May 2020 - luv-main - example.com

Squirrelmail
by Russell Coker 09 Jun '20

09 Jun '20

https://en.wikipedia.org/wiki/SquirrelMail The Wikipedia page indicates that Squirrelmail is no longer maintained. I've had a problem reported with it after upgrading to PHP 7.3. Is there a good replacement for Squirrelmail? Something simple and lightweight that just gets the job done. Not Horde. -- My Main Blog http://etbe.coker.com.au/ My Documents Blog http://doc.coker.com.au/

4 6

Re: Open Bug Bounty
by Russell Coker 28 May '20

28 May '20

On Thursday, 28 May 2020 9:13:27 AM AEST James McGlashan via luv-main wrote: > (Redacted retransmit. Unsure if Russell received unredacted version after no > response and the issue remaining unpatched. Added a note about TeamHash' > low prices and implied low time.) Yes I got it thanks. I'm dealing with things as I have time and energy for them. While the pandemic has given me more free time it has given me less energy for significant things. -- My Main Blog http://etbe.coker.com.au/ My Documents Blog http://doc.coker.com.au/

1 0

Re: Open Bug Bounty
by James McGlashan 28 May '20

28 May '20

(On-list resend; accidentally replied directly to Sam.) On Thu, May 28, 2020 at 10:15:32AM +1000, Sam Varghese via luv-main wrote: > Where do you get this figure from? It sounds wildly optimistic, even > after > passage of the Code on Wages Act, 2019 in August last year. Ah, I misread a resource. Discard that figure. Thanks for pointing this out Sam. I can't find any hourly rate. Hmm. Wikipedia states the minimum daily wage can vary between cities; as low as 160 rupees to as high as 750. This approximates to 1.33, 6.66, and 2.66 days respectively for 750. Or with the lower rate of 160; 6.24, 31.24, and 12.5 days. However, we're talking about a team of security researchers; not a single person and not a low paying job. I don't know enough about Indian work to say; so take all this with a grain of salt. https://en.wikipedia.org/wiki/List_of_minimum_wages_by_country

1 0

Upgrading LUV server
by Russell Coker 27 May '20

27 May '20

In preparation for a Drupal upgrade I've just upgraded to MariaDB on the LUV server. Everything seems to work ok, if you notice anything broken then let me know. Also I removed the DNS entry for members.luv.asn.au as that wasn't configured in the web server. -- My Main Blog http://etbe.coker.com.au/ My Documents Blog http://doc.coker.com.au/

2 1

Open Bug Bounty
by Russell Coker 26 May '20

26 May '20

https://www.openbugbounty.org/reports/1170432/ Is this some kind of scam? The web page in question is a static page with an embedded Google search field. Unless there's a problem with the Google search (which would probably be more of a problem for Google than for me) then I can't imagine what the issue might be. Details aren't provided, presumably they want me to pay for that. -- My Main Blog http://etbe.coker.com.au/ My Documents Blog http://doc.coker.com.au/

2 2

iotop and kvm
by Russell Coker 21 May '20

21 May '20

When I run iotop on a kvm server I see the qemu processes for KVM being listed as between 1MB/s and 8MB/s for writes, the aggregate of those processes is about 10MB/s. This isn't an impossible number as the image files for KVM are stored on a RAID-1 array of SSDs. Reads tend to be about zero because there is a lot of cache and the system has been running for a while. When I run iostat on the block devices for the RAID-1 I see the average write speed reported at something that's always below 2MB/s and which is often below 1MB/s for a 10 second reporting period. Could iotop be considering writes to /dev/kvm or /dev/net/tun as disk IO? What else could be the explanation for this? How could I work this out? -- My Main Blog http://etbe.coker.com.au/ My Documents Blog http://doc.coker.com.au/

