Re: Something is injecting malware into my HTTP traffic
Traceroute is useless. Only traffic directed at port 80 is routed through the proxies. Nothing else, meaning that ICMP, used for traceroute, would got to the target directly. There are methods of identifying transparent proxies (you could probably ask Google about them), however, this is not one of them. Etzion On Sat, Mar 21, 2015 at 7:48 PM, E.S. Rosenberg esr+linux...@g.jct.ac.il wrote: Depending on the version of windows and it's network environment you freshly installed rootkits could be likely, but that is OT here. Note that different ISP in Israel is a fairly relative statement since there are basically just a few major players who own a bunch of the smaller ISPs and could have caching proxies on their international lines... Did you traceroute the connection both from working and non-working settings? Regards, Eliyahu - אליהו 2015-03-21 8:30 GMT+02:00 Amos Shapira amos.shap...@gmail.com: Just speculating, but could it be that your ISP uses a caching transparent proxy (which would explain why it doesn't happen on SSL) and its cache got corrupted? The other ISP case could be explained if it's actually upstream/downstream from your ISP, or they share a proxy cache for other reasons. On 21 March 2015 at 04:07, Roman Ovseitsev rom...@gmail.com wrote: Please forgive the slight off-topic, but I am experiencing a rather strange issue while downloading a certain file over HTTP. Instead of getting node.js installer as expected from here http://nodejs.org/dist/v0.12.0/node-v0.12.0-x86.msi I am receiving a completely different executable - an installer for Elcomsoft's Advanced EFS Password Recovery whatever that is. Both files are exactly the same size but SHA sums obviously don't match. SSL version of the link - https://nodejs.org/dist/v0.12.0/node-v0.12.0-x86.msi works as expected. i.e. downloads the correct node.js installer. I have verified this on three different machines running Fedora, CentOS, and Windows. None of these machines ever exchanged any files or used anything else but the default repos. In fact the windows machine is a 13 years old pc with a freshly installed OS. So presumably that dismisses any possibility of rootkits. It doesn't seems to be due to my router or ISP either. I am getting the wrong executable on two of my neighbours' Wi-Fi networks and at least one of them seems to be using a different ISP. However it doesn't happen on another Israeli nor a couple of US and UK servers I've tried so far. I am not using any proxies either. nodejs.org domain on all of the above resolves to the same IP. What's going on? Could be that the ISPs are the culprit? Considering that the application is relatively popular and I am the only one experiencing this issue it doesn't seem to be the case of nodejs.org server doing this on purpose (knowingly or not). ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il -- http://au.linkedin.com/in/gliderflyer ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Skimping on AWS EC2 bills
I believe that the time required for system start depends on the list of services. It could be shorted than two minutes, or longer. Depends. I used a condition - 'if' he can trim the image to startup in about 15 seconds, it becomes feasible. Etzion On Thu, Jan 15, 2015 at 8:11 AM, Orna Agmon Ben-Yehuda ladyp...@gmail.com wrote: Hi Amos, Etzion, You are talking about 15 seconds for bringing up the machine, and about shutting down the machine according to idleness detection. Last time I checked (and maybe I am not up-to-date), 1. It took about two minutes to bring up the machine. 2. Amazon charged per full hour. That is, if you use the instance for 20 minutes, shut it down and then bring it up for 20 minutes, you pay for two hours. So it might be beneficial to wait a bit, at least until the end of a full hour. Orna On Thu, Jan 15, 2015 at 2:33 AM, Amos Shapira amos.shap...@gmail.com wrote: Thanks Etzion. Yes you are on the same track as me. An unmapped Elastic IP will cost $3.65/month, which is a significant amount in comparison to the numbers I'm looking at skimming, so you are probably right about using a no-ip address. Finding the instance IP is a matter of a trivial curl call to the right URL, and no-ip can just use the current update requests source address automatically anyway. The next step would be to automatically identify idleness of the application for automatic shut down. Would people in the audience here see themselves using such a service (to fire up your server) if it was offered? --Amos On 15 January 2015 at 09:38, Etzion Bar-Noy eza...@tournament.org.il wrote: Hi Amos. It means you make use of an instance which is very quick to load. Removing non-esential services, or postponing them to after Jira starts, using a lightweight system, etc. If you can remove boot-time hogs, you can reach a fast-booting system. A script using Amazon API will prepare it for you. I wouldn't use the elastic IP because of its price (I get the feeling you seek something cheap). no-ip.com or other no-dns services could do the trick, except that the VM in Amazon network is unaware of its external IP (you might be able to query that using the API, BTW), and that it might take a few minutes (one, maybe more) before you could connect to the machine, because their update might no be immediate. Other than that - seems fine. Etzion On Tue, Jan 13, 2015 at 12:28 PM, Amos Shapira amos.shap...