Hey everyone -- I was wondering about getting htop installed throughout the infrastructure, as I think this will help with on-the-fly checking of issues. For example, if load average is high, it is very easy using htop to sort by CPU, and see exactly what is causing the issue. If RAM usage is abnormal, it's easy to sort by that. It also has some other features that could prove useful.
A prime example of when this could have been useful was when hosted01's load averages went up to near 60 last night. It would have made finding the issue quite a bit faster for me.
I asked on IRC what would have to be done to go about getting it installed, and was informed that the first step is to write to the mailing list and check for any objections, which is the purpose of this email. So, I'll formally ask: Are there any objections to getting htop installed globall throughout the infrastructure?
It's in EPEL and is 67k, disk-space-wise according to yum.
Thoughts?
I work with the EPEL maintainer and we use it quite often... So I'm for it.
Sent from my iPhone
On Jun 19, 2010, at 12:07, Ricky Elrod codeblock@elrod.me wrote:
Hey everyone -- I was wondering about getting htop installed throughout the infrastructure, as I think this will help with on-the-fly checking of issues. For example, if load average is high, it is very easy using htop to sort by CPU, and see exactly what is causing the issue. If RAM usage is abnormal, it's easy to sort by that. It also has some other features that could prove useful.
A prime example of when this could have been useful was when hosted01's load averages went up to near 60 last night. It would have made finding the issue quite a bit faster for me.
I asked on IRC what would have to be done to go about getting it installed, and was informed that the first step is to write to the mailing list and check for any objections, which is the purpose of this email. So, I'll formally ask: Are there any objections to getting htop installed globall throughout the infrastructure?
It's in EPEL and is 67k, disk-space-wise according to yum.
Thoughts?
infrastructure mailing list infrastructure@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/infrastructure
If you want to install this, make sure you do it in puppet, don't just install it in yum on a machine.
htop is useful, it's just policy for things to be installed via puppet, as it is much easier to manage.
Darren L. VanBuren ===================== http://theoks.net/
On Sat, Jun 19, 2010 at 10:07, Ricky Elrod codeblock@elrod.me wrote:
Hey everyone -- I was wondering about getting htop installed throughout the infrastructure, as I think this will help with on-the-fly checking of issues. For example, if load average is high, it is very easy using htop to sort by CPU, and see exactly what is causing the issue. If RAM usage is abnormal, it's easy to sort by that. It also has some other features that could prove useful.
A prime example of when this could have been useful was when hosted01's load averages went up to near 60 last night. It would have made finding the issue quite a bit faster for me.
I asked on IRC what would have to be done to go about getting it installed, and was informed that the first step is to write to the mailing list and check for any objections, which is the purpose of this email. So, I'll formally ask: Are there any objections to getting htop installed globall throughout the infrastructure?
It's in EPEL and is 67k, disk-space-wise according to yum.
Thoughts?
infrastructure mailing list infrastructure@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/infrastructure
Yep, that's what I'm planning so long as there are no objections. Thanks :)
On Jun 19, 2010, at 8:19 PM, Darren VanBuren wrote:
If you want to install this, make sure you do it in puppet, don't just install it in yum on a machine.
htop is useful, it's just policy for things to be installed via puppet, as it is much easier to manage.
Darren L. VanBuren
On Sat, Jun 19, 2010 at 10:07, Ricky Elrod codeblock@elrod.me wrote:
Hey everyone -- I was wondering about getting htop installed throughout the infrastructure, as I think this will help with on-the-fly checking of issues. For example, if load average is high, it is very easy using htop to sort by CPU, and see exactly what is causing the issue. If RAM usage is abnormal, it's easy to sort by that. It also has some other features that could prove useful.
A prime example of when this could have been useful was when hosted01's load averages went up to near 60 last night. It would have made finding the issue quite a bit faster for me.
I asked on IRC what would have to be done to go about getting it installed, and was informed that the first step is to write to the mailing list and check for any objections, which is the purpose of this email. So, I'll formally ask: Are there any objections to getting htop installed globall throughout the infrastructure?
It's in EPEL and is 67k, disk-space-wise according to yum.
Thoughts?
infrastructure mailing list infrastructure@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/infrastructure
infrastructure mailing list infrastructure@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/infrastructure
On 19/06/10 18:07, Ricky Elrod wrote:
Hey everyone -- I was wondering about getting htop installed throughout the infrastructure, as I think this will help with on-the-fly checking of issues. For example, if load average is high, it is very easy using htop to sort by CPU, and see exactly what is causing the issue. If RAM usage is abnormal, it's easy to sort by that. It also has some other features that could prove useful.
A prime example of when this could have been useful was when hosted01's load averages went up to near 60 last night. It would have made finding the issue quite a bit faster for me.
