You are not logged in.
The Azure environment had to be restarted this morning.
The sites are available for testing by NewMars members of record in September of 2023.
SpaceNut ... just FYI ... kbd512 is considering testing the backup he is taking of the forum database.
We might be able to restore it to the Azure test environment. The only caveat is that I have subscribed to the teeniest tinyest possible Virtual Machine, and it may not have the resources to perform a full restore of the full database.
It if ** works ** it will be a nice surprise.
(th)
Offline
The Azure environment was running but not responsive this morning. The log showed no changes in the past week except the most recent restart.
I decided to stop the VM at 12:06 UTC.
The deallocation normally takes on the order of 10 minutes.
I found it stopped at UTC 12:19 ... Deallocation took less than 13 minutes
Restarting now at 12:20 UTC...
Running at 12:21 UTC ... All sites are back up at 12:23 UTC
The entire refresh operation took from 12:06 UTC to 12:23 UTC.
The sites should now remain up for a couple of days.
They remain available for NewMars members to check out the updated software, or to evaluate the phpBB3 alternative.
Status of upgrade: The upgrade was postponed due to the Houston power outage.
Work will resume after the dust settles there in Houston.
(th)
Offline
The Azure environment was up this morning...
It remains available for anyone to test the new software kbd512 is in the process of installing.
We also have a version of phpBB3 available for members to test, for comparison with our upgraded FluxBB.
And we have a small demo program that shows what a "Like" feature might look like.
(th)
Offline
Azure sleepy but woke up ....
Azure was non-responsive when I touched it this morning at 11:00 UTC, but it was back in normal operation when I tested at 14:00 UTC.
I was expecting to have to restart the system because the OS backup failed four days ago, but on this occasion it came back up by itself, with a bit of time.
***
I'm starting to think about attempting to create a web based version of GW Johnson's rocket mission planning software.
This would be quite an ambitious undertaking, but ChatGPT4o tells me that I can run Python as a service called by PHP, which is the language we have installed in our Azure environment to support the updated FluxBB package and the test phpBB3 service.
Everything we have set up remains available to NewMars members for evaluation, while we wait for the conditions to be favorable for work to resume on upgrade of our software from Version 1.5.11 to 1.5.12.
You can see a simulation of the new "Like" feature at: http://40.75.112.55/forums/UpdateEdited.php?id=80882
When you click refresh, the count of the Edited field increments, just as it would do if to were wired to support "Like".
The idea i am thinking about is (possibly) implementing Dr. Johnson's rocket mission planning software (currently a spreadsheet) as a web based service. I don't know if it is possible for us to achieve this, but ChatGPT4o seems to think there is a chance.
(th)
Offline
Update at 12:25 .... it happened again .... the system started OK, but it locked up when I started too many instances.
I'll just let it go for now, to see if it resolves itself. This is probably an issue related to the bottom tier capabilities of this subscription.
It is disappointing to see the failure after months of operation with multiple instances active.
Update at 12:00 UTC .... it appears that the system fails if too many instances are started. This is a new problem.
The applications showed up initially, when I opened them on one machine. However, when I started opening the applications on a second machine, all the application instances froze up.
Update at 11:46 UTC ... I ** thought ** the system was up, but the three test sites are not responding.
The log only shows a successful restart. I'll run the process again, starting at 11:45 UTC.
Deallocation completed at 11:56 ... Restarting (again)
Azure was non-responsive this morning.
The log showed an OS update failed yesterday.
Install OS update patches on virtual machine Failed a day ago
Mon Aug 05 2024 01:55:18 GMT-0400 (Eastern Daylight Time)
Azure subscription 1
Stop order was issued at 11:00 UTC
De-allocation is in progress ...
Stop state is reported as of: 11:12 UTC
Start state reported... and the site is back up at 11:13:49
(th)
Offline
2024/08/16 After a pause due to external circumstances I decided to reopen work with the Azure environment to try to install a web site to support GW Johnson's course and book projects.
The online session started well enough. However, after installing the first of a number of packages for Python, the connection froze, and now the Azure applications are not responding.