@gmail.com wrote: Etzion, just a question: Amos 0 if you can customise your instance to be very very light, - what do you mean by that? Your description is close to what I have in mind. As for the changing IP address - this can be easily overcome using Elastic IP and/or no-ip.com and friends. Thanks, --Amos On 13 January 2015 at 08:11, Etzion Bar-Noy eza...@tournament.org.il wrote: Except that NUC costs about 700+ ILS (I have three. I know. This is the Celeron version). Amos 0 if you can customise your instance to be very very light, and it can startup in about 15 seconds or so, it is acceptable to have it on-demand. You can wrap it in a script (using AWS API and tools) to just start it up. Since it will be about 15 seconds boot/startup time, you will find that very economical, and very simple to achieve. In any case, considering your requirements, this does seem to be the most simple and easy solution. Note that your IP *will* change each time you start your instance, so your API interface should also tell you what's the IP address of the machine (or you could use some no-dns service, but it will probably be slower). Etzion On Mon, Jan 12, 2015 at 8:57 PM, E.S. Rosenberg esr+linux...@g.jct.ac.il wrote: I don't know what type of load JIRA presents but for low load private stuff a raspberrypi or something similar (for heavier but still fairly 'light' stuff maybe an Intel NUC system or a mini-itx system) at home + noip/dyndns or some other form of locating it by yourself can be more then enough 2015-01-08 11:37 GMT+02:00 Amos Shapira amos.shap...@gmail.com: I was thinking about running it on my own laptop, and perhaps I will. But that would mean leaving it on around the clock which I don't want to (I'm very conscious of power consumption, both economically and environmentally), and I don't carry it with me most of the time but would like to have access to my server from both my mobile and workplace. On 8 January 2015 at 19:59, Vitaly li...@karasik.org wrote: Amos, IMHO, it's not technical, but more human issue. For example, as far as you decide that you need Jira every last day of month, you can launch instance automatically. But typically Jira usage is more random, so I don't think there is technical solution exist. If you're the only Jira user, why don't run it from your own computer for free? And, BTW, AWS reserved instances allow you to modify everything; plus up-front pay isn't must anymore. regards
Re: Skimping on AWS EC2 bills
Hi Amos. It means you make use of an instance which is very quick to load. Removing non-esential services, or postponing them to after Jira starts, using a lightweight system, etc. If you can remove boot-time hogs, you can reach a fast-booting system. A script using Amazon API will prepare it for you. I wouldn't use the elastic IP because of its price (I get the feeling you seek something cheap). no-ip.com or other no-dns services could do the trick, except that the VM in Amazon network is unaware of its external IP (you might be able to query that using the API, BTW), and that it might take a few minutes (one, maybe more) before you could connect to the machine, because their update might no be immediate. Other than that - seems fine. Etzion On Tue, Jan 13, 2015 at 12:28 PM, Amos Shapira amos.shap...@gmail.com wrote: Etzion, just a question: Amos 0 if you can customise your instance to be very very light, - what do you mean by that? Your description is close to what I have in mind. As for the changing IP address - this can be easily overcome using Elastic IP and/or no-ip.com and friends. Thanks, --Amos On 13 January 2015 at 08:11, Etzion Bar-Noy eza...@tournament.org.il wrote: Except that NUC costs about 700+ ILS (I have three. I know. This is the Celeron version). Amos 0 if you can customise your instance to be very very light, and it can startup in about 15 seconds or so, it is acceptable to have it on-demand. You can wrap it in a script (using AWS API and tools) to just start it up. Since it will be about 15 seconds boot/startup time, you will find that very economical, and very simple to achieve. In any case, considering your requirements, this does seem to be the most simple and easy solution. Note that your IP *will* change each time you start your instance, so your API interface should also tell you what's the IP address of the machine (or you could use some no-dns service, but it will probably be slower). Etzion On Mon, Jan 12, 2015 at 8:57 PM, E.S. Rosenberg esr+linux...@g.jct.ac.il wrote: I don't know what type of load JIRA presents but for low load private stuff a raspberrypi or something similar (for heavier but still fairly 'light' stuff maybe an Intel NUC system or a mini-itx system) at home + noip/dyndns or some other form of locating it by yourself can be more then enough 2015-01-08 11:37 GMT+02:00 Amos Shapira amos.shap...