Does anyone know (Jordan, Denny maybe???) why Tomcat takes so long to
startup? I have a Linux server with 5 contexts on it and the Tomcat startup
time is around 150 seconds, so about 2 1/2 minutes, this seems crazy to me.
I’ve tried some of the suggestions from this page:
But to be honest they made pretty much no difference. I’m running on Tomcat
7 with Tomcat Native installed and Java 1.7. This is an install from the
Railo installer that was updated to Lucee, but the startup time has always
been like this, not just since I updated to Lucee. I’m sure when I
previously used Resin (years ago) it’s startup time was seconds, no more
than 4 or 5, not minutes.
+1 to Julian’s reply regarding Gavin’s article. The JarScanner in Tomcat can really make the contexts crawl at startup.
I’ve personally been running with it set to false for a while now and have had no issues alongside the faster start times.
I also looked at the startStopThreads attribute for each context and tested with various thread counts and also by setting its value to -1 but the gain in decreasing startup was minimal.
and was able to reduce my startup times down from around 150 seconds
to under 15.
Julian.On 29 April 2015 at 15:55, Andrew Dixon <@Andrew_Dixon> wrote:
Does anyone know (Jordan, Denny maybe???) why Tomcat takes so long to
startup? I have a Linux server with 5 contexts on it and the Tomcat startup
time is around 150 seconds, so about 2 1/2 minutes, this seems crazy to me.
Also, always check whether there is a permission issue. Sometimes Lucee tries to create files in places it has no right to do so. That happened with JPL a while ago and we had a 300s startup time there and then with the correct rights, we were able to bring it down to 6s.
Sincerely
Gert Franz
RASIA GmbH
Spittelgasse 7
5103 Moeriken-Wildegg
Email: @Gert_Franz
Skype: gert.franz
Phone Switzerland: +41 76 5680 231-----Ursprüngliche Nachricht-----
Von: lucee@googlegroups.com [mailto:lucee@googlegroups.com] Im Auftrag von Julian Halliwell
Gesendet: Mittwoch, 29. April 2015 17:07
An: lucee@googlegroups.com
Betreff: Re: [Lucee] Tomcat startup time
Hi Andrew
I followed Gavin’s instructions here:
and was able to reduce my startup times down from around 150 seconds to under 15.
Julian.
On 29 April 2015 at 15:55, Andrew Dixon <@Andrew_Dixon> wrote:
Does anyone know (Jordan, Denny maybe???) why Tomcat takes so long to
startup? I have a Linux server with 5 contexts on it and the Tomcat
startup time is around 150 seconds, so about 2 1/2 minutes, this seems crazy to me.
+1 to Julian’s reply regarding Gavin’s article. The JarScanner in Tomcat
can really make the contexts crawl at startup.
I’ve personally been running with it set to false for a while now and have
had no issues alongside the faster start times.
I also looked at the startStopThreads attribute for each context and
tested with various thread counts and also by setting its value to -1 but
the gain in decreasing startup was minimal.
This reduced my startup times on one server from 15s → 2s (15 lucee
contexts) and on another from 8s → 1s (3 lucee contexts)
skipping jar scanning should definitely be included in the installer config.On Thu, 30 Apr 2015 at 10:05 Gert Franz <@Gert_Franz> wrote:
What I will do in the next days is to start Tomcat without Lucee in order
to see where we could end at. Just to rule out a general Lucee issue…
Gert
Sent from somewhere on the road
Am 30.04.2015 um 10:51 schrieb Konstantinos Liakos <
@Konstantinos_Liakos>:
The performance on your dev machine is quite decent I think. Bear in mind
that each context has to lookup all these files in the WEB-INF/lucee
folder, so IMO the IO is pretty much the bottleneck.
For the live machine, m1 is previous generation and without ssd hard
drive. It would be expected to have bad performance, but not that much I
think.