1 0

Strange load average
by Russell Coker 15 May '20

15 May '20

At the bottom of this message is an extract from the monitoring system for the LUV server. Load average spikes to over 20, but at the time of monitoring there was only 1 D state process and nothing was using much CPU time or much RAM. At the same time other VMs didn't report high load so it wasn't an issue of the disk capacity of the hardware being saturated (which would be difficult for a RAID array of SSDs on a mostly quiet server). 09:55:01 CPU %user %nice %system %iowait %steal %idle 12:35:02 all 4.28 0.00 1.15 17.87 1.40 75.29 12:45:01 all 4.82 0.00 1.54 17.73 1.57 74.35 12:55:01 all 5.00 0.00 1.20 18.15 0.91 74.75 13:05:01 all 5.60 0.00 1.33 17.79 1.00 74.28 13:15:01 all 14.11 0.00 2.70 16.95 1.56 64.69 13:25:01 all 4.21 0.00 1.52 20.83 0.89 72.56 Average: all 5.15 0.00 1.31 19.99 1.09 72.45 Above is part of the sar output. Note that at around the time of the high load average there was a higher than usual amount of user CPU time. The iowait while not correlated with this issue was higher than I expected, I ran "iotop -o -d5 -b -P" which indicated that writes from mysqld was the main disk access. I ran "fatrace -f W" which indicated that mysqld was writing to deleted files in /tmp. | 68991 | luv_drupal | localhost | luv_drupal | Query | 0 | Creating sort index | SELECT v.vid, v.*, n.type FROM vocabulary v LEFT JOIN vocabulary_node_types n ON v.vid = n.vid WHERE | The only time I caught an access with the "show processlist;" SQL command was the above, might "Creating sort index" mean writing to deleted files in /tmp? ALERT itmustbe/loadavg: 21.27 6.84 3.01 >= 7 5 4 (Wed May 13 13:06:39) Summary output : 21.27 6.84 3.01 >= 7 5 4 Detailed text (if any) follows: ------------------------------- Here are D state processes: USER PID VSZ RSS TTY COMMAND root 157 0B 0B [jbd2/vda-8] Here are processes with the top CPU percentages: USER PID CPU TTY COMMAND mon 1630 7.0 /usr/bin/perl /usr/lib/mon/mon-local.d/ loadavg.monitor 7 5 4 mon 1629 6.0 /usr/bin/perl /usr/lib/mon/mon.d/msql- mysql.monitor --mode m www-data 1445 5.1 /usr/sbin/apache2 -k start Here are processes with the top RAM use: USER PID VIRT RES TTY COMMAND clamav 335 1.12GB 874MB /usr/sbin/clamd --foreground=true mysql 1392 1.7GB 613MB /usr/sbin/mysqld --basedir=/usr -- datadir=/var/lib/mysql --p spamassassin 20347 101MB 89.3MB spamd child spamassassin 29351 97MB 84.9MB spamd child root 1047 94.5MB 84.1MB /usr/bin/perl -T -w /usr/sbin/spamd -d --pidfile=/var/run/sp Swap Used: 7.8MB / 256MB -- My Main Blog http://etbe.coker.com.au/ My Documents Blog http://doc.coker.com.au/

2 3

Re: server stats collection
by Andrew Mather 14 May '20

14 May '20

Hi, Hope the reply doesn't confuse threaded readers, I only get the Digest. > Message: 2 > Date: Wed, 13 May 2020 18:00:39 +1000 > From: Russell Coker <russell(a)coker.com.au> > To: luv-main(a)luv.asn.au > Subject: server stats collection > Message-ID: <14453232.tgyrW5GtJc@liv> > Content-Type: text/plain; charset="us-ascii" > > https://www.datadoghq.com/ > > I want to do something like what DataDog does, but with free software. > The > aim is to address the LUV server load average issue as well as other > similar > things. Below is a bunch of links to things I'm considering. I welcome > comments about any of the below or general comments about the issue that > don't > reference the below stuff. So if you have some experience to report and > don't > want to bother reading the below then please let me know. > <snip> In $LIFE-1 we used to use Ganglia for this type of monitoring on our HPC cluster (CentOS). You can install the monitor daemon on your nodes/VM's etc and it creates a series of graphs, usually with load-1 as the topmost level for all monitored machines. You can then drill down to each node and see more detail and other metrics. It monitors a bunch of stuff out of the box and is fairly extensible if you want metrics that aren't among the defaults. Uses RRDTool to store and create the graphs. The first few of the links you included looked like slightly more modern variations. May or may not be useful for this, but worth mentioning. -- - https://picasaweb.google.com/107747436224613508618 -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- "Do not go where the path may lead, go instead where there is no path and leave a trail." *-Ralph Waldo Emerson, * "Unless you're a train....that's not good advice for trains" *-Andrew Mather*

1 1

Drupal experts needed
by Russell Coker 13 May '20

13 May '20

Could anyone here help with upgrading Drupal on the LUV server? -- My Main Blog http://etbe.coker.com.au/ My Documents Blog http://doc.coker.com.au/

1 0

server stats collection
by Russell Coker 13 May '20

13 May '20

https://www.datadoghq.com/ I want to do something like what DataDog does, but with free software. The aim is to address the LUV server load average issue as well as other similar things. Below is a bunch of links to things I'm considering. I welcome comments about any of the below or general comments about the issue that don't reference the below stuff. So if you have some experience to report and don't want to bother reading the below then please let me know. https://gitlab.com/gitlab-org/gitlab-foss/-/issues/51876 Seems that Datadog was based on or inspired by statsd. https://www.tecmint.com/monitorix-a-lightweight-system-and-network-monitori… This one seems like an all in one thing that's easy to install. But the downside is that it's for only a single system. It would be useful to for example be able to display all disk IO in all VMs on the one host and see where the load is coming from when the disks can't keep up. https://wiki.openstack.org/wiki/Gnocchi Gnocchi (packaged for Debian as gnocchi-statsd) seems to be an implementation of the statsd protocol from Etsy. The idea of statsd is that you have UDP based data collection with arbitrary "bucket" names and various ways of analysing it (total of a series of numbers, number of unique IDs, etc). Then you have different tools for analysing it. https://thenewstack.io/collecting-metrics-using-statsd-a-standard-for-real-… Here's an overview of statsd operation. It mentions the ability to add monitoring to any code, there are statsd modules in Debian for Perl, Python, Ruby, Haskell, and Go. https://github.com/statsd/statsd/wiki Here is a list of links to various pieces of statsd software. https://github.com/talebook/statsd-client-cpp Here's a statsd client that monitors the basics of load, cpu, processes, network transfer, etc. -- My Main Blog http://etbe.coker.com.au/ My Documents Blog http://doc.coker.com.au/

1 0