@gmail.com: I was thinking about running it on my own laptop, and perhaps I will. But that would mean leaving it on around the clock which I don't want to (I'm very conscious of power consumption, both economically and environmentally), and I don't carry it with me most of the time but would like to have access to my server from both my mobile and workplace. On 8 January 2015 at 19:59, Vitaly li...@karasik.org wrote: Amos, IMHO, it's not technical, but more human issue. For example, as far as you decide that you need Jira every last day of month, you can launch instance automatically. But typically Jira usage is more random, so I don't think there is technical solution exist. If you're the only Jira user, why don't run it from your own computer for free? And, BTW, AWS reserved instances allow you to modify everything; plus up-front pay isn't must anymore. regards, Vitaly On Thu, Jan 8, 2015 at 4:40 AM, Amos Shapira amos.shap...@gmail.com wrote: Yes I'm well aware of the RI option. It can save up to %70 for high-load (i.e. machines which are up 24/7), but much less saving compared to something that you can keep bringing up and down on demand. Also the up-front cost is not cheap, and commits you to that type of instance (as far as I remember, you can't buy switch or upgrade an RI slot, what's paid is paid). On 8 January 2015 at 12:47, Aviram Jenik avi...@jenik.com wrote: I'm not an AWS expert and would love to hear from those who are. But we do have a few (dozen) instances on AWS. We have them running 24/7. I get that you could start and stop on demand, but don't get how you would do that without changing the way you work in a drastic way (compared to a physical machine). To save costs, buy a 'reserved instance'. You are paying up front for 1-3 years (I recommend 3 years) and then paying a very very low cost per hour. If your load is low, buy the 'low load' machine to save even more costs (but then you pay hire fees if you cross the threshold). I don't know how this works well enough - we always buy the 'high load' instance and buy them for 3 years; the total average cost is equivalent to what we would have paid for the hosting and so the hardware is free. - Aviram On Wed, Jan 7, 2015 at 7:33 PM, Amos Shapira amos.shap...@gmail.com wrote: Hi, Do people here keep EC2 instances running? Do you leave it running 24/7 or do you fire them up when you need them? I'd like to run my own EC2 instance running $10 Jira + $10 Confluence (+$10 some
Re: Skimping on AWS EC2 bills
Except that NUC costs about 700+ ILS (I have three. I know. This is the Celeron version). Amos 0 if you can customise your instance to be very very light, and it can startup in about 15 seconds or so, it is acceptable to have it on-demand. You can wrap it in a script (using AWS API and tools) to just start it up. Since it will be about 15 seconds boot/startup time, you will find that very economical, and very simple to achieve. In any case, considering your requirements, this does seem to be the most simple and easy solution. Note that your IP *will* change each time you start your instance, so your API interface should also tell you what's the IP address of the machine (or you could use some no-dns service, but it will probably be slower). Etzion On Mon, Jan 12, 2015 at 8:57 PM, E.S. Rosenberg esr+linux...@g.jct.ac.il wrote: I don't know what type of load JIRA presents but for low load private stuff a raspberrypi or something similar (for heavier but still fairly 'light' stuff maybe an Intel NUC system or a mini-itx system) at home + noip/dyndns or some other form of locating it by yourself can be more then enough 2015-01-08 11:37 GMT+02:00 Amos Shapira amos.shap...@gmail.com: I was thinking about running it on my own laptop, and perhaps I will. But that would mean leaving it on around the clock which I don't want to (I'm very conscious of power consumption, both economically and environmentally), and I don't carry it with me most of the time but would like to have access to my server from both my mobile and workplace. On 8 January 2015 at 19:59, Vitaly li...@karasik.org wrote: Amos, IMHO, it's not technical, but more human issue. For example, as far as you decide that you need Jira every last day of month, you can launch instance automatically. But typically Jira usage is more random, so I don't think there is technical solution exist. If you're the only Jira user, why don't run it from your own computer for free? And, BTW, AWS reserved instances allow you to modify everything; plus up-front pay isn't must anymore. regards, Vitaly On Thu, Jan 8, 2015 at 4:40 AM, Amos Shapira amos.shap...@gmail.com wrote: Yes I'm well aware of the RI option. It can save up to %70 for high-load (i.e. machines which are up 24/7), but much less saving compared to something that you can keep bringing up and down on demand. Also the up-front cost is not cheap, and commits you to that type of instance (as far as I remember, you can't buy switch or upgrade an RI slot, what's paid is paid). On 8 January 2015 at 12:47, Aviram Jenik avi...