vy32's questions

I'm using AWS Linux (which is a Centos variant). I've used SNMP before. I'm trying to change the community name from public to something1234 (for example). I'm using the standard netsnmp package that one gets by running sudo yum install snmp.

My test to see if snmp giving me access control or not is to use snmpwalk. If this works, the community is public or is being ignored:

snmpwalk -v 1 localhost -c public .1.3.6

If that gives no output but this works, I'm golden:

snmpwalk -v 1 localhost -c something1234 .1.3.6

I have yet to be golden.

The distribution creates the file /etc/snmp/snmpd.conf.

The top of the file has this:

# First, map the community name "public" into a "security name"

#       sec.name  source          community
com2sec notConfigUser  default       public

I've tried changing public in that line to something1234 but that doens't help. I've also tried adding this to the end of the file:

view    systemview      included .1.3.6.1.2
view    systemview      included .1.3.6.1.4
rocommunity something1234 localhost

That doesn't work either.

I've tried running snmpconf which is supposed to create and modify SNMP configuration files. It deletes my file /etc/snmp/snmpd.conf and creates a new file /usr/share/snmp/snmpd.conf which has a lot less information ... and which also doesn't work for something1234.

So how do I change the RO community to something1234?

We have a system with more than 4TB of swap space. I want to measure it's memory with SNMP (because I'm using MRTG to graph it). SNMP is reporting negative numbers for memTotalSwap. I suspect that this is because it's being reported with a 32-bit number:

$ snmpwalk -v 2c -c ******** localhost Memory
UCD-SNMP-MIB::memIndex.0 = INTEGER: 0
UCD-SNMP-MIB::memErrorName.0 = STRING: swap
UCD-SNMP-MIB::memTotalSwap.0 = INTEGER: -205520900 kB
UCD-SNMP-MIB::memAvailSwap.0 = INTEGER: -205520900 kB
UCD-SNMP-MIB::memTotalReal.0 = INTEGER: 784103952 kB
UCD-SNMP-MIB::memAvailReal.0 = INTEGER: 730528620 kB
UCD-SNMP-MIB::memTotalFree.0 = INTEGER: 525007720 kB
UCD-SNMP-MIB::memMinimumSwap.0 = INTEGER: 16000 kB
UCD-SNMP-MIB::memShared.0 = INTEGER: 0 kB
UCD-SNMP-MIB::memBuffer.0 = INTEGER: 412400 kB
UCD-SNMP-MIB::memCached.0 = INTEGER: 21600812 kB
UCD-SNMP-MIB::memSwapError.0 = INTEGER: noError(0)
UCD-SNMP-MIB::memSwapErrorMsg.0 = STRING:

Is there any easy workaround?

I've been using CGI since 1996 or so, so it's probably time to move to FastCGI or something else. (Flask?)

I'm using Dreamhost. I did some googling and found this article, but the article claims to be out of date and, sure enough, it doesn't work.

One of the things I really like about CGI is being able to change the script and instantly see the results. I know that flask leaves a single Python script running and re-uses it, but as a result you need to modify some file to tell flask to re-read your Python file. That seems odd---why can't flask see that the file has been modified? But I guess it makes sense for large programs that have literally hundreds of python includes.

Is FastCGI just not a thing anymore?

Amazon's marketing materials claim that the m4.16xlarge node has 64 vCPU.

When I look at /proc/cpuinfo on the system, however, I get the following information:

• 64 CPUs (CPU 0 .. 63)
• cpu cores: 16
• Processor type: Intel(R) Xenon(R) CPU E5-2686 v4 @ 2.30Ghz

So this makes no sense to me. This means that I have 64 CPUs with 16 cores each, or a total of 1024 cores.

However, online documentation for the Intel E5-2686 v6 claims that it has 36 cores and hyperthreading, for 72 virtual cores.

What's going on? How many cores are there?

An Amazon Machine Image contains an EBS volume. Is there a way to get the EBS volume out of the AMI without booting the AMI?

RHEL 7 includes rh-php70-php in the official repository. However, this module does not install mod_php.

I have tried all of the approaches for installing rh-php70-* and none of them installs and registers the appropriate php module. How does one do it?

What are the network protocols that can be used to measure the system clock (time) of a remote server? So far I have:

• Query with NTP
• TCP Timestamps
• HTTP headers

Are there any other?

It appears that Hive, Impala, Pig, and others all provide SQL or SQL-like access to data stored on Hadoop clusters. They all seem to have support for HDFS, S3, and other forms.

So why are there so many different ways for accessing Hadoop information by SQL, how are they different, and how does their performance compare?