On 30 April 2015 at 11:42, Andrew Dixon <@Andrew_Dixon> wrote:
Ok I have made the change on my production machine and it came down from
150 seconds to 50 seconds, so much better but still nearly a minute, so
still not that happy with that.
@Konstantinos - In production I have 6 contexts running on a m1.small AWS
instance. On my dev machine (Macbook Pro 2.4Ghz i5) I have 9 contexts and
it takes 20 seconds. I’m guessing therefore this is a resource issue. The
only other difference is the production machine has New Relic installed as
well, maybe this is part of the problem.
This took the time from around 60 seconds to 18 - 20 seconds, which is
better but still seems quite slow. I will try my production machine later
and report back on that.
Is this something that should be configured by default in the
installers?
Warm Regards,
Jordan Michaels
On 04/29/2015 04:12 PM, Igal @ Lucee.org wrote:
The jarsToskip makes all the difference.
+1
in Tomcat/conf/catalina.properties set:
tomcat.util.scan.StandardJarScanFilter.jarsToSkip=*.jar
Igal Sapir
Lucee Core Developer
Lucee.org <http://lucee.org/>
On 4/29/2015 11:50 AM, Konstantinos Liakos wrote:
The jarsToskip makes all the difference. Follow Gavin's
article.
--
You received this message because you are subscribed to the
--
You received this message because you are subscribed to the Google
Groups "Lucee" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to lucee+un...@googlegroups.com
<mailto:lucee%2Bu...@googlegroups.com>.
To post to this group, send email to lu...@googlegroups.com
<mailto:lu...@googlegroups.com>.
To view this discussion on the web visit
Just awesome!Am Mittwoch, 29. April 2015 16:55:50 UTC+2 schrieb Andrew Dixon:
Hi All,
Does anyone know (Jordan, Denny maybe???) why Tomcat takes so long to
startup? I have a Linux server with 5 contexts on it and the Tomcat startup
time is around 150 seconds, so about 2 1/2 minutes, this seems crazy to me.
I’ve tried some of the suggestions from this page:
But to be honest they made pretty much no difference. I’m running on
Tomcat 7 with Tomcat Native installed and Java 1.7. This is an install from
the Railo installer that was updated to Lucee, but the startup time has
always been like this, not just since I updated to Lucee. I’m sure when I
previously used Resin (years ago) it’s startup time was seconds, no more
than 4 or 5, not minutes.
Warm Regards,
Jordan MichaelsOn 04/29/2015 04:47 PM, Igal Sapir wrote:
Yep.
On Apr 29, 2015 16:40, “Jordan Michaels” <@Jordan_Michaels mailto:Jordan_Michaels> wrote:
Is this something that should be configured by default in the
installers?
Warm Regards,
Jordan Michaels
On 04/29/2015 04:12 PM, Igal @ Lucee.org wrote:
The jarsToskip makes all the difference.
+1
in Tomcat/conf/catalina.properties set:
tomcat.util.scan.StandardJarScanFilter.jarsToSkip=*.jar
Igal Sapir
Lucee Core Developer
Lucee.org <http://lucee.org/>
On 4/29/2015 11:50 AM, Konstantinos Liakos wrote:
The jarsToskip makes all the difference. Follow Gavin's article.
--
You received this message because you are subscribed to the Google
Groups "Lucee" group.
To unsubscribe from this group and stop receiving emails from
it, send
an email to lucee+unsubscribe@googlegroups.com
<mailto:lucee%2Bunsubscribe@googlegroups.com>
<mailto:lucee+unsubscribe@googlegroups.com
<mailto:lucee%2Bunsubscribe@googlegroups.com>>.
To post to this group, send email to lucee@googlegroups.com
<mailto:lucee@googlegroups.com>
<mailto:lucee@googlegroups.com <mailto:lucee@googlegroups.com>>.
To view this discussion on the web visit
https://groups.google.com/d/msgid/lucee/55416552.9040304%40lucee.org
<https://groups.google.com/d/msgid/lucee/55416552.9040304%40lucee.org?utm_medium=email&utm_source=footer>.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google
Groups "Lucee" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to lucee+unsubscribe@googlegroups.com
<mailto:lucee%2Bunsubscribe@googlegroups.com>.