@jenik.com wrote: I'm not an AWS expert and would love to hear from those who are. But we do have a few (dozen) instances on AWS. We have them running 24/7. I get that you could start and stop on demand, but don't get how you would do that without changing the way you work in a drastic way (compared to a physical machine). To save costs, buy a 'reserved instance'. You are paying up front for 1-3 years (I recommend 3 years) and then paying a very very low cost per hour. If your load is low, buy the 'low load' machine to save even more costs (but then you pay hire fees if you cross the threshold). I don't know how this works well enough - we always buy the 'high load' instance and buy them for 3 years; the total average cost is equivalent to what we would have paid for the hosting and so the hardware is free. - Aviram On Wed, Jan 7, 2015 at 7:33 PM, Amos Shapira amos.shap...@gmail.com wrote: Hi, Do people here keep EC2 instances running? Do you leave it running 24/7 or do you fire them up when you need them? I'd like to run my own EC2 instance running $10 Jira + $10 Confluence (+$10 some extra useful add-ons) (to clarify - these are one-off $10 for each product), but can't justify running a $30/month small EC2 (and perhaps more, Jira alone requires 1.5-2GB of RAM) just to be used at most a few hours a month if not less. But logging in to the console to fire it up (or through aws cli, or using an Android based app) every time I want to access it also would be inconvenient. So is there another way? Thanks, --Amos ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il -- http://au.linkedin.com/in/gliderflyer ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il -- http://au.linkedin.com/in/gliderflyer ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list
Re: Question about Linux installation and selecting a language
US International. Etzion On Mon, Jan 12, 2015 at 9:14 PM, Dan Shimshoni danshi...@gmail.com wrote: Thanks , Oleg use a somewhat older Fedora, so YMMV, but in KDE open System Settings - Input Devices - Keyboard Settings - Layouts and select what suits you (I assume American + Israeli?). There is no American English there! Since the installation was with British English! what should I do so that American English will be available there ? regards, Dan On Mon, Jan 5, 2015 at 9:19 PM, Oleg Goldshmidt p...@goldshmidt.org wrote: Dan Shimshoni danshi...@gmail.com writes: I had installed Fedora 21 on x86_64 (server, and with kde). When installation started, I selected the language as british english and not American English. I am guessing the installation chose a British keyboard for you. This is a bit surprising because Red Hat, unlike some other distros (we'll protect the guilty) usually ask a separate queston about keyboard layout. Now when I log in and press some keys, I get unwanted results. For example, when pressing @ I get the character. [snip] Any ideas how this problem can be resolved ? I use a somewhat older Fedora, so YMMV, but in KDE open System Settings - Input Devices - Keyboard Settings - Layouts and select what suits you (I assume American + Israeli?). -- Oleg Goldshmidt | p...@goldshmidt.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: sending to same dest via different interfaces
Sorry. My default was not 'send to all'. Now posting to the list as well. On Wed, Mar 5, 2014 at 2:58 PM, Elazar Leibovich elaz...@gmail.com wrote: Another idea from colleague, is to bind the source address of the socket to the address of the desired netwrokr interface. While it doesn't guarantee anything, he said that in practice the kernel routed the packets through the desired network interface. On Tue, Mar 4, 2014 at 10:02 AM, Erez D erez0...@gmail.com wrote: Hello I have 2 external interfaces via two eth cards, both connected to the internet I want to send a udp packet to same host:port, but choose dynamically which interface to use. can this be done with linux, and how ? 10x erez. ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: sending to same dest via different interfaces
Eventually I will master the secret art of sending e-mail messages to the right persons, with the right content. A secret and holy art it is... Assuming you want the paths to be selected dynamically based on load, you can use iputils2 'ip route' to create a multi-homed routes. You will not have a clean control over how a specific packet goes, however, you will be able to split the communication between the two interfaces to whichever ratio you desire. Lots of info on the Internet. Ez On Thu, Mar 6, 2014 at 12:41 AM, Etzion Bar-Noy eza...@tournament.org.ilwrote: Sorry. My default was not 'send to all'. Now posting to the list as well. On Wed, Mar 5, 2014 at 2:58 PM, Elazar Leibovich elaz...@gmail.comwrote: Another idea from colleague, is to bind the source address of the socket to the address of the desired netwrokr interface. While it doesn't guarantee anything, he said that in practice the kernel routed the packets through the desired network interface. On Tue, Mar 4, 2014 at 10:02 AM, Erez D erez0...@gmail.com wrote: Hello I have 2 external interfaces via two eth cards, both connected to the internet I want to send a udp packet to same host:port, but choose dynamically which interface to use. can this be done with linux, and how ? 10x erez. ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Astrerisk question, anyone sell a cellphone that can be used to make calls?
