Tags: mysql

06/30/13

  10:42:00 am, by The Dreamer   , 902 words  
Categories: FreeBSD

Catching up on port update backlog....

This weekend, I decided it was time that I checked on port updates in my /compat/i386 FreeBSD 'system'. Which primarily exists to provide me some ports that don't build on 64-bit, namely emulators/wine-devel and net/nxserver. Don't recall the last time I used nx since I got it working, probably should check to see whether it is still working or not (probably okay on my home system, but might be broke on work one....and might see about setting it up on other work computer too).

Hmmm, hadn't updated ports since May 5th. Start with working through /usr/ports/UPDATING, run into a problem that on 20130609: AFFECTS: users of audio/flac and any port that depends on it, in that there it thinks perl depends on it (kind of an annoyance I have with dependencies....there can be miles of separation between one port and another port, but everything get's marked as depending on that very bottom port, when it in fact didn't or doesn't... Was annoying in trying to figure out why a port was marked BROKEN / DEPRECATED and not get any attention except that people should stop using it...when 100's of ports on my system depend on it. When it turns out that its one or two ports, had an option set that caused it to depend on it. While the other ports generally don't care what options are enabled in that port, just that the command exists for it...or other reason. Though there are some ports that do care about what options were used, which I had ranted about earlier...and I ran into Thunderbird also having that dependency, resulting in this kluge patch:

Code

--- Makefile.orig  2013-06-26 06:01:34.000000000 -0500
+++ Makefile  2013-06-27 20:07:04.142845537 -0500
@@ -98,6 +98,8 @@
.endif
post-patch:
+  @${REINPLACE_CMD} -e '/with SQLITE_SECURE_DELETE/s/_ERROR/_WARN/' \
+    ${WRKSRC}/mozilla/configure.in
  @${REINPLACE_CMD} -e 's|%%LOCALBASE%%|${LOCALBASE}|g' \
    ${WRKSRC}/mail/app/nsMailApp.cpp
.if ${PORT_OPTIONS:MENIGMAIL}

But, I let the portmaster -r flac run aways, with the suspicion that it would break later because perl modules that depend on perl (and not flac) wouldn't get picked up as needed to be re-installed or upgraded, due to 20130612: AFFECTS: users of lang/perl* and any port that depends on it. But, would break the re-install or upgrade of a port somewhere and abort. Which is what I found when I checked on it this morning.

So, I did a portmaster -R -r perl, and noticed that it seemed to include most of the ports that the previous portmaster hadn't done. In fact it included all of them. I also peeked in /usr/local/lib/perl5/5.14.2 and /usr/local/lib/perl5/site_perl/5.14.2 to see what perl modules had gotten missed....mainly the p5-XML-* ones that caused the previous portmaster to abort.

Though I probably should've looked to see if the second portmaster was going to address those, instead of doing them while it was asking to proceed. Because that caused it to abort when re-installing those perl modules (that I had done while it was waiting), but restarting it got things done.

That leaves the latest entry 20130627: AFFECTS: users of ports-mgmt/portmaster, which is just informational and not currently applicable.

Before running in to the flac entry, there had been "20130527: AFFECTS: users of lang/ruby18" which was pretty straight forward, since it only exists as a dependency to ports-mgmt/portupgrade, which I seldom use now...but I have other scripts that use binaries that come as part of it (namely portsclean), which I could probably replace with the portmaster way or something else. But, its not really a priority, plus who knows if I won't decide to go back to using portupgrade...which has options in its pkgtools.conf that I haven't found equivalents for with portmaster, though isn't currently an issue right now. Except perhaps that I'm holding back on updating to the latest emulators/virtualbox-ose, since I've gotten warnings from various sources to stay away from it.

The other big one is what's the portmaster equivalent to portupgrade's ALT_PKGDEP?

