FrOSCon 2010 is over .. waiting now for 2011 :-)
Posted by fabian.arrotin in CentOS, Fun on August 26th, 2010
It was the first time that I was at FrOSCon and I admit I enjoyed it. Not only because I can always see in real life some other CentOS contributors (thanks again Andreas, Sarah, Didi and Christoph), but also because I can see some other people really happy with CentOS. Last year (even if i was not there), CentOS was used to ‘power’ some TFT screens at the entrance. So we did the same this year and just because Dag asked it, we took a picture this year too 
Automatic network switcher on Nokia E71
Posted by fabian.arrotin in Fun, Nokia on August 19th, 2010
I’ve always loved Nokia products and always have been satisfied by my last two E-series (E51 and E71). But there are those little things that can bring you a better life when using it. For example, you can decide which data access point (Wireless or 3G) you want to use when launching an application. That’s fine and of course I prefer using my WLAN at home than the 3G connection. But what if you schedule something to happen automagically on your phone, like calendar and contacts sync (I sync those with the integrated Mail for Exchange application, even if, obviously, i don’t have an Exchange server, but rather a Zarafa OpenSource server with z-push installed). You have to define a ’sync plan’ and choose which connection it will use in the background. Wait, I’d like it to select my Wireless AP when at home (or with some pre-defined wlans from friends, etc ..) if available and then switch to 3G if no wlans available. That’s where SmartConnect helps you a lot : you can define a ‘fake’ access point which is in fact a group that contains your connections (wlans/3G) with priorities and it will use the first available one. Use that ‘fake’ access point on your mobile, whatever the application (even tested with Putty for Symbian). Great and useful.
You can download it from the official website, or directly install it from your mobile through the OVI app installer.
CentOS promo team @ FrOSCon 2010
Posted by fabian.arrotin in CentOS on August 4th, 2010
Hi all … just to let you know that some members of the CentOS promo team (including myself) will be at the ‘Free and OpenSource Software Conference‘ in Germany in (more than) two weeks …
Feel free to come at the CentOS booth to discuss with some of us !
DRBD backported (or not) to 2.6.32 in EL6 ?
Posted by fabian.arrotin in CentOS, Cluster, linux-ha on July 20th, 2010
As some of you already know it, DRBD is now (since kernel 2.6.33) part of the mainline/upstream kernel. Some were expecting RHEL6 to come with that kernel (used for Fedora 13). The latest RHEL6beta2 still comes with 2.6.32, which doesn’t include DRBD support. Of course we still don’t know what the ‘frozen’ RHEL6 kernel will be but on the other hand, we know that Red Hat quite often ‘backports’ modules from newer kernel into the RHEL kernel. What about DRBD ? At the time of writing this blog post, it seems still undecided, but you can follow the DRBD RFE on Upstream Bugzilla to get a clue, or even comment on it if you have a bugzilla account to make hear your voice. On the other hand, you can still be sure that even if DRBD isn’t part of EL6, CentOS will still ship it in the Extras repository, like for EL4/EL5 …
Following build status through Twitter …
Posted by fabian.arrotin in CentOS, Fun on May 31st, 2010
The other day, Russ pointed me/us to a cli client for Twitter : TTYter. Even if i’m not myself a Twitter (ab)user, I thought it would be funny to create a tweet that can be followed by those Twitter abusers and wanting to follow the build status of the RPMforge PPC rpm packages that I build. I’ve quickly modified my automated build scripts to post a build status after each build and you can follow it here : http://twitter.com/rpmppcbuilder . It would be nice also to have such stuff for other build machines as well … While i’m talking about those PPC packages, if you still have a PPC machine doing nothing and that you want to throw away, please redirect it to me instead (that can even be an IBM js2x blade that I would be able to have it hosted somewhere) 
samba preexec and postexec to the rescue …
Posted by fabian.arrotin in CentOS on April 9th, 2010
I had yesterday to migrate a server running CentOS 4.x to a new xen domU running CentOS 5.4 . While the migration was the easy part (including the data), the ‘problem’ I had was the backup part. In fact the backup application used is a proprietary one (Computer Associates Arcserve, to keep it ’secret’ .. ). Well, that proprietary backup agent (v11.5, old but still used ‘here’) doesn’t work on CentOS 5.4 (glibc issue) and i didn’t want to find a workaround that problem (with LD_LIBRARY_PATH or some other tweak). That’s where Samba came to the rescue : the Arcserve backup server can backup Windows nodes (and so Samba nodes too) through UNC path . OK, but for the previous system i had prebackup and postbackup jobs/scripts on the CentOS node (doing specific things, including creating a lvm snapshot and removing it after the backup job). How could i do that now through the samba-backup-method approach ?. Well a standard cron job on the CentOS machine was of course always doable but I (we ?) prefered some kind of ‘triggering action’. That’s where i used for the first time the preexec and postexec options for a samba share. Basically it still uses my prebackup and postbackup scripts (doing the snapshot and mount it, as well as the reverse , obviously) and expose that mounted snapshot during backup time for the backup server. .. Working nicely and maybe it can give you some ideas ….