Sorry. Should have been reply all. Not to mention that Hot sells phone lines with unlimited plan at 69 ILS/month, if I'm not mistaken, which leaves you with a simpler solution. Way simpler and cheaper than connecting cellular line... Ez On Sun, Jun 10, 2012 at 4:50 PM, Nadav Har'El n...@math.technion.ac.ilwrote: On Sun, Jun 10, 2012, shimi wrote about Re: Astrerisk question, anyone sell a cellphone that can be used to make calls?: Be advised that this probably qualifies as non fair-use usage, which is forbidden in a clause in every unlimited contract I've seen to date (with Also, I wonder, why are you trying to do this anyway? In the old days, calls from Bezeq landlines to cellphones were very expensive - around 1 shekel a minute - while cellphones had deals to call the same network (e.g., 50 agorot a minute to call from cellcom to cellcom) - so it made sense for a company to route their outgoing cellphone calls through physical cellular phones, and similar tricks. But today - the cost of calling a cellphone is down to (I believe) 6 agorot a minute - about the same as calling a regular landline. So why do you need to make calls through actual cellphones? Nadav. -- Nadav Har'El|Sunday, Jun 10 2012, n...@math.technion.ac.il |- Phone +972-523-790466, ICQ 13349191 |[I'm] so full of action, my name should http://nadav.harel.org.il |be a verb -- Big Daddy Kane (Raw, 1987) ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Preparing to convince to shift to non-propriety documents formats
Because most households in israel do not buy their office... It would be stupid to assume they do. Moreover - the school headmaster does not assume that either. He/she knows most people just have their office installed, and they care nothing about it. Ez On Feb 5, 2012 9:12 AM, Nadav Harapos;El n...@math.technion.ac.il wrote: On Sun, Feb 05, 2012, Jonathan Ben Avraham wrote about Re: Preparing to convince to shift to non-propriety documents formats: Hi Boaz, The time is not ripe. Don't waste your energy. Your school principal will not know what you are talking about and will dismiss you as a hopeless geek. I am not sure about the time not being ripe. In the last year, I installed for two non-technical family members a copy of OpenOffice (one a full fledged Linux, but the other a compromise Windows+OpenOffice). They both faced a few hardships when people sent them Microsoft Office documents and they didn't look exactly as expected, but I was able to convince them that it was in fact the other person who is behind the times ;-) And the documents *were* readable, even if didn't look perfect. And for users, this is a saving of 500 shekels (last time I checked). I don't see how this fact can be ignored in Israel after the summer's protests. This is actually the reason why I installed OpenOffice in these two cases - it's hard to justify adding 500 shekels to the price of a computer which cost around 1000 shekels (plus a few hundred more for the legal Microsoft Windows). A slightly more productive line might be to claim that you are no longer using desktop computers - only mobile devices, and for these you need either PDF or Google docs. I believe my Android can read Microsoft Office documents out of the box :( But it's true, with all these non-Microsoft devices around, Microsoft's stranglehold on the word processor document seems to be coming to an end. -- Nadav Har'El| Sunday, Feb 5 2012, n...@math.technion.ac.il |- Phone +972-523-790466, ICQ 13349191 |It's fortunate I have bad luck - without http://nadav.harel.org.il |it I would have no luck at all! ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Preparing to convince to shift to non-propriety documents formats
I agree with the notion. At best - the school headmaster will ask you what
to open it with, and how would the students with ms office would handle it.
This is hopeless, unfortunately.
Ez
On Feb 5, 2012 8:48 AM, Jonathan Ben Avraham y...@tkos.co.il wrote:
Hi Boaz,
The time is not ripe. Don't waste your energy. Your school principal will
not know what you are talking about and will dismiss you as a hopeless geek.
A slightly more productive line might be to claim that you are no longer
using desktop computers - only mobile devices, and for these you need
either PDF or Google docs.
I found that this approach, mobile devices, works. For example, at the
American School in Even Yehuda it helped convince teachers to accept and
give assignments in PDF or via Google docs. The techers made this head
switch about three years ago when the younger students who were the early
technology adopters demanded it. It didn't come from the principal, and not
from the parents either, both groups being generally clueless.
- yba
On Sun, 5 Feb 2012, Boaz Rymland wrote:
Date: Sun, 5 Feb 2012 07:20:39 +0200
From: Boaz Rymland boaz.ryml...@gmail.com
To: linux-il linux-il@cs.huji.ac.il
Subject: Preparing to convince to shift to non-propriety documents formats
Hi all,
I'm about to meet my daughter's school principal on the subject of the
formats of documents the school spreads around
routinely, like the weekly schedule. In short - they are using .DOC MS
Word format and I don't like it as I cannot
cleanly open those documents on my computer (which runs Ubuntu).
Although I'm quite old in the Linux world and probably heard over the
years most of them - still I prefer having a
refreshment of all the arguments in favor of moving to more open or at
least affordable (e.g. PDF) document formats.
Any pointers/text will be appreciated.
Thanks,
Boaz.
--
EE 77 7F 30 4A 64 2E C5 83 5F E7 49 A6 82 29 BA~. .~ Tk Open
Systems
=}**ooO--U--**
Ooo{=
- y...@tkos.co.il - tel: +972.2.679.5364, http://www.tkos.co.il -
___
Linux-il mailing list
Linux-il@cs.huji.ac.il
http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
___
Linux-il mailing list
Linux-il@cs.huji.ac.il
http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
___
Linux-il mailing list
Linux-il@cs.huji.ac.il
http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Virtual Server - Consult...