Update a few minutes later .... the system came back and I restarted the installation. It got further before freezing again.
The freezing of the terminal session happens simultaneously with freezing of the applications we have installed on the site.
I'm used to this but am less than happy to see it happening again.
Update later ... the session came back briefly. The previously installed packages are reported as present, and the installer continued working on dependencies, until it started Collecting jupyter-server>=1.24.0, at which point it froze again.
On the next try we got a bit further. This time the failure occurred as we were working on jupyter-server 1.24.0, and downloading jupyter_server-2.14.2-py3. The file is 383.6 kb and the download reached 174.1 kb.
Update later: The system is back attempting to load the modules for the project... it is slower than previously.
Apparently the jupyter_server is installed, but the process froze after reporting "Using cached" for the 383 kb file.
I could probably avoid all this by paying for more memory for the virtual machine.
The monthly fee is under $15, and I can afford that. The next level up is a great deal more.
Update later ... the retry got a bit further ... Collecting decorator
Collecting exceptiongroup
Collecting pygments
Collecting platformdirs
Collecting anyio
Collecting websocket-client
Collecting jupyter-server-terminals
Collecting nbformat
Collecting overrides
Collecting prometheus
Collecting send2Trash
Collecting argon2
Collecting (missed)
Collecting beautifulsoup4
Collecting async_timeout
Collecting aiohappyeyeballs
And the freeze occurred here
Surprise! the procedure picked up ...
Collecting multidict
Collecting frozenlist
Collecting aiosignal (long pause here)
Collecting yarl (301 kb)
There is a progress bar for this download, standing at 0.0/301.6 kb
eta is given as -:--:--
Nice!
Update later ... beautifulsoup4 just showed up on the retry ....
frozenlist just showed up ...
aiosignal should be next ...
This is worse than waiting for a pot to boil.... the fire keeps going out.
Retrying again ... we are back to frozenlist
Now we are back to yarl
For a change of pace, an orange message appeared reporting network problems
However, yarl now shows as completed
On to asynch-timeout
And back to aiohappyeyeballs
Collecting jupyterlab-widgets
Pausing here ... on to widgetsnbextension
On to notebook-shim
On to jupyterlab 11.6 mb (oh oh! )
Got it! surprise
On to sniffio
On to argon2
On to webencodings
On to parso
On to jsonschema
Collecting referencing
... on to babel - 9.6 mb - 3.6 reached ... up to 4.9 ... the red completed line is crawing to the right
eta is given as 1:34 ... I doubt that !!!
Amazing! It completed babel ...
What's next ???
Collecting requests ... got 64.9 kb
I'm wondering if the delays are at the Python server?
This message appeared: host='pypi.org' along with a timeout report
I have no way of knowing how many more modules are needed ...
*** starting over
If the delays ** are ** in the python server, then the Azure system should find all the modules already installed and start work on any remaining ones.
However, the delay is obviously NOT just at the python server, because the connection to the Azure instance is hanging up at open-session time.
Starting over at 17:46 UTC ...
Last login was reported as 17:10 so 36 minutes were consumed in the session that completed babel
The pace is a bit faster this time... the virtual machine is apparently getting a bit more machine time at Azure
On the other hand, all the modules have been cached, so there are no calls to the python server
I'm watching for beautifuleyeballs ...
Well, beautifulsoup showed up
... long pause at asynch_timeout
frozenlist should be next to show up ...
There it is!
There is happyeyeballs ...
Now we're hanging on jupyterlab-widgets...
It'll be a while before we get back to babel...
The pause lasted longer than my patience ... starting over at 18:17 UTC
Once again the connection to Azure is hanging...
The restart is underway at 18:25 UTC ... bleach just went by again
beautifulsoup4 just showed up again
happyeyeballs is back ...
babel is back ... on to json5 and requests
On to wcwidth ... soupsieve
On to executing
On to h11
Now we are downloading again ...
