \
\1reserved_memory = 1024/' ${initdir}/etc/lvm/lvm.conf &>/dev/null
Nitpick, this can be changed to 80 columns alignment, like move sed -i -e in a seperate line.
}
1.8.3.1 _______________________________________________ kexec mailing list -- kexec@lists.fedoraproject.org To unsubscribe send an email to kexec-leave@lists.fedoraproject.org
Thanks Dave
On 03/29/2017 at 03:10 PM, Dave Young wrote:
Hi Xunlei,
On 03/29/17 at 02:38pm, Xunlei Pang wrote:
We replace "reserved_memory= 8192" with "reserved_memory = 1024" in /etc/lvm/lvm.conf used by "lvm2", this can save 7MB peak memory consumption, so we lower the possibility of kdump OOM.
For kdump, we don't have too many lvm targets, lvm2 locates in the RAM(rootfs), so don't need that much memory, as discussed with lvm people, they agree that we can use 1MB under kdump as long as there are not so many lvm targets invloved.
We modify /etc/lvm/lvm.conf when "99kdumpbase" install() is executed, because it is parsed after "90lvm" by dracut.
Signed-off-by: Xunlei Pang xlpang@redhat.com
dracut-module-setup.sh | 5 +++++ 1 file changed, 5 insertions(+)
diff --git a/dracut-module-setup.sh b/dracut-module-setup.sh index 1f96bb8..674f3ee 100755 --- a/dracut-module-setup.sh +++ b/dracut-module-setup.sh @@ -743,4 +743,9 @@ install() { # target. Ideally all this should be pushed into dracut iscsi module # at some point of time. kdump_check_iscsi_targets
- # For the lvm type target under kdump, in /etc/lvm/lvm.conf we can safely replace
- # "reserved_memory = 8192" with "reserved_memory = 1024" to lower memory pressure
The code is setting reserved_memory = 1024 unconditionally, not only for relacing 8192 with 1024, need update the commit to match the code. Also the line is too long as well.
It only does replacing actually, but if lvm.conf doesn't exist, the code actually does nothing.
- sed -i -e 's/(^[[:space:]]*)reserved_memory[[:space:]]*=[[:space:]]*[[:digit:]]*/ \
\1reserved_memory = 1024/' ${initdir}/etc/lvm/lvm.conf &>/dev/nullNitpick, this can be changed to 80 columns alignment, like move sed -i -e in a seperate line.
ok, but it's a little hard for shell, we have many code very long including dracut, what about the following, I make each independent part one line? sed -i -e \ 's/(^[[:space:]]*)reserved_memory[[:space:]]*=[[:space:]]*[[:digit:]]*/\1reserved_memory = 1024/' \ ${initdir}/etc/lvm/lvm.conf &>/dev/null
Regards, Xunlei
}
1.8.3.1 _______________________________________________ kexec mailing list -- kexec@lists.fedoraproject.org To unsubscribe send an email to kexec-leave@lists.fedoraproject.org
Thanks Dave
On 03/29/17 at 03:26pm, Xunlei Pang wrote:
On 03/29/2017 at 03:10 PM, Dave Young wrote:
Hi Xunlei,
On 03/29/17 at 02:38pm, Xunlei Pang wrote:
We replace "reserved_memory= 8192" with "reserved_memory = 1024" in /etc/lvm/lvm.conf used by "lvm2", this can save 7MB peak memory consumption, so we lower the possibility of kdump OOM.
For kdump, we don't have too many lvm targets, lvm2 locates in the RAM(rootfs), so don't need that much memory, as discussed with lvm people, they agree that we can use 1MB under kdump as long as there are not so many lvm targets invloved.
We modify /etc/lvm/lvm.conf when "99kdumpbase" install() is executed, because it is parsed after "90lvm" by dracut.