I asked on IRC what would have to be done to go about getting it installed, and was informed that the first step is to write to the mailing list and check for any objections, which is the purpose of this email. So, I'll formally ask: Are there any objections to getting htop installed globall throughout the infrastructure?
It's in EPEL and is 67k, disk-space-wise according to yum.
Thoughts?
infrastructure mailing list infrastructure@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/infrastructure
I strongly suggest you wait until you get at least 3-4 +1s from the sysadmin-main folks. Keeping in mind, that any additional application is an additional potential attack vector.
Regards,
Tristan
Yep, which is why I'm asking here. It's not a daemon or long-running service or anything, but I do understand what you're saying. And yes, I'm planning on getting some -main's to +1 it, but I was told the first step was throwing an email out on the list.
Thanks
On Jun 19, 2010, at 8:38 PM, Tristan Santore wrote:
On 19/06/10 18:07, Ricky Elrod wrote:
Hey everyone -- I was wondering about getting htop installed throughout the infrastructure, as I think this will help with on-the-fly checking of issues. For example, if load average is high, it is very easy using htop to sort by CPU, and see exactly what is causing the issue. If RAM usage is abnormal, it's easy to sort by that. It also has some other features that could prove useful.
A prime example of when this could have been useful was when hosted01's load averages went up to near 60 last night. It would have made finding the issue quite a bit faster for me.
I asked on IRC what would have to be done to go about getting it installed, and was informed that the first step is to write to the mailing list and check for any objections, which is the purpose of this email. So, I'll formally ask: Are there any objections to getting htop installed globall throughout the infrastructure?
It's in EPEL and is 67k, disk-space-wise according to yum.
Thoughts?
infrastructure mailing list infrastructure@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/infrastructure
I strongly suggest you wait until you get at least 3-4 +1s from the sysadmin-main folks. Keeping in mind, that any additional application is an additional potential attack vector.
Regards,
Tristan
-- Tristan Santore BSc MBCS TS4523-RIPE Network and Infrastructure Operations InterNexusConnect Mobile +44-78-55069812 Tristan.Santore@internexusconnect.net
Former Thawte Notary (Please note: Thawte has closed its WoT programme down, and I am therefor no longer able to accredit trust)
For Fedora related issues, please email me at: TSantore@fedoraproject.org _______________________________________________ infrastructure mailing list infrastructure@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/infrastructure
On Sun, 20 Jun 2010, Ricky Elrod wrote:
Yep, which is why I'm asking here. It's not a daemon or long-running service or anything, but I do understand what you're saying. And yes, I'm planning on getting some -main's to +1 it, but I was told the first step was throwing an email out on the list.
If it doesn't pull any crazy deps in I'm not necessarly against it but it's not clear to me what it buys us over the tools that are on there now. You mentioned load (you can get with 'w'). CPU and RAM can both be ordered by the top command already installed.
-Mike
Thanks
On Jun 19, 2010, at 8:38 PM, Tristan Santore wrote:
On 19/06/10 18:07, Ricky Elrod wrote:
Hey everyone -- I was wondering about getting htop installed throughout the infrastructure, as I think this will help with on-the-fly checking of issues. For example, if load average is high, it is very easy using htop to sort by CPU, and see exactly what is causing the issue. If RAM usage is abnormal, it's easy to sort by that. It also has some other features that could prove useful.
A prime example of when this could have been useful was when hosted01's load averages went up to near 60 last night. It would have made finding the issue quite a bit faster for me.
I asked on IRC what would have to be done to go about getting it installed, and was informed that the first step is to write to the mailing list and check for any objections, which is the purpose of this email. So, I'll formally ask: Are there any objections to getting htop installed globall throughout the infrastructure?
It's in EPEL and is 67k, disk-space-wise according to yum.
Thoughts?
infrastructure mailing list infrastructure@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/infrastructure
I strongly suggest you wait until you get at least 3-4 +1s from the sysadmin-main folks. Keeping in mind, that any additional application is an additional potential attack vector.
Regards,
Tristan
-- Tristan Santore BSc MBCS TS4523-RIPE Network and Infrastructure Operations InterNexusConnect Mobile +44-78-55069812 Tristan.Santore@internexusconnect.net
Former Thawte Notary (Please note: Thawte has closed its WoT programme down, and I am therefor no longer able to accredit trust)
For Fedora related issues, please email me at: TSantore@fedoraproject.org _______________________________________________ infrastructure mailing list infrastructure@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/infrastructure
infrastructure mailing list infrastructure@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/infrastructure
On Sun, 20 Jun 2010, Mike McGrath wrote:
On Sun, 20 Jun 2010, Ricky Elrod wrote:
Yep, which is why I'm asking here. It's not a daemon or long-running service or anything, but I do understand what you're saying. And yes, I'm planning on getting some -main's to +1 it, but I was told the first step was throwing an email out on the list.