Collecting rpds-py
Collecting isodduration
Collecting webcolors
On to charset-normal ... download 142.1 kb
On to cffi
On to pycparser 117.6 kb
On to arrow - 66 kb
On to types-python 66.4 kb
Collecting types-python-dateutil
Well, we seem to be getting a bit further...
Now we have a new message: Installing collected packages: the list is long and it ends with vpython.
Well, ... now we have a warning: wsdump is installed in /home/azure/.local/bin which is not on PATH. Consider adding this directory to PATH
I don't see wsdump in the list of modules being installed, but perhaps it is a dependency of one of the modules that are visible in the list.
pip install numpy scipy matplotlib astropy vpython plotly flask
(th)
Offline
After a while ChatGPT4o and I decided to break the installation into tiny steps. I installed each module one at a time, and the installation succeeded.
The test program appears to be running, but it is a web server, and I don't know how to access it.
My time has run out for use of ChatGPT4o so I'll have to resume this later.
Update a bit later ... Python uses Port 5000 for it's web site service, so i had to log into Azure and Add a new inbound port rule.
However, then I ran out of gas in the ChatGPT4o account and will have to wait until tomorrow to get more quality time.
it's late in any case, and I'm happy with the results to date. The new web site may begin working tomorrow. We'll see.
The web site is set up to run a simulation of an orbit. This all started when GW Johnson was working on the rocket for SpinLaunch.
I'd like to be to offer NewMars members a chance to see GW's computations in animation.
(th)
Offline
Microsoft is requiring all Azure users to implement two factor authentication.
The process is unbelievably complex.
However, I think I can understand the problem... Microsoft is serving a global audience, and they are under attack by every evil force on the planet.
I tried to sign up for the Extra ID service, and was forced to create a new account, separate from the one that I use to manage the existing subscription.
It appears that I need to set up the primary account as a "tenant" in the new "secure" account.
In the mean time, the Asure test environment has been non-responsive all day. I just initiated a restart procedure.
19:07 UTC Stop command issued
This time it appears the deallocation is going to run long ... 19:17 still in progress (10 minutes)
Stopped at 19:18 ... ok! starting: 19:18:37 >> Status updating ... Running at 19:19:19 ...
And! The test sites are running again.
Restarting the VM again at 21:16 UTC .... the new access rule is supposed to allow access via port 5000, but port 5000 is not responding.
I moved the Apache2 default web site from port 80 to 5000, and it was not accessible from outside. However, it ** was ** accessible from inside, on 10.0.0.4.
That means the Apache2 restart worked correctly, and that the server is listening on port 5000.
Subsequently I restored port 80, and everything is back to normal, except that the Python server is still not working.
(th)
Offline
The Azure environment is responding this morning, which is good.
My objectives for today include:
Complete configuration of Python to be served by Apache2 like all the other web services.
Demonstrate the "Hello World" Python web site running from the Azure environment.
***
Add two more images for the "space pilot" book series...
***
Begin looking into how to enhance the "Hello World" program to show an orbit.
(th)
Offline
Today's report is about a breakthrough in the Azure environment that appears to be having a positive effect on performance.
I have reported from time to time on my favorable experiences using the upgraded version of ChatGPT, currently called ChatGPT4o.
In this post, I will report on the astonishing (to me for sure) computation by ChatGPT4o that led to the discovery that there was no swap file installed in the Virtual Machine I am renting from Azure. It would never have occurred to me to think about that, because a swap file is ** always ** installed in every installation of Ubuntu on a local machine, so I mistakenly thought/expected it would be installed at Azure.
ChatGPT4o listened to my tale of woe about the problems I have been seeing in working with my least-cost bottom-rung installation on Azure, and it asked if I had a swap file? When I revealed I didn't know, it provided the command to find out.
When that command revealed there ** was ** no swap file, it provided the commands to create one and to activate it.
Shortly afterward, performance improved dramatically. ChatGPT4o ** then ** provided the commands to find out how well the swap file was doing, and ** that ** command revealed that of my 1 Gig of paid-for memory, only 160 Mb was left to do useful work after MySql and all the applications software was loaded. The new swap file took up the slack, and of the 1 Gig allocated, less than half was in use.