Signed-off-by: Xunlei Pang xlpang@redhat.com
dracut-module-setup.sh | 5 +++++ 1 file changed, 5 insertions(+)
diff --git a/dracut-module-setup.sh b/dracut-module-setup.sh index 1f96bb8..674f3ee 100755 --- a/dracut-module-setup.sh +++ b/dracut-module-setup.sh @@ -743,4 +743,9 @@ install() { # target. Ideally all this should be pushed into dracut iscsi module # at some point of time. kdump_check_iscsi_targets
- # For the lvm type target under kdump, in /etc/lvm/lvm.conf we can safely replace
- # "reserved_memory = 8192" with "reserved_memory = 1024" to lower memory pressure
The code is setting reserved_memory = 1024 unconditionally, not only for relacing 8192 with 1024, need update the commit to match the code. Also the line is too long as well.
It only does replacing actually, but if lvm.conf doesn't exist, the code actually does nothing.
Yes, I got that, what I means is if someone set it as reserved_memory=4096 we still set it to 1024, so the comment just say 8192 -> 1024. We can just say we will unconditionally set it as 1024.
- sed -i -e 's/(^[[:space:]]*)reserved_memory[[:space:]]*=[[:space:]]*[[:digit:]]*/ \
\1reserved_memory = 1024/' ${initdir}/etc/lvm/lvm.conf &>/dev/nullNitpick, this can be changed to 80 columns alignment, like move sed -i -e in a seperate line.
ok, but it's a little hard for shell, we have many code very long including dracut, what about the following, I make each independent part one line? sed -i -e \ 's/(^[[:space:]]*)reserved_memory[[:space:]]*=[[:space:]]*[[:digit:]]*/\1reserved_memory = 1024/' \ ${initdir}/etc/lvm/lvm.conf &>/dev/null
Yes, it is better, maybe less indent for the last two lines, it will be within 80 columns
Regards, Xunlei
}
1.8.3.1 _______________________________________________ kexec mailing list -- kexec@lists.fedoraproject.org To unsubscribe send an email to kexec-leave@lists.fedoraproject.org
Thanks Dave
kexec mailing list -- kexec@lists.fedoraproject.org To unsubscribe send an email to kexec-leave@lists.fedoraproject.org
On 03/29/2017 at 03:51 PM, Dave Young wrote:
On 03/29/17 at 03:26pm, Xunlei Pang wrote:
On 03/29/2017 at 03:10 PM, Dave Young wrote:
Hi Xunlei,
On 03/29/17 at 02:38pm, Xunlei Pang wrote:
We replace "reserved_memory= 8192" with "reserved_memory = 1024" in /etc/lvm/lvm.conf used by "lvm2", this can save 7MB peak memory consumption, so we lower the possibility of kdump OOM.
For kdump, we don't have too many lvm targets, lvm2 locates in the RAM(rootfs), so don't need that much memory, as discussed with lvm people, they agree that we can use 1MB under kdump as long as there are not so many lvm targets invloved.
We modify /etc/lvm/lvm.conf when "99kdumpbase" install() is executed, because it is parsed after "90lvm" by dracut.
Signed-off-by: Xunlei Pang xlpang@redhat.com
dracut-module-setup.sh | 5 +++++ 1 file changed, 5 insertions(+)
diff --git a/dracut-module-setup.sh b/dracut-module-setup.sh index 1f96bb8..674f3ee 100755 --- a/dracut-module-setup.sh +++ b/dracut-module-setup.sh @@ -743,4 +743,9 @@ install() { # target. Ideally all this should be pushed into dracut iscsi module # at some point of time. kdump_check_iscsi_targets
- # For the lvm type target under kdump, in /etc/lvm/lvm.conf we can safely replace
- # "reserved_memory = 8192" with "reserved_memory = 1024" to lower memory pressure
The code is setting reserved_memory = 1024 unconditionally, not only for relacing 8192 with 1024, need update the commit to match the code. Also the line is too long as well.
It only does replacing actually, but if lvm.conf doesn't exist, the code actually does nothing.
Yes, I got that, what I means is if someone set it as reserved_memory=4096 we still set it to 1024, so the comment just say 8192 -> 1024. We can just say we will unconditionally set it as 1024.
Indeed, will update soon.
Regards, Xunlei
-
Dave Young -
Xunlei Pang -
Xunlei Pang