If it doesn't pull any crazy deps in I'm not necessarly against it but it's not clear to me what it buys us over the tools that are on there now. You mentioned load (you can get with 'w'). CPU and RAM can both be ordered by the top command already installed.
That seems like it would be useful to some, I'm not sure I would personally use it (I'm more of a sar man myself), but it's a non-suid tiny app that others might be able to use, fine by me.
-Mike
On Sun, Jun 20, 2010 at 9:37 PM, Mike McGrath mmcgrath@redhat.com wrote:
That seems like it would be useful to some, I'm not sure I would personally use it (I'm more of a sar man myself), but it's a non-suid tiny app that others might be able to use, fine by me.
sar is great for historical data (though I prefer collectl now that my $DAYJOB uses it - it is AWESOME!). htop is good for realtime data as a top replacement. However, it does have a number of unfortunate bugs in how it handles multiple CPU's :(
Hm, I haven't noticed any bugs with that in using it personally, or do you mean from a lower/code level?
I'm confused as to whether you're for or against it. :)
On Jun 21, 2010, at 12:56 AM, Jon Stanley wrote:
On Sun, Jun 20, 2010 at 9:37 PM, Mike McGrath mmcgrath@redhat.com wrote:
That seems like it would be useful to some, I'm not sure I would personally use it (I'm more of a sar man myself), but it's a non-suid tiny app that others might be able to use, fine by me.
sar is great for historical data (though I prefer collectl now that my $DAYJOB uses it - it is AWESOME!). htop is good for realtime data as a top replacement. However, it does have a number of unfortunate bugs in how it handles multiple CPU's :( _______________________________________________ infrastructure mailing list infrastructure@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/infrastructure
On Mon, Jun 21, 2010 at 1:58 AM, Ricky Elrod codeblock@elrod.me wrote:
Hm, I haven't noticed any bugs with that in using it personally, or do you mean from a lower/code level?
When you're trying to bind CPU's or look at the CPU utilization data at the top, it starts from CPU1 rather than the correct CPU0 - the authors simply need to learn to count :). I'm fine with it being installed though, htop is da bomb!
Alright, how does this look? +1 worthy? :)
--- a/manifests/services/global.pp +++ b/manifests/services/global.pp @@ -171,6 +171,7 @@ class global { "bind-utils", "man", "traceroute", + "htop", ]: ensure => installed, }
On Jun 20, 2010, at 9:37 PM, Mike McGrath wrote:
On Sun, 20 Jun 2010, Mike McGrath wrote:
On Sun, 20 Jun 2010, Ricky Elrod wrote:
Yep, which is why I'm asking here. It's not a daemon or long-running service or anything, but I do understand what you're saying. And yes, I'm planning on getting some -main's to +1 it, but I was told the first step was throwing an email out on the list.
If it doesn't pull any crazy deps in I'm not necessarly against it but it's not clear to me what it buys us over the tools that are on there now. You mentioned load (you can get with 'w'). CPU and RAM can both be ordered by the top command already installed.
That seems like it would be useful to some, I'm not sure I would personally use it (I'm more of a sar man myself), but it's a non-suid tiny app that others might be able to use, fine by me.
-Mike _______________________________________________ infrastructure mailing list infrastructure@lists.fedoraproject.org https://admin.fedoraproject.org/mailman/listinfo/infrastructure
On Mon, Jun 21, 2010 at 12:18 AM, Ricky Elrod codeblock@elrod.me wrote:
Alright, how does this look? +1 worthy? :)
--- a/manifests/services/global.pp +++ b/manifests/services/global.pp @@ -171,6 +171,7 @@ class global { "bind-utils", "man", "traceroute",
- "htop",
]: ensure => installed, }
After using it today I have no objections to having this installed. It seemed rather useful.
On Mon, 2010-06-21 at 21:31 -0600, Stephen John Smoogen wrote:
On Mon, Jun 21, 2010 at 12:18 AM, Ricky Elrod codeblock@elrod.me wrote:
Alright, how does this look? +1 worthy? :)
--- a/manifests/services/global.pp +++ b/manifests/services/global.pp @@ -171,6 +171,7 @@ class global { "bind-utils", "man", "traceroute",
"htop",After using it today I have no objections to having this installed. It seemed rather useful.
+1. I <3 htop.
Aside from nice CPU/memory usage views, I also use it to quickly strace processes (s) along with sending them signals (k).
luke
infrastructure@lists.fedoraproject.org
-
Darren VanBuren -
Jon Stanley -
Luke Macken -
Mike McGrath -
Ricky Elrod -
Rob Marti -
Stephen John Smoogen -
Tristan Santore