Robert passed CIW Database Design Specialist certifcation

Lagniappe Internet’s founder Robert Porter has passed the Certified Internet Web Professional, aka CIW, Database Design Specialist exam today.
CIW descibes this exam…

CIW v5 Database Design Specialists have mastered the knowledge and theory of database design that applies to the most popular database platforms. These professionals help solve the problem of poorly designed databases. Aimed at database programmers and administrators alike, this vendor-neutral certification focuses on universal database design principles and SQL. The CIW v5 Database Design Specialist exam validates foundational knowledge of databases in general, such as Oracle, IBM, DB2, MySQL and others.

CIW v5 Database Design Specialist certification is valuable for individuals working in fields such as IT, database development, application development and other areas that depend on Web-enabled systems for productivity. To become a CIW Database Design Specialist, the candidate must pass one required CIW exam AND complete the CIW Certification Agreement by logging in to the CIW Candidate Information Center.

Together with the previously passed CIW Associate requirements, gives Robert the “CIW Professional” certification as well. This is in a long list of certifications gained over the years including A+, Network+, Security+, MCSE+Inet, and CCNA.

Short downtime while migrating vm’s

We’ll be experiencing some short downtimes while migrating some virtual machines – the largest of which is our own lagniappeinternet.com website. It was down for 9 minutes while migrating. Most others should see no more than a few seconds of downtime. The vm’s that are being migrated will be housed on faster nodes… so this is a good move that has been planned for a while.

GT20-1 issues

Since the Christmas eve event where all servers, pdu, switches, etc briefly lost power in Dallas, the GT20-1 server has not been behaving well. Virtbiz in Dallas worked hard to get the server back up and going, but it’s simply not reliable. Any change to CMOS settings and it doesn’t want to boot… So we don’t want to put it back into production. They’re going to ship the server back to us, and we’re going to migrate off of that chassis to one of the spares we have here. We’ll likely do an signifigant upgrade to the hardware at the same time.

For what it’s worth… Our decision of virtualizing all customer facing servers and backing them up and copying the backups to other servers worked. We’re looking at refining that backup procedure to decrease the downtime in the future. We apologize for the downtime and will be contacting affected customers shortly.

Dallas nodes … power

All nodes, power distribution system, etc. lost power for a moment. They are restarting now, and should be available shortly.

Update: gt24-2, x335, x306 are back up, and all virtual servers on them are running.
Sent support page to the d.c. for gt20-1

Update 2: Virtbiz’s support and company websites are down. The gt20-1 node is taking a long time to come back up.

Update 3: I’ve paged a virtbiz support tech, but have not gotten any responses. Their support site is down. The This affects the DirectAdmin virtual server plus a couple others. We apologize for the downtime, and will post Virtbiz’s response when we get one.

Update 4: Since we can’t seem to get in touch with Virtbiz (we will be getting a LONG-TERM resolution to that problem!) in any way shape or form. We’re restoring last night’s backups from gt20-1 to alternate nodes. It will take apx 30 minutes to have the virtual servers done. We apologize, and will be contacting the affected customers in the coming days. One quick note: DirectAdmin customer weren’t affected by gt20-1 extended outage. That virtual server had been migrated to gt24-2 previously for load balancing.

GT24-2 node problem

At apx. 9:55pm central time, GT24-2 had an issue and kernel panic’ed. The node is coming back up now, though it is taking some time for drive array checks, etc. All vps’es on it should be run again in the next 10 minutes but with reduced speed while the array consistency check completes. It should be back to normal within the next couple hours.

Reboot of GT24-2 node

GT24-2 is being rebooted for a (hopefully) quick fsck of a filesystem.
This affects several virtual servers including the DirectAdmin reseller.
I dont expect it to be long provided the fsck repairs are not too extensive.

9:30 : The server has been renooted and one fsck done on the node’s filesystems… It wasn’t pretty. We’ll need to reboot again. I’m going to migrate the vps containers to another node. so that gt24-2 can be diagnosed and issues addressed properly.

10:30 : The server is up and fscks are now clean… I’ve stopped the migration of the vps containers. We’ll monitor the server and if we need we’ll live migrate the containers to another node.

GT24-1 offline issue

At about 5:15pm central time, we had an issue where gt24-1 went offline. We can be pretty sure of when it went offline as I was logged into the server as part of the migration of vps accounts to SolusVM. I had gone to get something to drink. When I came back I had an instant message about email on an account on a vps on that server not working… I went to check and my SSH session was dead. After some initial testing and determining the server was completely unresponsive, I put in a ticket with Tulix support. They had to come out to reboot the server. The server was at a kernel panic. I suspect it had to do with the tweaking we had to do to convert the server over to a SolusVM node. I’ll continue to monitor the node. A second Atlanta node server is currently being prepped for deployment later this year… We may have to alter our timelines. I have a email in to Tulix management asking about the new charge for a reboot, as well as how we can best go forward… for example, is a remote PDU is order.

As of 7:15 the server was coming back up. By 7:25pm all vps containers we up and running.

SolusVM … vps renumbering

We are working to implement the SolusVM VPS management system… it has come a long way in the past few months and is taking a quick lead over the other offerings. This is however requiring us to rearrange the vps container numbering. This is a number behind the scenes, but SolusVM needs it a certain way for management purposes. Each vps will be done individually and should only take a few minutes each. During the conversion the VPS will be taken offline. This is true for reseller accounts as well, as reseller servers are virtualized.

This is an entirely internal process. Nothing on your logins, ip addresses, etc. will change. VPS customers will be given a login to the new management console after their vps has been converted.

We thank you for your understanding in this process. We’re expecting to convert all vpses to SolusVM over the next 2 to 3 weeks.

GT24-1 drive error – Updated

GT24-1 has alerted us to a hard drive error… The server uses a RAID-10 array, and the other 3 drives are functional. So the server is still online. We are in the process of running a manual secondary rsync of the nightly vzdump files to another server as a precaution. A replacement drive has been ordered for overnight delivery to the datacenter. We’ll update this post as information is available.

9/23 2pm: The drive has been replaced and we are actively resyncing the arrays at this time. We did have to reboot the server a couple of times during the procedure as linux had decided to ‘renumber’ the drive and we needed back at /dev/sda.

DA1 – restarted

The DirectAdmin virtual server da1 was restarted due to a runaway process count.
At 8:42am, we received a nagios warning about the number of processes running on the gt24-2 node (almost 3 times its normal).
Upon closer inspection, 1/3rd of all processes belonged to the da1 VM.
At at 8:47am, we elected to restart the VM — some held connections were evident during the restart.
By 8:49am, the da1 VM was available and the number of processes back to normal.