Extending (live) a SR (storage repository) on XenServer 5.5
Posted by fabian.arrotin in CentOS on March 4th, 2010
For my new job I have to learn how to deal with Citrix XenServer (yeah, because of a mixed workload of CentOS domU’s and Windows TSE servers, for which XenServer has been optimized). I liked the fact that I’m directly feeling “like home” , as Citrix XenServer dom0 is based on CentOS (still 5.3 at this time though). One of the things i had to do was to extend a Storage Repository served from an IBM DS3200 through dual HBAs, and using mpp/rdac (the default on XenServer 5.5 when it sees a rdac disk storage backend). Great, I’ve never had problem doing this on plain RHEL or CentOS machines, so after having extended the LUN on the IBM DS3200, I was back on the XenServer side. I always like to read the official documention before doing something (and it’s even faster when you know what you’re searching for) and I found this on the Citrix XenServer documentation : “How to resize a Storage repository after changing the size of an LVM-base storage” . Hmmm, WTF ? Their recipe is : “live migrate the guests, restart the host and proceed for each host”! . No, it has to work without a reboot, we’re not Windows admins, right ? Here is what i did : (that was tested on a test machine !)
We have first to list the current status/size :
[root@xen1 ~]# xe sr-list
uuid ( RO) : c945d1bb-2432-36ac-2766-ebd2bc7f2e81
name-label ( RW): Hardware HBA virtual disk storage
name-description ( RW): Hardware HBA SR [IBM - /dev/sdb]
host ( RO): xen1
type ( RO): lvmohba
content-type ( RO):
[root@xen1 ~]# xe sr-param-list uuid=c945d1bb-2432-36ac-2766-ebd2bc7f2e81|grep physical-size
physical-size ( RO): 85886763008
[root@xen1 ~]# pvscan|grep c945d1bb-2432-36ac-2766-ebd2bc7f2e81
PV /dev/sdb VG VG_XenStorage-c945d1bb-2432-36ac-2766-ebd2bc7f2e81 lvm2 [79.99 GB / 16.12 GB free]
Now we’ll extend with the IBM DS StorageManager script editor : “set logicalDrive ["XenPool1"] addcapacity=139 GB;”
Back on the xen host we have to rescan for the new size (using a MPP device presented as /dev/sdb on the xen host) and confirm with dmesg|tail
[root@xen1 device]# echo 1 >/sys/block/sdb/device/rescan ; dmesg|tail
sdb: detected capacity change from 85899345920 to 235149459456
[root@xen1 device]# pvresize /dev/sdb
Physical volume “/dev/sdb” changed
1 physical volume(s) resized / 0 physical volume(s) not resized
[root@sicxen1 device]# pvscan
PV /dev/sdb VG VG_XenStorage-c945d1bb-2432-36ac-2766-ebd2bc7f2e81 lvm2 [218.99 GB / 155.12 GB free]
PV /dev/sda3 VG VG_XenStorage-9c1e7a2a-2fc0-83eb-3e32-7cea2c9e9d93 lvm2 [60.59 GB / 60.59 GB free]
Total: 2 [279.58 GB] / in use: 2 [279.58 GB] / in no VG: 0 [0 ]
Rescan now that SR :
[root@xen1 device]# xe sr-scan uuid=c945d1bb-2432-36ac-2766-ebd2bc7f2e81
[root@xen1 device]# xe sr-param-list uuid=c945d1bb-2432-36ac-2766-ebd2bc7f2e81|grep physical-size
physical-size ( RO): 235136876544
Done ! and i confirm that the CentOS domU’s were still running after that …
PS : while talking about Citrix XenServer, I have to add that I used only ssh/xe to manage it, as their XenCenter gui app is a Windows only GUI (relying on .Net). But I found several days ago an interesting GPL project: OpenXenCenter, something to keep an eye on as it’s still alpha but quickly involving …
WARNING: mismatch_cnt is not 0 on /dev/md0 through LogWatch
Posted by fabian.arrotin in CentOS on February 21st, 2010
Recently I received (through logwatch) several weekly reports about a mismatch in synchronized block on my md0 (/boot) device :
WARNING: mismatch_cnt is not 0 on /dev/md0
`cat /proc/mdstat` was normal though.
A `echo repair >/sys/block/md0/md/sync_action` followed by a `echo check >/sys/block/md0/md/sync_action` seems to have corrected it. Now `cat /sys/block/md0/md/mismatch_cnt` returns 0 …
Fosdem 2010 is over …
Posted by fabian.arrotin in CentOS, Fun on February 11th, 2010
Yes i know, it’s even over for more than 4 days, but i was too busy with other stuff to write a small report of the event , which i’ll do later if `locate free_time` returns something useful.
In the meantime, a (funny) picture from the CentOS booth at Fosdem 2010 
CentOS @ Fosdem 2010
Posted by fabian.arrotin in CentOS on January 23rd, 2010
CentOS promotion
http://www.centos.org