To post to this group, send email to lucee@googlegroups.com
<mailto:lucee@googlegroups.com>.
To view this discussion on the web visit
https://groups.google.com/d/msgid/lucee/55416BEF.5000502%40viviotech.net.
For more options, visit https://groups.google.com/d/optout.
Ok I have made the change on my production machine and it came down from
150 seconds to 50 seconds, so much better but still nearly a minute, so
still not that happy with that.
@Konstantinos - In production I have 6 contexts running on a m1.small AWS
instance. On my dev machine (Macbook Pro 2.4Ghz i5) I have 9 contexts and
it takes 20 seconds. I’m guessing therefore this is a resource issue. The
only other difference is the production machine has New Relic installed as
well, maybe this is part of the problem.
This took the time from around 60 seconds to 18 - 20 seconds, which is
better but still seems quite slow. I will try my production machine later
and report back on that.
Is this something that should be configured by default in the
installers?
Warm Regards,
Jordan Michaels
On 04/29/2015 04:12 PM, Igal @ Lucee.org wrote:
The jarsToskip makes all the difference.
+1
in Tomcat/conf/catalina.properties set:
tomcat.util.scan.StandardJarScanFilter.jarsToSkip=*.jar
Igal Sapir
Lucee Core Developer
Lucee.org <http://lucee.org/>
On 4/29/2015 11:50 AM, Konstantinos Liakos wrote:
The jarsToskip makes all the difference. Follow Gavin's
article.
--
You received this message because you are subscribed to the
--
You received this message because you are subscribed to the Google
Groups "Lucee" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to lucee+un...@googlegroups.com
<mailto:lucee%2Bu...@googlegroups.com>.
To post to this group, send email to lu...@googlegroups.com
<mailto:lu...@googlegroups.com>.
To view this discussion on the web visit
While we’re talking I/O, and since Vagrant is becoming more prevalent, I
wanted to point out how much difference one’s method of folder sharing can
make:
Spoiler: NFS reduced my startup times from 390s to 45s (single context, and
yes, skipping the scanned jars).
It is previous generation, but it is running on SSD, but if IO is the
bottleneck then I can see why. I’m going to upgrade to a newer generation
machine once my reservation is up in a couple of months.
On 30 April 2015 at 09:51, Konstantinos Liakos <@Konstantinos_Liakos wrote:
The performance on your dev machine is quite decent I think. Bear in mind
that each context has to lookup all these files in the WEB-INF/lucee
folder, so IMO the IO is pretty much the bottleneck.
For the live machine, m1 is previous generation and without ssd hard
drive. It would be expected to have bad performance, but not that much I
think.
On 30 April 2015 at 11:42, Andrew Dixon <@Andrew_Dixon> wrote:
Ok I have made the change on my production machine and it came down from
150 seconds to 50 seconds, so much better but still nearly a minute, so
still not that happy with that.
@Konstantinos - In production I have 6 contexts running on a m1.small
AWS instance. On my dev machine (Macbook Pro 2.4Ghz i5) I have 9 contexts
and it takes 20 seconds. I’m guessing therefore this is a resource issue.
The only other difference is the production machine has New Relic installed
as well, maybe this is part of the problem.
This took the time from around 60 seconds to 18 - 20 seconds, which is
better but still seems quite slow. I will try my production machine later
and report back on that.
Is this something that should be configured by default in the
installers?
Warm Regards,
Jordan Michaels
On 04/29/2015 04:12 PM, Igal @ Lucee.org wrote:
The jarsToskip makes all the difference.
+1
in Tomcat/conf/catalina.properties set:
tomcat.util.scan.StandardJarScanFilter.jarsToSkip=*.jar
Igal Sapir
Lucee Core Developer
Lucee.org <http://lucee.org/>
On 4/29/2015 11:50 AM, Konstantinos Liakos wrote:
The jarsToskip makes all the difference. Follow Gavin's
article.