Will eventually run into a port that has a line one of these:

Code

RUN_DEPENDS+=   mysql-server>=0:${PORTSDIR}/databases/mysql${MYSQL_VER}-server
RUN_DEPENDS+=   mysql-server>=0:${PORTSDIR}/databases/mysql${MYSQL_VER}-server  
RUN_DEPENDS+=   mysqld_safe:${PORTSDIR}/databases/mysql55-server
RUN_DEPENDS+=   ${LOCALBASE}/bin/mysqld_safe:${PORTSDIR}/databases/mysql55-server
RUN_DEPENDS=    ${LOCALBASE}/libexec/mysqld:${PORTSDIR}/databases/mysql${MYSQL_VER}-server
RUN_DEPENDS+=   ${LOCALBASE}/libexec/mysqld:${PORTSDIR}/databases/mysql${MYSQL_VER}-server
RUN_DEPENDS+=   ${LOCALBASE}/libexec/mysqld:${PORTSDIR}/databases/mysql${MYSQL_VER}-server
RUN_DEPENDS+=   mysql-server>=0:${PORTSDIR}/databases/mysql${WANT_MYSQL_VER}-server
RUN_DEPENDS+=   mysql-server>=0:${PORTSDIR}/databases/mysql${MYSQL_VER}-server
RUN_DEPENDS+=   mysql-server>=0:${PORTSDIR}/databases/mysql${WANT_MYSQL_VER}-server
LIB_DEPENDS+=   mysqlclient:${PORTSDIR}/databases/mysql55-client
RUN_DEPENDS+=   mysqld_safe:${PORTSDIR}/databases/mysql${MYSQL_VER}-server

Where I'm using databases/percona55-server & databases/percona55-client now....

Somehow expected there would be more than 12 ports wanting either client or server... probably missing the occurrences in multiline RUN_DEPENDS or some other way to specify the depend. Since pkg_info says there are 103 ports that depend on the client, and two ports that depend on the server (neither being www/owncloud or mail/roundcube, which are the ports that I'm running on 'zen' using the mysql server. On cbox, there are 73 ports that depend on it, some are obviously true...like net-mgmt/cacti and net-mgmt/cacti-spine, but nothing is depending on the server...though it is obviously being used by cacti. I left dbox with the default databases/mysql55-client...there are 71 ports depending on it.

Meanwhile I have postgres server running on zen, which was a depend of something else that I had since removed....but I haven't stopped or removed postgres yet....

In fact after dealing with ruby, flac and perl....the only ports left to update are:

dialog4ports-0.1.3 < needs updating (index has 0.1.5_1)
freeglut-2.8.0 < needs updating (index has 2.8.1)
portmaster-3.16 < needs updating (index has 3.17)

Might be time to see what else I can get working under wine.

Wonder when I had last updated ports in the /compat/i386 on my system at work? And, do I want to tackle that from home, on a Sunday instead of other important tasks/projects....

But first...lets see what port got updated since yesterday....

ports-mgmt/portupgrade

Odd, I seemed to have missed updating to lang/ruby19 from lang/ruby18 on cbox and dbox....

06/14/13

  08:36:00 am, by The Dreamer   , 959 words  
Categories: Software, FreeBSD, CFEngine

Perl update + cfengine3 and subversion == great big mess on FreeBSD

So being up ridiculously early this morning, because I got up at the right time during the night...but forgot to take my second dose, and too late to take it when I woke again.

So, I thought I would tackle updating to the latest perl on zen. Choose not to just do a blind 'portmaster -r perl', since that would include anything that depends on being able to run perl or depends on something that depends on perl...possibly many levels down. Somebody should come up with a simple way to only rebuild the ports that directly depend on another port....

What I decided to do what after updating perl, was 'portmaster p5-*' then do a pkg_libchk to see what ports are missing libperl.so and redo those.

I did this first by logging into my machine at work (mew) and it went quickly and worked pretty cleanly. Though there are only half the number p5- ports, and less ports missing libperl.so as well.

So, back on zen...I updated perl and then started the 'portmaster p5-*', when it stopped. A port is dependent on /usr/local/bin/perl5.14.2, which is not found because I just updated to 5.14.4!

How do you have ports that are dependent on a specific version, especially since its possible to have a different version of perl. When I first installed zen, the default perl was 5.12.x...but I had elected to upgrade to 5.14.x when that became the new default. OTOH, when I setup cbox/dbox, I opted to go with 5.16.x...and at work, mew had started before zen, so it was also 5.12.x initially...but I choose to jump that up to 5.16.x (actually no .x at first, but it was very quickly followed by a .1...which inflicted a lot of pain again....though in this latest update they have switched to just major.minor for module path, which means no more pain until its time to upgrade to a newer minor release....