Do we have so many different versions because all of the projects were started at the same time for more or less the same reason? If so, is there an advantage to knowing more than one of them?

I have found several articles that attempt to explain the differences (e.g. 10 ways to query hadoop with SQL and Selecting the right SQL on Hadoop, but mostly they just list features.

Is there a way to read the certificate in a consumer credit card EMV chip and use this to authenticate to an Apache web server using the client-side certificate support in SSL/TLS?

I am trying to get a modern GCC to compile on Centos 6.4. The problem is that Centos does not have a modern glibc and GCC 4.8.x and 4.7.x keep giving me the following compile error:

... -DL_gcov -c ../../.././libgcc/libgcov.c
In file included from /usr/include/features.h:385:0,
                 from /usr/include/stdio.h:28,
                 from ../../.././libgcc/../gcc/tsystem.h:88,
                 from ../../.././libgcc/libgcov.c:29:
/usr/include/gnu/stubs.h:7:27: fatal error: gnu/stubs-32.h: No such file or directory

The problem here is the gnu/stubs-32.h is part of the modern glibc and Centos 6.4 doesn't seem to have it. I've tried building my own glibc but as soon as it gets installed and in my local LD_LIBRARY_PATH I can't run any other programs, because all of the existing executables on the system try to link against it and they fail.

I want to use the new compiler because it has dramatically better handling of C++ STL code, and because the optimizer in GCC 4.8 makes my code run in 1/2 the time as the GCC 4.4.7 compiler that comes with Centos.

Any suggestions on how to do this?

We are using Lustre in a cluster with approximately 200TB of storage, 12 Object Storage Targets (that connect to a DDN storage system using QDR Infiniband), and roughly 160 quad and 8-core compute notes. Most of the users of this system have no problems at all, but my tasks are I/O intensive. When I run an array job that has 250-500 processes that are simultaneously pounding the file system typically between 10 and 20 of my processes will fail. The log files indicate that the load on the OSTs are going over 2 and that the Lustre client is returning either bad data or failed read() function calls.

Currently the only way we have of resolving my problem is to run fewer simultaneous jobs. This is unsatisfactory, because there is no way to know in advance if my workload will be CPU-heavy or I/O heavy. Besides, just turning down the load isn't the way to run a supercomptuer: we would like it to run slower when running under load, not produce incorrect answers.

I'd like to know how to configure Lustre so that clients block when the load on the OSTs goes too high, rather than having the clients get bad data.

How do I configure Lustre to make the clients block?

We would like to have a network backup system with a user that can read any file on our servers but not write any file. Is there any way to do this under Linux (and specifically Fedora)? We would rather not have a remote root that can erase any file...

We're currently using Mailman as a mailing list manager. Mailman modifies the content of mail messages. The problem is that some of our users are sending digitally signed messages and the modification makes the signature break. I've seen this behavior with Apple Mail, Outlook, and Thunderbird.

The problem seems to be this: S/MIME signed messages are implemented with a Content-Type: multipart/signed; MIME Content-Type. Mailman wraps this inside a Content-Type: multipart/mixed MIME Content-Type. None of the mail readers look inside the outer mixed for the inner signed.

We won't be able to get the clients fixed. Is there anyway to modify Mailman so that it doesn't have this behavior?

I'm new to SUSE and do not understand how to install packages with zypper. Or, rather, I understand you use the "install" subcommand, but it frequently can't find what I'm looking for.

For example, I need libpcap-devel to get the pcap.h headers and the linkable library. The SUSE website says that the package is available, but it doesn't show ups hen I try to install. Meanwhile zypper says it is installed, but the .h file isn't there:

$ sudo zypper install libpcap-dev
Refreshing service 'nu_novell_com'.
Loading repository data...
Reading installed packages...
'libpcap-dev' not found.
Resolving package dependencies...

Nothing to do.
$ sudo zypper search pcap
Refreshing service 'nu_novell_com'.
Loading repository data...
Reading installed packages...

S | Name            | Summary                        | Type      
--+-----------------+--------------------------------+-----------
  | libpcap         | A Library for Network Sniffers | srcpackage
i | libpcap0        | A Library for Network Sniffers | package   
i | slessp1-libpcap | Recommended update for libpcap | patch     
$ sudo zypper install libpcap
Refreshing service 'nu_novell_com'.
Loading repository data...
Reading installed packages...
'libpcap0' providing 'libpcap' is already installed.
Resolving package dependencies...

Suggestions?

Is there a Unix command-line program that implements ZIP with ZIP64 support? The standard zip program won't ZIP files bigger than 2**32 bytes.

I don't want to use .tar.gz format, because with ZIP I can access the files without decompressing the entire archive.