Mind you that the level of user experience would depend on the speed if the display adapter. This is not a simple requirement in a virtual environment. None of the desktop-level virtualization solutions would give you that. Display will be slow, and with it - the entire user experience. You need something with vga pass through. This is what you need to search for. Etzion On Jan 23, 2012 12:36 AM, Nadav Harapos;El n...@math.technion.ac.il wrote: On Sun, Jan 22, 2012, Robert Wallner wrote about Re: Virtual Server - Consult...: I think it depends on what operating system will run inside those VMs. Another option would be also qemu. You're probably thinking of qemu with KVM, in which case it's the same option as the KVM option raised already. Using qemu *only*, without KVM (or Xen) is not a sensible option on today's hardware. It is very slow, and KVM significantly speeds it up by using hardware virtualization (Intel VMX or AMD SVM) which is present on all x86 PCs manufactured in the last 5 or so years. Nadav. -- Nadav Har'El|Monday, Jan 23 2012, n...@math.technion.ac.il |- Phone +972-523-790466, ICQ 13349191 |Bumper sticker on stealth bomber: If you http://nadav.harel.org.il |can read this, we wasted a lot of money! ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: question: discount bank web site - linux+firefox friendly?
To my knowledge this bank uses the international bank backbend systems, and it works with linux quite well. Etzion On Jan 19, 2012 10:51 AM, Nadav Harapos;El n...@math.technion.ac.il wrote: On Thu, Jan 19, 2012, guy keren wrote about question: discount bank web site - linux+firefox friendly?: i'm considering switching to Discount bank (www.discountbank.co.il) - and i want to make sure that their online banking works with firefox on linux for these operations: .. if anyone here is doing one or more of these operations successfully with firefox *on linux* - could you share your experience? I have exactly the same question, about Bank Otsar Hahayal - Does it work properly with Firefox? And like Guy, by properly I don't mean just the front-page, but all the various operations starting with viewing the balance and ending with buying and selling stocks. Thanks, Nadav. -- Nadav Har'El| Thursday, Jan 19 2012, n...@math.technion.ac.il |- Phone +972-523-790466, ICQ 13349191 |Learn from mistakes of others; you won't http://nadav.harel.org.il |live long enough to make them all yourself ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: [OFFTOPIC] Finding the next lucrative niche (was: Re: Goodbye, Lingnu)
True. However, the best support can be given by a pro who has a cooperation of the hosting company. Most of them will never cooperate, as they earn better (despite their awful quality of work) with an external person/company. So, your main channel - the hosting supplier, will probably not like you using it as a marketing channel, and you can hardly ever know (without some extensive search or cracking skills) which hosting service holds which customers. Ez 2011/11/14 Hetz Ben Hamo het...@gmail.com Hi, I'll give you one: Supporting hosting customers. Explanation: there are over 40 businesses here in Israel which provides VPS and dedicated server renting solutions. Most of those businesses (including mine) provide the machine to the customers as unmanaged, which means: the support you'll get relates *only* to network (if your VPS/Server) can communicate to and from the server The other part of support: shutting down/restart/pause/reset the server. Anything else - you're on your own. You can purchase a bank of hours for support or you can get emergency support for a higher price. There are, of course, some hosting businesses which will give you managed service, but in most case I would strongly recommend to get a freelance who will do the job for the client. Why? because they *SUCK AT IT* ! spreading root privileges all over the place, changing permissions of directories and files to 777, opening the internal firewall (if it exists, most of the time it's not) to everyone+his dog, and other nonsense stuff, and I'm also talking about the *BIG* providers which are also ISP - they suck, period! I think that this niche is mostly empty and people can get some jobs/contracts with such customers. If anyone want to register himself, I have a neutral forum for that. see here: http://hosts-forum.com/index.php Hope this helps, Hetz On Mon, Nov 14, 2011 at 3:02 PM, Omer Zak w...@zak.co.il wrote: We need some process for identifying the next niche to pursue, taking into account current skill set, customers/contacts, and effortexpense incurred in acquiring the competencies relevant for the next niche. Did anyone blog about such a process? -- *חץ בן חמו חץ-ביז *השכרה ואירוח של שרתים פיזיים מעוניין להשתמש בשרותים שחסומים לגולש הישראלי? Hulu? NetFlix? Pandora? Google Voice? אם כן, היכנס לכאן http://vps.net.bz/?p=406. ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: [OFFTOPIC] Finding the next lucrative niche (was: Re: Goodbye, Lingnu)