I am happy to report that Azure was up this morning.
***
The problem ChatGPT4o and I are having with trying to run the Python Flask application persists. The application runs inside the Azure environment, which is proven by using a tool called "curl" to poke at the web site. However, the site is still not visible to the outside world, so more work is needed.
The whole purpose of the exercise is to try to provide animation online to support Dr. Johnson's course material, which he is attempting to adapt for an elementary school audience from the late high school/early college level it is now.
The Flask application itself is running in a simple demo mode. It is able to compute and display an orbit in a 3D grid, which displays in HTML on any browser. This is clearly a promising direction to be heading, but the outside access problem remains to be solved.
ChatGPT4o has suggested an alternative way to deliver the web site. Our first attempt failed, but ChatGPT4o evaluated the results and recommended a possible solution. That solution is to create more than one virtual web site under Apache2, which I had no idea was possible.
The theory of the recommendation appears to be that Apache2 can serve requests on two different ports (80 and 5000) if they are separated into "virtual" web sites.
(th)
Offline
This report is a follow up to the earlier one on the addition of a swap file to the Azure environment.
In this afternoon's session, ChatGPT4o and I tackled the backlog of updates that have NOT been applied to the OS.
This backlog has caused failure of the VM once a week when the environment tries to apply updates. ChatGPT4o gave me the option of performing updates module by module, but we decided to try going for broke with all of them, and to my delight they all applied. The automatic process then restarted a number of services including MySql. My hope is that the hangups that have been occurring will stop happening. The proof will be seen in a week, when the VM tries to apply updates according to it's regular schedule.
Now that ** this ** problem has been solved, I intend to return to the problem of how to persuade Azure to serve the Python Flask web site to outside visitors. We know that the application works inside the VM, because a tool called "curl" proved that the website was serving data.
However, it is break time in Azure City.
(th)
Offline
This is a follow up to Post #186 .... the new swap file seems to be working. The Azure environment has been up since the updates were applied on the 22nd. The command to show how memory is being used is:
free -h
This command was entered this evening, before I start work on the various todo items that remain. Here is the display:
free -h
total used free shared buff/cache available
Mem: 833Mi 701Mi 55Mi 15Mi 76Mi 16Mi
Swap: 0B 0B 0B
The VM I am renting has 1 Gb of RAM, and only 55 Mb remains free with the existing applications all running. I am surprised to find the swap file empty.
While working with Flask (the Python website) the swap file loaded up, but apparently housecleaning must happen automatically, because the traces of that activity appear to be gone.
The purpose of the current todo list is to (at least try to) implement a web site to support Dr. Johnson's existing course with animation, and beyond that, to support a new and ambitious book project, to try to bring his instruction to a 9-12 years age group. I am hoping the Azure environment can be persuaded to support such a web site, and ChatGPT4o tells me it should be possible.
However, our first attempts to run Flask failed. The default method is to run Flask as a Python app using port 5000, However, despite opening a firewall port for 5000, I was never able to access Flask from outside, even though the application was proved to be running inside the firewall.
The second approach bypasses the entire port 5000 issue by enlisting Apache2 to serve the Flask application via port 80. However, ** that ** attempt failed in a dismal way, because the existing port 80 applications stopped working. ChatGPT4o thought about that for a nanosecond or two, and decided that our problem might be solved by setting up two "virtual" connections in Apache2. The first exists now and it serves the PHP applications on port 80, as well as the default web site. A new "virtual" server would (presumably) pick up requests for the Flask application and route them properly, without bothering the existing Port 80 applications. That would be nice, if we can get it to work.
I could spend time enhancing the Flask application, but see no point in spending time on it until the access problem is solved.
(th)
Offline
And this post comes after the work session this evening ...
ChatGPT4o and I seem to be making progress, but the Flask website is not yet running and visible from the outside.
The big improvement this evening is that the Apache2 configuration file now contains code to support flask, while keeping the existing web sites running.
When I started the flask application however, the entire environment froze up, but I was able to stop the flask application and everything is back to normal.