The first casualty was devel/p5-B-Keywords ...and I find its depended on by textproc/p5-Perl-Critic and textproc/p5-Test-Perl-Critic....with the last port being the leaf. So that seems to be that it was probably just a build depend for something I had installed long ago (since it doesn't sound like anything I could call in any of my own perl scripts.) So, I figure I'll just delete those packages and carry on.

Nope, the next port also fails with the same strange demand.

&#58;&#35;&#35;

After hunting around a bit, to try and figure out what is making this port think it depends on the previous version of perl...it dawns on me that the perl port updates /etc/make.conf as to what version of perl is installed.

And, I have

/etc/make.conf

being managed by cfengine3. &#58;&#111;&#111;&#112;&#115;&#58;

So, I go update the file, and try to svn commit the change....which blows up because the perl modules needed for the commit hook haven't been updated yet.

Well, guess I'll stop cfengine3 from reverting my /etc/make.conf (by disabling the promise from the root side) Though IIRC now, the commit hook is mainly to prevent root from doing commits into my subversion repository, instead of the put subversion on an NFS filesystem and have rootsquash prevent root from being able to write into the repository, that we do at work.

Phew....I can commit updated /etc/make.conf and let cfengine promise it. Perhaps in a future project, I'll see if there's some way to have cfengine set that dynamically or something.

Though should I have cfengine promising all of /etc/make.conf ? There's a block in /etc/make.conf that is the same across all my FreeBSD systems, since its the ports that I come across that don't like 'make -j#'. And, it was intended to have cfengine promise that part, though it there have been other additions that I want the same on all my FreeBSD systems, like the override on modules in net-snmp, that I mention when I ran into some cacti. Though there's value in having cfengine having the whole file....or rather its that the files are in subversion.

Next up...nagios has alerted me that spamassassin has stopped, yeah, I guess that would happen. Which means there's going to be a big chunk of spam in all my mailboxes (126...I need to find a good way to aggregate them back so that I can read them all ... from roundcube) module rebuilds are done, and spamassassin seems to be running again (probably because cfengine3 has a promise to keep for that) Though might work better if I do an 'sa-compile' to make sure that part is right...though seemed to me it was only major.minor....

Full story »

04/21/13

  10:52:00 pm, by The Dreamer   , 2431 words  
Categories: Software, Computer, Ubuntu, FreeBSD, CFEngine

Home server migration ran into some cacti

The home server migration that I wrote about on April 7th, hit a delay .... I started working on migrating cacti and nagios.

I probably should've started with nagios, since I don't think that would've taken as long as cacti has.

I had already been monitoring the new servers using my old cacti installation. I had pretty much decided that moving the old installation to the new servers wasn't going to straightforward.... partly because of versions, and no easy intermediary. But, I wasn't too worried about the historical data in my old cacti....

I figured that once I got things up and running, I'd just export the templates and import them into my new system and I'd be done.

But, then I hit a hitch....the squid templates I had weren't working on the new system....all I could find were old results about issues with doing SNMP to ports other than 161, and possibly due to newer versions of net-snmp....though that later turned out to be a wild goose.

Anyways...the work around was to use the proxy option in net-snmp. Though I recall having tried net-snmp before discovering bsnmpd on FreeBSD, but I gave it a shot.

Before I got to testing the proxy...I soon saw that it wasn't giving the same information as bsnmpd...specifically, for the HOST-RESOURCES-MIB and parts of UCB-SNMP-MIB. So, I decided that I could proxy net-snmp to bsnmpd and get those. But, that didn't work.....after some reading the answer was I needed to either map bsnmpd in somewhere else or exclude those areas from net-snmp.

Well, during the build of net-snmp, it did make reference to being able to set some variables in make.conf -- such as NET_SNMP_WITH_MIB_MODULE_LIST and NET_SNMP_WITHOUT_MIB_MODULE_LIST. And, by default NET_SNMP_WITH_MIB_MODULE_LIST contained "host disman/event-mib smux mibII/mta_sendmail mitII/tcpTable ucd-snmp/diskio sctp-mib if-mib"

So, I tried setting NET_SNMP_WITH_MIB_MODULE_LIST without host and ucb-snmp/diskio and tried to exclude the rest of ucb-snmp in NET_SNMP_WITHOUT_MIB_MODULE_LIST. Which got me a strange error about host being in both lists.