It doesn't matter. Because a customer needs a method of estimating a work, does not make you his bitch(tm). You can estimate your work, but you seldom work in a vacuum. You are working on his servers, on his setups, his storage devices, around his network equipment. Under most cases, when I get delayed, it's because of the customer, and not me. Example. One of the newer cellular suppliers has demanded that I install a large system in their new server farm (16 blades and a storage). I have prepared the system, and on the delivery day (which was very time-pressed due to the customer), I have discovered that they had not installed 220v AC there yet. Can't do anything with the servers. The funny thing is that during that time, two or three weeks since, the system is just sitting there, when the server room is being built around it. Ez On Mon, Nov 14, 2011 at 4:33 PM, Nadav Har'El n...@math.technion.ac.ilwrote: On Mon, Nov 14, 2011, geoffrey mendelson wrote about Re: [OFFTOPIC] Finding the next lucrative niche (was: Re: Goodbye, Lingnu): Long ago I got out of the PC repair/support business when the 50 NIS an hour people took over. Am I the only one bothered by all this per hour talk? What prevents a 50-NIS-per-hour consultant from telling you he worked for 4 hours on your problem, which it actually took him just one hour in his lab? Or a 200-NIS-per-hour consultant from finding your problem in 10 minutes, because he solved exactly this problem a week ago or wrote the necessary code already? So how is the price per hour a meaningful variable when comparing two consultants? -- Nadav Har'El|Monday, Nov 14 2011, n...@math.technion.ac.il |- Phone +972-523-790466, ICQ 13349191 |Cats aren't clean, they're just covered http://nadav.harel.org.il |with cat spit. ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Checkpoint VPN client for Linux - is there any?
Snx is for their ssl snx product. It will not work where officeconnect should be deployed. Ez On Oct 27, 2011 7:13 PM, Noam Meltzer tsn...@gmail.com wrote: If I remember correctly CP has some kind of plugin/extension/some other kind of lie called snx. Or at least snx was the utility for linux which was the VPN client. You need some kind of license per user for that in the firewall. It worked for me several years ago with Ubuntu 08-04 (after a little hacking, of adding some missing '.so') 2011/10/27 dyasny dya...@gmail.com On Thu, Oct 27, 2011 at 6:49 PM, Oleg Goldshmidt p...@goldshmidt.orgwrote: Hi everyone, I would like to find a VPN client solution for Linux that is compatible with what I currently use on Windows (in a VM hosted on Linux): Checkpoint's Endpoint Connect NGX R73. It does not look like Checkpoint offer a Linux VPN client (of any kind? the VPN Client Download page has a Linux tab but there is nothing there). Is this correct? I do not know enough to decide for myself (after googling around) whether 1. the whole thing is hopeless and there is nothing compatible; 2. there may be 3rd party/FOSS clients that are compatible with some Chekpoint solutions and not others and more information is needed to pick one or more to try; 3. one can just grab this thingy here and see if it works. Just try to use raccoon or vpnc, I remember having either work better than the original CP client (which was only available for RH7, so no wonder it's ancient and horrible) I don't own the server side but I can ask questions - the trick is to know what to ask. Any advice? Thanks! -- Oleg Goldshmidt | p...@goldshmidt.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: remote directory/partition
Nutrino will enforce connection-less transport. I am not sure it is desired.
He will have to emulate connection tracking by software, which will pay
these few nanoseconds earned earlier. Not worth it, I think.
Gravity generator. The speed of gravity is still unknown, but might be very
very fast, and will enforce low-latency and cross-barrier transport across
continents. I don't think, however, you can find a gravity generator in the
commodity market. Not sure if you can find any at all :-)
Ez
On Sun, Oct 23, 2011 at 10:24 AM, Jonathan Ben Avraham y...@tkos.co.ilwrote:
On Sun, 23 Oct 2011, Yedidyah Bar-David wrote:
Date: Sun, 23 Oct 2011 09:27:34 +0200
From: Yedidyah Bar-David linux...@didi.bardavid.org
To: Hetz Ben Hamo het...@gmail.com
Cc: ILUG linux-il@cs.huji.ac.il
Subject: Re: remote directory/partition
On Sat, Oct 22, 2011 at 11:03:49PM +0200, Hetz Ben Hamo wrote:
Hi,
Here is a theoretical question:
Lets say I have a Linux server in Israel, and I have a block of storage
(lets say iSCSI partition for this example) in USA, and I want to mount
it
on my server in Israel.
iSCSI over such a long distance and with big latency (thanks to our
ISP's)
Not sure it's mainly the ISPs, BTW. You do also depend on the physics of
speed of light.