Meanwhile, my available time with ChatGPT4o has expired, so I'll have to pursue this tomorrow.
(th)
Offline
Azure is running this morning (Sunday 2024/08/25)...
That is definitely encouraging, because the code to support the Flask application is still present, and the PHP applications are still running.
There is still work to be done to enable the Flask application to run (if it can).
An attempt to run the Flask application last night resulted in the Azure environment entering a non-responsive state, but the system recovered when the Flask application was stopped.
One possible explanation may be that the Flask application is consuming all available CPU cycles, which would be the case if a loop is built into the software, and ** that ** might be the case, since this is an animation application.
For anyone who might not have thought about animation before, the ability to simulate continuous movement I similar to that we experience in a movie theater with old fashioned celluloid frames passing through a projector. The frames are fixed but our brains provide continuity from one frame to the next, so the movie appears to be continuous.
Something similar happens in a computer animation ... the computer works on each "frame" during a time interval, and it releases the output to the screen for the viewer while it works on the next frame.
My (tentative) work plan for today is to attempt to measure the CPU usage of the Flask application on a development PC where it runs just fine. The development PC has 8 gigs of RAM compared to the paltry 1 gig on the Azure virtual machine.
***
There is another possible strategy that might work for animation ....
JavaScript is a language that runs in browsers on practically every browser type.
JavaScript runs on the local machine, and it might be suitable for rendering animation as a way of taking that burden off the remote server.
If this is a solution (and that remains to be seen) it needs to run on cell phones as well as tablets and laptops.
(th)
Offline
The OS update failed again last night.
I had hoped that by manually applying updates, and by adding a swap file, this persistent failure would go away, but obviously it has not.
Restart is in progress. Deallocation usually takes on the order of 10 minutes, and restart is about 30 seconds, but it takes considerably longer for the VM to refresh itself, including restarting MySQL which always seems to take a while.
Running at 13:39:40
phpBB3 is up at 13:40
FluxBB 1.5.12 is up
Port 80 Apache is balky >> The system locked up again
Stopping again ... this is getting old...
This time I logged into the terminal window and stopped Apache2 quickly, before the system had time to fail.
In diagnosis, it turned out that there was an error in configuration of the new code to support Flask.
ChatGPT4o provided suggestions to correct the error, and now the system appears to be serving the default applications correctly.
Update at 19:06 UtC ... after it's struggles this morning, caused by an error in configuration of software to support Flask, Azure seems to be holding on.
There is at least one more hurdle to solve before the Flask application starts working.
(th)
Offline
Azure seems healthy this morning.
(th)
Offline
If anyone is curious to see the current state of play in testing of a flask application to run on Azure, here is a command you can enter:
The external viewer sees a complaint message, but the inside viewer sees a cascade of errors that all come down to the application not finding a file that is in plain sight. ChatGPT4o and I battled this mystery until I exhausted my allowance again. We are just going in circles at this point, but we ** did ** make some progress by removing the entire wsgi package and re-installing it.
It's time to take a break.
(th)
Offline
To my amazement, Flask is now working from Azure. ChatGPT4o ran out my allotted time again but it put me within the last 10 yards to the goal, and I was able to install the needed modules where they needed to be.
The program is now running from http://40.75.112.55/flask/
The display shows a simple orbit path. I've been holding off on working on the program until the ability to deliver it to the public is working.
The larger goal is to try to support GW's kid's book project with animation.
(th)
Offline
The trial of Azure is coming to an end...
Microsoft has been generous in providing hardware and software at no charge, and they have only charged for CPU and disk and possibly Internet traffic, which I have been able to afford.
Buy a subscription to keep using your product
InboxMicrosoft <microsoft-noreply@microsoft.com>
5:30AM (8 hours ago)
to meMicrosoft
Your trial is ending—buy a subscription to keep using your product
You have a trial that will expire on September 17, 2024. To continue using your products, buy a paid subscription in the Microsoft 365 admin center.Purchase a subscription >
If you don’t purchase a new subscription by September 17, 2024, your trial data will be available to all people in your organization for about 30 days. At that time, the subscription status changes to Disabled, and only admins will have access to trial data for about another 90 days. After that, your trial is permanently deleted.Your trial that will expire:
Product Total licenses Start date
Microsoft Entra Id Governance Trial 25 August 17, 2024
Additional information
How toextend a trial.