I delved into the Makefile, and found while the other settable NET_SNMP parameters were done as '?=' in the Makefile, the NET_SNMP_WITH_MODULE_LIST was done as '+='...with conditionals that '+=' the last two modules.

OSVERSION >= 700028 adds 'sctp-mib' and the port option MFD_REWRITES adds 'if-mib'....I had started looking at what the fix might be, but decided that all I needed to do was remove all these lines...since I'm going to have my own definition in my /etc/make.conf file.

Trying to exclude all of ucd-snmp wouldn't make things work....but I did an snmpwalk comparing bsnmpd and net-snmp, and decided that the two areas that were lacking were ucd-snmp/diskio and ucd-snmp/disk_hw. So, I recreated the 'original' NET_SNMP_WITH_MODULE_LIST in /etc/make.conf, without 'host' and 'ucd-snmp/diskio' and put 'ucd-snmp/disk_hw' in NET_SNMP_WITHOUT_MODULE_LIST. The build grumbled, but finished.

I that worked.....all my ucd/snmp host graphs were working on m new cacti server in the same detail that I was getting before (IE: the CPU Utilization gave traces for each of the 8 vCPUs...instead of just one.... I could see all the ZFS filesystems, not just the the single zroot.

So, I went back to looking at getting squid graphs to work....that didn't work.

Pages: 1· 2· 3

07/10/11

  03:39:00 pm, by The Dreamer   , 770 words  
Categories: General, Stuff

I have my own YOURLS

Link: http://lkc.me/yourls

A while back I saw that my initials were available in the .me domain...and I had always wanted to get my intials.

Though, its kind of a pain switching domains and stuff. Was bad enough going from the old URL for this blog to the current one, and I haven't moved any of my email addresses from the old domain (have created some new ones on this domain though).

But, I snapped up the domain name anyways....because I had envisioned that some day I would migrate off of the old 'chen-online.net' domain. I had gotten it before I knew that domains with dashes are associated with spammers, and often get blocked on the sole basis that there's a dash in the domain name. Though, making the decision to change my blog's URL from 'lawrence.chen-online.net' to 'lawrencechen.net' was kind of a no-brainer to me.

Didn't find that out until I pushed hard on why there were a couple of blogs where I left lots of comments on, but they never seemed to clear moderation. One site admin said explained this to me, the other's was using a hosted solution and never saw my posts come in for moderation...and didn't know why. Seemed odd though since they were hosting for an official site under the domain of where I work....could just be everybody likes to type as little as possible so they just haven't run into the email address filter....

Anyways....I didn't have any plan on what I would do web-wise with the domain. And, while I do have Adsense, I'm not sure what kind of ads they would throw up for my initials and whether it would all be appropriate or not (if only to me). Plus I may yet decide to use it for something...someday.

But, then while I was sleeping last night...the idea came to me that 'lkc.me' is pretty short URL, wonder if I could set up my own, private URL shortening service.

This morning I did a quick google search on what's out there....found a matrix that somebody had compiled....quickly decided that it needed to be standalone, private, custom URL capable.

That resulted in only one choice. YOURLS &#58;&#99;&#111;&#111;&#108;&#58;

And, that's what I did....

http://lkc.me/blog

Full story »

Now instead of subjecting some poor random forum to a long rambling thought, I will try to consolidate those things into this blog where they can be more easily ignored profess to be collected thoughts from my mind.

Latest Poopli Updaters -- http://lkc.me/poop

bloglovin

There are 20 years 5 months 28 days 5 hours 36 minutes and 20 seconds until the end of time.
And, it has been 4 years 7 months 8 hours 26 minutes and 36 seconds since The Doctor saved us all from the end of the World!

Search

July 2017
Mon Tue Wed Thu Fri Sat Sun
 << <   > >>
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30
31            
Google

Linkblog

  XML Feeds

Who's Online?

  • Guest Users: 0
This seal is issued to lawrencechen.net by StopTheHacker Inc.
powered by b2evolution

hosted by
Green Web Hosting! This site hosted by DreamHost.

monitored by
Monitored by eXternalTest
SiteUptime Web Site Monitoring Service
website uptime