--
You received this message because you are subscribed to the
It is previous generation, but it is running on SSD, but if IO is the
bottleneck then I can see why. I’m going to upgrade to a newer generation
machine once my reservation is up in a couple of months.
The performance on your dev machine is quite decent I think. Bear in mind
that each context has to lookup all these files in the WEB-INF/lucee
folder, so IMO the IO is pretty much the bottleneck.
For the live machine, m1 is previous generation and without ssd hard
drive. It would be expected to have bad performance, but not that much I
think.
On 30 April 2015 at 11:42, Andrew Dixon <@Andrew_Dixon> wrote:
Ok I have made the change on my production machine and it came down from
150 seconds to 50 seconds, so much better but still nearly a minute, so
still not that happy with that.
@Konstantinos - In production I have 6 contexts running on a m1.small AWS
instance. On my dev machine (Macbook Pro 2.4Ghz i5) I have 9 contexts and
it takes 20 seconds. I’m guessing therefore this is a resource issue. The
only other difference is the production machine has New Relic installed as
well, maybe this is part of the problem.
This took the time from around 60 seconds to 18 - 20 seconds, which is
better but still seems quite slow. I will try my production machine later
and report back on that.
Is this something that should be configured by default in the
installers?
Warm Regards,
Jordan Michaels
On 04/29/2015 04:12 PM, Igal @ Lucee.org wrote:
The jarsToskip makes all the difference.
+1
in Tomcat/conf/catalina.properties set:
tomcat.util.scan.StandardJarScanFilter.jarsToSkip=*.jar
Igal Sapir
Lucee Core Developer
Lucee.org <http://lucee.org/>
On 4/29/2015 11:50 AM, Konstantinos Liakos wrote:
The jarsToskip makes all the difference. Follow Gavin's
article.
--
You received this message because you are subscribed to the
--
You received this message because you are subscribed to the Google
Groups "Lucee" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to lucee+un...@googlegroups.com
<mailto:lucee%2Bu...@googlegroups.com>.
To post to this group, send email to lu...@googlegroups.com
<mailto:lu...@googlegroups.com>.
To view this discussion on the web visit
What I will do in the next days is to start Tomcat without Lucee in order to see where we could end at. Just to rule out a general Lucee issue…
GertSent from somewhere on the road
Am 30.04.2015 um 10:51 schrieb Konstantinos Liakos <@Konstantinos_Liakos>:
The performance on your dev machine is quite decent I think. Bear in mind that each context has to lookup all these files in the WEB-INF/lucee folder, so IMO the IO is pretty much the bottleneck.
For the live machine, m1 is previous generation and without ssd hard drive. It would be expected to have bad performance, but not that much I think.
On 30 April 2015 at 11:42, Andrew Dixon <@Andrew_Dixon> wrote:
Ok I have made the change on my production machine and it came down from 150 seconds to 50 seconds, so much better but still nearly a minute, so still not that happy with that.
@Konstantinos - In production I have 6 contexts running on a m1.small AWS instance. On my dev machine (Macbook Pro 2.4Ghz i5) I have 9 contexts and it takes 20 seconds. I’m guessing therefore this is a resource issue. The only other difference is the production machine has New Relic installed as well, maybe this is part of the problem.
On 30 April 2015 at 09:13, Konstantinos Liakos <@Konstantinos_Liakos> wrote:
How many contexts do you have? For 5-6 contexts this time is fairly good for a medium range dev machine.
On Thursday, 30 April 2015 11:10:47 UTC+3, Andrew Dixon wrote:
Not tried this on my production machine yet, but on my dev machine, running Tomcat 7.0.59 I had to change three properties, as Gavin describes:
This took the time from around 60 seconds to 18 - 20 seconds, which is better but still seems quite slow. I will try my production machine later and report back on that.
Is this something that should be configured by default in the
installers?