How tocancel a trial.
Account information
Organization name: Mars Society NewMars forumsExpiration date: September 17, 2024
Primary Tenant Id:
(th)
Offline
For SpaceNut re Pothole Scan on Azure version of FluxBB....
For as long as I've been using the Azure environment, I've had a problem with the environment going to sleep due to lack of activity, and then taking a long time to wake up. Now that the Pothole Scanner is back in robust operation, I decided to try running it on the Azure version of FluxBB. That is the new software that kbd512 is working on installing at the Mars Society web server. I am hoping the long running application will keep the Azure system awake.
(th)
Offline
For SpaceNut ....
my little experiment running the Pothole Scan against the Azure system revealed a problem, but at this point I don't have enough information to know what happened. The symptom is an unhappy message from MySQL.
Azure itself is not responding. Even just Apache2 without MySQL is hung up.
Well! This is a free demo that is set up with the absolute minimum capabilities it is possible to have.
While it's disappointing to have the job doing 100,000 lookups fail, on the ** other ** hand, that ** is ** asking a lot of a free demo.
The Pothole Scanner stopped with a Big Red Screen! It was working on the 400th lookup at the time of failures.
Here is the ugly SQL error:
OK ... I can't show the MySQL message due to AISE.
I'll break up the text:
W a r n i n g: m y s q l i::__c o n s t r u c t(): E rror while reading greeting packet. PID=501424 in / v a r / w w w / h t m l/f o r u m s/F l u x B B/c o n f i g.p h p on line 23
Fatal error: Uncaught m y s q li_s q l_exception: M y S Q L server has gone away in / v a r/ w w w/ h t ml/ f o r u m s/F l u x BB/co n fig.p hp:23 Stack trace: <snip> on line 23
One possible explanation is that the MySql instance ran out of memory. I added a 1 gig swap file recently, but that may not have been enough for this workload.
From the application log we have:
Completed Sequence for ID: 214399
Starting Sequence for ID: 214400
Smart Wait Address Retry Count = 2Report of status at compare failure:
Row counter: 8 Row contents: WEPLogged in as tahanson43206
Expected string: 0
Received string: Actual
(th)
Offline
For SpaceNut ... the Azure environment is still running but today is the end of the free trial of one year. During this past year I've been paying for disk, CPU and Internet services, but Microsoft has provided the software at no charge. They've notified me that they intend to pull the plug if I don't pay for the software, but so far they've made it impossible for ** me at least ** to find the button to click to upgrade.
I submitted a request for help on the community bulletin board, and I did get a response but it was for the procedure I followed a year ago, not the one that is needed now.
In addition to ** that ** dilemma, there is the oncoming locomotive of conversion to Two Factor authentication. Most sites make this extremely easy, but Microsoft is serving a global customer base of primarily major corporations and government agencies. These entities need a complex management system to manage potentially hundreds of employees, and the interface they need is far in excess of what I need.
Fortunately, kbd512 has made excellent progress in moving toward implementing the updated FluxBB software, thanks to timely assistance from Mr. Burk here and there. We appear to have a duplicate database ready to go in our own test directory, and after our Google Meeting on Sunday, kbd512 reported successful delivery of the software components to the directory. There are now some detail adjustments needed before we can begin testing the new version at the Mars Society server. There is a file to be adjusted, and a table in the database, and there may be something else I might have forgotten.
All this progress is timely because the Azure environment may be going away in coming weeks.
As a reminder to all NewMars members.... We've had the new software running and available for everyone to test for many months.
The address is: http://40.75.112.55/forums/FluxBB/
The logins are the same as in our production environment.
