Mark's blog - Comments

Scott: Apache mod_proxy balancing with PHP sticky sessions

nospam@example.com (Scott) — Thu, 02 Oct 2008 13:59:19 +0100

My cookie value is getting URL encoded. In my log I see:

proxy: BALANCER: Found value balancer%2Ewww1 for stickysession BALANCEID

So the route is never getting
discovered.

I'm setting the cookie with an ASP page. Anyone know whats going on?

anil: Apache mod_proxy balancing with PHP sticky sessions

nospam@example.com (anil) — Thu, 25 Sep 2008 16:10:19 +0100

in your above examle you are using http like below.

BalancerMember http://www1.example.com
BalancerMember http://www2.example.com

in my case i have two apache as BalanceMember and they are listing at secure port 443.
so my links are

BalancerMember https://www1.example.com
BalancerMember https://www2.example.com

this is not working????? why??
more over in my real servers i am re writing the urls too.

and while i request my loadbalancer apache it just redirect my request to real server and in my browser i see the address of my real server. which i dont like.

can anybdy help me.

Dennis Clarke: Blastwave is dead

nospam@example.com (Dennis Clarke) — Thu, 11 Sep 2008 17:28:34 +0100

lost ?

it is sitting there .. never moved.

# grep "^mark" /etc/passwd
mark:x:16434:101:Mark Round:/export/home/mark:/bin/ksh

What you were caught in was me trying to stop the total lift of everything out of Blastwave by three people that simply announced they were suddenly up and running somewhere else with everythign they could carry.

Mark Round: Blastwave is dead

nospam@example.com (Mark Round) — Wed, 27 Aug 2008 14:39:00 +0100

Sorry, but I've lost all my work, build notes, and scripts.

However, if you grab the PHP source it looks like adding dBase support is pretty simple : http://uk2.php.net/manual/en/dbase.installation.php

Pavel Lisy: Blastwave is dead

nospam@example.com (Pavel Lisy) — Wed, 27 Aug 2008 14:26:59 +0100

This is really sad. I've used Blastwave long time.

I have question to you. We use your packages (mod_php) from Blastwave. But we need dbase compiled in it. Can I get source of this solaris package from you? I want try compile dbase support in it. Can you help me even if Blastwave build stack is down?

Mark Round: Blastwave is dead

nospam@example.com (Mark Round) — Mon, 11 Aug 2008 12:21:17 +0100

Login and the build stack is still down, mailing lists are still down, the csw subdomain (which was where the packaging effort was going to move to) is down...

It's definitely not back up.

Joshua Gregory: Blastwave is dead

nospam@example.com (Joshua Gregory) — Fri, 08 Aug 2008 18:46:38 +0100

Looks like it's back up now.

Hardayal Patil: Apache mod_proxy balancing with PHP sticky sessions

nospam@example.com (Hardayal Patil) — Sun, 03 Aug 2008 19:12:37 +0100

I am trying to use load_balancer for HTTP load balancing without using servlets or Tomcat. Has anyone tried this?

I have 5 prescription servers in the back-end that support HTTP tunneled requests. They have an internal sessionid called CustSessID that is coming from cookie. I want stickyness so that the requests from same browser go to the same Prescription Server. I also have URL rewriting that connects to Generic server or BrandName Server.

Uwe: Apache mod_proxy balancing with PHP sticky sessions

nospam@example.com (Uwe) — Fri, 01 Aug 2008 16:00:00 +0100

hi,

I checked this and it work so far.

But what are you doing, when a server fails? I mean completely.
So far I've got a so called healthcheck script which checks the BalancerMember and reconfigures the apache.
Now, the cookie is still set and the proxy sends the request to the failed server. Do you have any idea how to get into that request and send to another BalancerMember?

Best Regards,
Uwe

Mads: ZFS Replication

nospam@example.com (Mads) — Mon, 16 Jun 2008 20:30:07 +0100

Thanks a lot for this post. I didn't realize zfs send actually supported incremental sends before reading your article This is just what I need for remote backup of my zfs based home server.

suresh: ZFS Replication

nospam@example.com (suresh) — Fri, 13 Jun 2008 05:02:58 +0100

HI in the above example iam facing a problem while performing the task

i have take the first snapshot but during second snapshot i got error like

"cannot receive:destination has been modified since most recent snapshot"

can u please rectify the problem..

aidan: ZFS as a volume manager

nospam@example.com (aidan) — Mon, 10 Mar 2008 03:36:59 +0000

Thanks for this.

I've just set up a raidz zpool and then created a volume as described for UFS.

This way I get the required UFS filesystem over 4 disks without the RAID5 write-hole problem i would have faced using Solaris Volume Manager.

Plus it was a snap to configure thanks to the zfs commands.

Naveenkanth: Apache mod_proxy balancing with PHP sticky sessions

nospam@example.com (Naveenkanth) — Thu, 31 Jan 2008 19:17:50 +0000

Mark,

Thanks for the post.

Paulo,

JRun's JSessionID will not help for session stickiness here. So I would propose writing an extra filter in your web application(in your case CF) and add a custom cookie in both the instances such as cookie.node1, cookie.node2 Then add this custom cookie name in the stickysession=${Custom_cookie}

This should help to maintain session stickiness in the balancer, though it will break in case of fail over.

-JR

kon: LigHTTPd and Apache - Symfony benchmarks

nospam@example.com (kon) — Thu, 11 Oct 2007 13:17:58 +0100

Hello Mark

Have you already re-run the tests with both daemons running fcgi?

Phil Harman: ZFS as a volume manager

nospam@example.com (Phil Harman) — Tue, 11 Sep 2007 11:04:40 +0100

It's pretty cool. However, you do need to take more care when you snapshot a mounted UFS. The filesystem metadata should be consistent (as long as you haven't turned off logging -- in which case you really ought to use fsck(1M) before mounting).

The main issue is that some file data may still be waiting to be flushed from the page cache to the parent zvol (so some file contents may be randomly stale or zero in the snapshot).

In your example, if you do the following ...

# date >/mnt/testvol/testfile; zfs snapshot test/testvol@1

... the snapshot will probably contain the file, but will probably NOT contain the expected data (i.e. the date). This is because the metadata describing the file gets written immediately, but the file data has to wait for fsflush.

By default, fsflush aims to scan the entire page cache every 30 seconds. This means that unless you sync the file contents explicitly, you could be waiting for up to 30 seconds or more for new file data to make it to non-volatile storage.

On systems with a large amount of RAM, or lots of dirty pages, or both, fsflush may not manage to clean the entire page cache within the default 30 seconds. Indeed, on systems with lots of RAM fsflush is often configured (via the /etc/system parameters autoup and tune_t_fsflushr) to take much longer (typically 5 minutes) so that it doesn't use too much CPU.

If your file data hasn't made it to storage at the point the snapshot was taken, you just won't see it in the snapshot (or ever again in the original filesystem if the system loses power or crashes before the flush occurs).

However, if you do something like this ...

# date >/mnt/testvol/testfile; lockfs -wf /mnt/testvol; zfs snapshot test/testvol@1; lockfs -u /mnt/testvol

... all should be well. The downside is that the snapshot operation now takes much longer, such that it may disrupt other users of the source filesystem.