Warm Regards,
Jordan Michaels
On 04/29/2015 04:12 PM, Igal @ Lucee.org wrote:
The jarsToskip makes all the difference.
+1
in Tomcat/conf/catalina.properties set:
tomcat.util.scan.StandardJarScanFilter.jarsToSkip=*.jar
Igal Sapir
Lucee Core Developer
Lucee.org <http://lucee.org/>
On 4/29/2015 11:50 AM, Konstantinos Liakos wrote:
The jarsToskip makes all the difference. Follow Gavin's article.
--
You received this message because you are subscribed to the Google
Groups "Lucee" group.
To unsubscribe from this group and stop receiving emails from
it, send
an email to lucee+un...@googlegroups.com
<mailto:lucee%2Bu...@googlegroups.com>
<mailto:lucee+un...@googlegroups.com
<mailto:lucee%2Bu...@googlegroups.com>>.
To post to this group, send email to lu...@googlegroups.com
<mailto:lu...@googlegroups.com>
<mailto:lu...@googlegroups.com <mailto:lu...@googlegroups.com>>.
To view this discussion on the web visit
https://groups.google.com/d/msgid/lucee/55416552.9040304%40lucee.org
<https://groups.google.com/d/msgid/lucee/55416552.9040304%40lucee.org?utm_medium=email&utm_source=footer>.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google
Groups "Lucee" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to lucee+un...@googlegroups.com
<mailto:lucee%2Bu...@googlegroups.com>.
To post to this group, send email to lu...@googlegroups.com
<mailto:lu...@googlegroups.com>.
To view this discussion on the web visit
https://groups.google.com/d/msgid/lucee/55416BEF.5000502%40viviotech.net.
For more options, visit https://groups.google.com/d/optout.
How many contexts do you have? For 5-6 contexts this time is fairly good
for a medium range dev machine.On Thursday, 30 April 2015 11:10:47 UTC+3, Andrew Dixon wrote:
Not tried this on my production machine yet, but on my dev machine,
running Tomcat 7.0.59 I had to change three properties, as Gavin describes:
This took the time from around 60 seconds to 18 - 20 seconds, which is
better but still seems quite slow. I will try my production machine later
and report back on that.
Is this something that should be configured by default in the
installers?
Warm Regards,
Jordan Michaels
On 04/29/2015 04:12 PM, Igal @ Lucee.org wrote:
The jarsToskip makes all the difference.
+1
in Tomcat/conf/catalina.properties set:
tomcat.util.scan.StandardJarScanFilter.jarsToSkip=*.jar
Igal Sapir
Lucee Core Developer
Lucee.org <http://lucee.org/>
On 4/29/2015 11:50 AM, Konstantinos Liakos wrote:
The jarsToskip makes all the difference. Follow Gavin's
article.
--
You received this message because you are subscribed to the
--
You received this message because you are subscribed to the Google
Groups "Lucee" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to lucee+un...@googlegroups.com <javascript:>
<mailto:lucee%2Bu...@googlegroups.com <javascript:>>.
To post to this group, send email to lu...@googlegroups.com
This took the time from around 60 seconds to 18 - 20 seconds, which is
better but still seems quite slow. I will try my production machine later
and report back on that.
On Apr 29, 2015 16:40, “Jordan Michaels” <@Jordan_Michaels mailto:Jordan_Michaels> wrote:
Is this something that should be configured by default in the
installers?
Warm Regards,
Jordan Michaels
On 04/29/2015 04:12 PM, Igal @ Lucee.org wrote:
The jarsToskip makes all the difference.
+1
in Tomcat/conf/catalina.properties set:
tomcat.util.scan.StandardJarScanFilter.jarsToSkip=*.jar
Igal Sapir
Lucee Core Developer
Lucee.org <http://lucee.org/>
On 4/29/2015 11:50 AM, Konstantinos Liakos wrote:
The jarsToskip makes all the difference. Follow Gavin's
article.