***
In other related news .... I restarted the pothole finder on the Azure test site, but adjusted the delay between cycles from 5 seconds to 30. The idea is to keep the Azure environment from going to sleep, without causing a major ding to my (limited) budget. The last time I tried this the run failed, and according to the log the cause was most likely another failed attempt to apply OS updates.
The run was set for 600 posts at 30 second intervals, or 300 minutes or 50 hours. It will be interesting to see how far we get this time.
The advantage of this procedure (assuming it works) is that our members will find the Azure system responsive if they decide to test the new software.
(th)
Offline
For SpaceNut .... the keep-alive run on the Azure system failed after 209 cycles...
Fatal error: Uncaught mysqli_sql_exception: MySQL server has gone away in /var/www/html/forums/FluxBB/include/dblayer/mysqli.php:75 Stack trace: #0 /var/www/html/forums/FluxBB/include/dblayer/mysqli.php(75): mysqli_query() #1 /var/www/html/forums/FluxBB/include/functions.php(266): MysqlDBLayer->query() #2 /var/www/html/forums/FluxBB/include/functions.php(131): set_default_user() #3 /var/www/html/forums/FluxBB/include/common.php(146): check_cookie() #4 /var/www/html/forums/FluxBB/edit.php(10): require('...') #5 {main} thrown in /var/www/html/forums/F
To my surprise, Azure itself had not locked up, so I'll just restat and see what happens.
224609 on the way to 225000
(th)
Offline
For SpaceNut re Azure test environment....
I've been beating pretty hard on the new FluxBB system running in Azure, and some problems have shown up. I assume they are all due to the weakness of the Azure environment, which is a free bottom tier learning system.
Here is an error that showed up recently:
Fatal error: Uncaught mysqli_sql_exception: Table './fluxbb_newmars/fluxbb_posts' is marked as crashed and last (automatic?) repair failed in /var/www/html/forums/listtopic.php:19 Stack trace: #0 /var/www/html/forums/listtopic.php(19): mysqli->query() #1 {main} thrown in /var/www/html/forums/listtopic.php on line 19
The implication is that the posts table in the test system has been damaged.
I haven't been posting anything ** to ** the table, but I have been doing lookups.
Update later.... here's another version of the problem report:
Fatal error: Uncaught mysqli_sql_exception: Table './fluxbb_newmars/fluxbb_posts' is marked as crashed and last (automatic?) repair failed in /var/www/html/forums/FluxBB/include/dblayer/mysqli.php:75 Stack trace: #0 /var/www/html/forums/FluxBB/include/dblayer/mysqli.php(75): mysqli_query() #1 /var/www/html/forums/FluxBB/viewtopic.php(52): MysqlDBLayer->query() #2 {main} thrown in /var/www/html/forums/FluxBB/include/dblayer/mysqli.php on line 75.
Update 2024/09/19 ... My guess is that the frequent crashes of the Azure environment have caused whatever problem exists with the Posts table. It turned out that when I added a swap file to the VM, I did not make it permanent, so the crashes continued until yesterday. The swap file will now persist over a restart, so hopefully the OS updates will complete, and thus not take the system down.
However, we have a learning opportunity. The Posts table is apparently marked as bad/crashed, so php procedures trying to use it are blocked my MySQL.
I hope to learn about the flag, and what to do to reset it.
(th)
Offline
As our year long experiment with Azure draws to a close, I am able to report that for the first time, the VM was able to apply OS updates on it's own without crashing. In addition, the ping every 30 seconds is keeping the VM awake and thus responsive to requests from NewMars members who might wish to test either the updated FluxBB package, or the phpBB3 package we set up for evaluation. In addition, if a NewMars member feels on the adventurous side, we have a simple Python Flask service running. it doesn't appear to be doing much, but ** boy!** it sure was a ** lot ** of work to get it to show up at all.
The only ongoing issue is with the posts table in the FluxBB test site. The frequent crashes prior to today appear to have caused damage to the posts table, and I am hoping to tackle that today.
As a reminder for folks who might like to see these services before they go away, here are the URL's:
http://40.75.112.55/forums/FluxBB/
(th)
Offline