I don't want to use 7zip because there is no native Python support for 7zip archives.

Thanks.

I have a VPS running apache that keeps running out of memory at Dreamhost. Dreamhost says that the problem is the wordpress configuration, but I think that's their stock answer. I've looked and the main web hosting is coming from mediawiki. I sometimes have 20-30 http processes running, and they are all running php, and mediawiki is powering my top two sites.

So I am looking for suggestions on ways to reduce the memory footprint of mediawiki. I'm currently running version 1.16.4, which I see is significantly behind the current version. (Dreamhost is supposed to upgrade it for me, but apparently they haven't.)

• Does version 17.2 have lower footprint than 16.2?
• Is there a clever way that I could use caching to reduce the amount of memory?
• Are there configuration options that will reduce the memory?
• Why do I have both apache httpd running and php5.cgi running?
• Is there an easy way to find out which mediawiki parts are using the most ram?
• Is there a way to reduce the number of files that are fetched? My web logs are filled with fetches to user.gif, bullet.gif, external.png, document.png --- how come mediawiki's themes don't use sprites?

Thanks!

I want to download all of the cygwin files for a computer that is not connected to the Internet. I know I can download all of the cygwin files with rsync, but I'm not quite sure of the right incantation. Can you help me?

I made the mistake of purchasing a Brother MFC-9480CDW, a multi-function scanner printer. There is a "scan" button on the printer that you can configure to send a file to an FTP server. The problem that I am having is that the Brother does not put the FTP server into Binary mode, so RFC-compliant FTP servers receive the file in TEXT and new newline translation. This is causing the received PDFs to become corrupted.

I can't figure out how this could have happened. Perhaps the developers at Brother only tested this with FTP servers running on Windows? I have no idea. But anyway, I need to configure my FTP server (which is running on a Mac) to default to binary mode, not text. Alternatively I could run a different FTP server.

Anyone have a clue what to do?

My supercomputing center recently moved from SGE to pbs/Torque. Now, when I schedule my array jobs, only half of the jobs in the array get scheduled. When they finish, the other half get scheduled. This happens despite the fact that they are largely under utilized.

For example, I just scheduled an array with 10 jobs. Here is the qstat output 10 minutes later:

[myuserna@sub ~]$ qstat -t
Job id                    Name             User            Time Use S Queue
------------------------- ---------------- --------------- -------- - -----
3100[1].systemm2           ...-to-work.sh-1 myuserna        00:07:40 R short          
3100[2].systemm2           ...-to-work.sh-2 myuserna        00:07:32 R short          
3100[3].systemm2           ...-to-work.sh-3 myuserna        00:09:55 R short          
3100[4].systemm2           ...-to-work.sh-4 myuserna        00:09:44 R short          
3100[5].systemm2           ...-to-work.sh-5 myuserna        00:09:07 R short          
3100[6].systemm2           ...-to-work.sh-6 myuserna               0 Q short          
3100[7].systemm2           ...-to-work.sh-7 myuserna               0 Q short          
3100[8].systemm2           ...-to-work.sh-8 myuserna               0 Q short          
3100[9].systemm2           ...-to-work.sh-9 myuserna               0 Q short          
3100[10].systemm2          ...to-work.sh-10 myuserna               0 Q short          
[myuserna@sub ~]$ 

Any clues how to fix the scheduler?

Here is the relevant portion of the scheduler config:

create queue short
set queue short queue_type = Execution
set queue short Priority = 10000
set queue short max_user_queuable = 500
set queue short max_running = 200
set queue short resources_max.walltime = 24:00:00
set queue short resources_default.nodes = 1
set queue short max_user_run = 50
set queue short enabled = True
set queue short started = True
#

#
# Set server attributes.
#
set server scheduling = True
set server acl_hosts = systemm2
set server acl_roots = root@*
set server managers = [email protected]
set server operators = [email protected]
set server default_queue = route
set server log_events = 511
set server mail_from = adm
set server resources_default.walltime = 01:00:00
set server scheduler_iteration = 600
set server node_check_rate = 150
set server tcp_timeout = 6
set server mom_job_sync = True
set server keep_completed = 300
set server submit_hosts = submit-1
set server submit_hosts += submit-0
set server auto_node_np = True
set server next_job_number = 6217
set server max_job_array_size = 512
set server max_slot_limit = 5

We have a Macintosh OS10.6 server and a big storage system accessible via SMB.

Right now we can mount partitions, but when we mount them they are only accessible to the user that did the mount. All other users get "Access Denied."

It there any straightforward way to mount so that all users will be able to access the files? (We want all users mapped to a single SMB UID.)