--
You received this message because you are subscribed to the Google
Groups "Lucee" group.
To unsubscribe from this group and stop receiving emails from
it, send
an email to lucee+unsubscribe@googlegroups.com
<mailto:lucee%2Bunsubscribe@googlegroups.com>
<mailto:lucee+unsubscribe@googlegroups.com
<mailto:lucee%2Bunsubscribe@googlegroups.com>>.
To post to this group, send email to lucee@googlegroups.com
<mailto:lucee@googlegroups.com>
<mailto:lucee@googlegroups.com <mailto:lucee@googlegroups.com>>.
To view this discussion on the web visit
--
You received this message because you are subscribed to the Google
Groups "Lucee" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to lucee+unsubscribe@googlegroups.com
<mailto:lucee%2Bunsubscribe@googlegroups.com>.
To post to this group, send email to lucee@googlegroups.com
<mailto:lucee@googlegroups.com>.
To view this discussion on the web visit
The performance on your dev machine is quite decent I think. Bear in mind
that each context has to lookup all these files in the WEB-INF/lucee
folder, so IMO the IO is pretty much the bottleneck.
For the live machine, m1 is previous generation and without ssd hard drive.
It would be expected to have bad performance, but not that much I think.On 30 April 2015 at 11:42, Andrew Dixon <@Andrew_Dixon> wrote:
Ok I have made the change on my production machine and it came down from
150 seconds to 50 seconds, so much better but still nearly a minute, so
still not that happy with that.
@Konstantinos - In production I have 6 contexts running on a m1.small AWS
instance. On my dev machine (Macbook Pro 2.4Ghz i5) I have 9 contexts and
it takes 20 seconds. I’m guessing therefore this is a resource issue. The
only other difference is the production machine has New Relic installed as
well, maybe this is part of the problem.
This took the time from around 60 seconds to 18 - 20 seconds, which is
better but still seems quite slow. I will try my production machine later
and report back on that.
Is this something that should be configured by default in the
installers?
Warm Regards,
Jordan Michaels
On 04/29/2015 04:12 PM, Igal @ Lucee.org wrote:
The jarsToskip makes all the difference.
+1
in Tomcat/conf/catalina.properties set:
tomcat.util.scan.StandardJarScanFilter.jarsToSkip=*.jar
Igal Sapir
Lucee Core Developer
Lucee.org <http://lucee.org/>
On 4/29/2015 11:50 AM, Konstantinos Liakos wrote:
The jarsToskip makes all the difference. Follow Gavin's
article.
--
You received this message because you are subscribed to the
--
You received this message because you are subscribed to the Google
Groups "Lucee" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to lucee+un...@googlegroups.com
<mailto:lucee%2Bu...@googlegroups.com>.
To post to this group, send email to lu...@googlegroups.com
<mailto:lu...@googlegroups.com>.
To view this discussion on the web visit
Just so you know: in the next version of mod_cfml, the jar scanning for every new web context will be disabled by default.
To be continued…
Kind regards,
Paul KlinkenbergOp 30 apr. 2015, om 13:59 heeft Patrick Heppler <@Patrick_Heppler> het volgende geschreven:
Great article. Result:
INFORMATION: Server startup in 4089 ms
Just awesome!
Am Mittwoch, 29. April 2015 16:55:50 UTC+2 schrieb Andrew Dixon:
Hi All,
Does anyone know (Jordan, Denny maybe???) why Tomcat takes so long to startup? I have a Linux server with 5 contexts on it and the Tomcat startup time is around 150 seconds, so about 2 1/2 minutes, this seems crazy to me. I’ve tried some of the suggestions from this page:
But to be honest they made pretty much no difference. I’m running on Tomcat 7 with Tomcat Native installed and Java 1.7. This is an install from the Railo installer that was updated to Lucee, but the startup time has always been like this, not just since I updated to Lucee. I’m sure when I previously used Resin (years ago) it’s startup time was seconds, no more than 4 or 5, not minutes.