If you use IP over nutrino-based transport you might be able to shave a few
nanoseconds off the speed of light, see this: http://www.wired.com/**
wiredscience/2011/09/**neutrinos-faster-than-light/http://www.wired.com/wiredscience/2011/09/neutrinos-faster-than-light/
Shavua tov,
- yba
is a big no no, it's too slow. NFS is also not a good idea (here's
whyhttp://goo.gl/vn4GM
).
I can take this storage, format it and export it from my server in USA,
but
which protocol would give me:
1. All (or almost all) functionality of a local mounted device
Do you need it read/write on both sides? If so, you are going to have
big problems if the link is cut.
2. Can work with long distance latencies
3. won't kill the machine if the remote directory is disconnected /
disappeared
4. If possible - supported (either directly or using 3rd party driver)
on
Windows 2008 (Linux is the main concern, Windows is optional)
I used drbd on a LAN, and know that it can theoretically work rather well
on larger distance when used as read-write on one side only. They also
have a pay-for tool to do this asyncronously called drbd proxy. This
implies using a local copy and have drbd sync it. You can choose between
three what they call Protocols to affect the perceived local latency.
--
EE 77 7F 30 4A 64 2E C5 83 5F E7 49 A6 82 29 BA~. .~ Tk Open
Systems
=}**ooO--U--**
Ooo{=
- y...@tkos.co.il - tel: +972.2.679.5364, http://www.tkos.co.il -
__**_
Linux-il mailing list
Linux-il@cs.huji.ac.il
http://mailman.cs.huji.ac.il/**mailman/listinfo/linux-ilhttp://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
___
Linux-il mailing list
Linux-il@cs.huji.ac.il
http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: New (First!) smart phone (OT_
Welcome to the world of the cellular. First - there are various keyboards available for the device. Their key size might differ, layout might differ, and ease of use, for you in person, would differ as well. About stylus - you will need a device designed for capacitive screen. Although I am not familiar with your device in particular, most devices today are capacitive, meaning they react to the human electrical charge, our something like that. You can purchase such a device for very cheap at deal extreme, or for very expensive in Israel. Service quality differ depending on many factors. Usually, passing cars are not part of them. It could be that your friend's land line phone was a wireless one, and it's all his fault. Could be that you are in the area of overlapping cellular antennas, and your call switched between cells. Try to wait before judging the system just a little while. Most people do not feel that land lines are superior compared to cellular communication. This could be because, from the perspective of the caller - it is not. Good luck and enjoy your new device. Ez On Oct 4, 2011 11:28 PM, Stan Goodman stan.good...@hashkedim.com wrote: As per Subject, this is my first smart phone, and I am engaged in trying to discover if I myself as smart as it is; the outlook is bleak. More than it should be because the documentation seems to be written for a reader who already knows the score and needs only a bit of memory nudging to recall the drill. As OT as this is, I hope I can ask here for answers to a few questions. The instrument is HTC Aria. The two questions that are bugging me at the moment are: 1) The on-screen keyboard for writing messages defeats me, because the keys are absurdly narrow, certainly more so than my fingers, and there is absolutely no chance that I would ever be able to peck out even a short coherent message with them. This suggests that there must be a way to type with a stylus, rather then directly with fingers. I have tried to use objects made of various materials, but so far nothing works. How do people type on these things? 2) I succeeded in telphoning to a friend this evening, to his land-line phone. For a while, the sound I got was quite good and distinct, but after a bit, it began to fade in and out, so that cconversation became intermittent and impossible. There is a cell-phone tower less than a kilometer away from my home. Is this the level of service that I have to look forward too? The fading suggests that the communication may be affected by passing vehiles; is this possiblle over such a short distance? -- Stan Goodman Qiryat Tiv'on Israel ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: what was the name of the app...
OpenMosix, but it I's hardly usefull for most usages, old, not really maintained, and very expensive. Why do you need it? Ez On Aug 14, 2011 9:07 PM, Hetz Ben Hamo het...@gmail.com wrote: Hi, Few years ago I heard about an app which can connect few servers behind it - and show itself as a single cpu (single machine), so if you ran an application on this app, it would do the magic of dividing parts to other servers and combining them back. Anyone remember the application name or URL for it? Thanks, Hetz ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Finally - A RMS talk in Tel-Aviv. Including details
+1 to that. Etzion On Jul 17, 2011 8:07 PM, Dima (Dan) Yasny dya...@gmail.com wrote: On Sun, Jul 17, 2011 at 7:46 PM, Orr Dunkelman orr.dunkel...@gmail.com wrote: On Sun, Jul 17, 2011 at 7:19 PM, Michael Shiloh michaelshiloh1...@gmail.com To be more general, it is unheard of for a prospective guest to set political conditions for his hosts. Aside of any other aspects, the man is a boor, and should have been ignored. I'm not sure that's true. Haven't musicians and other performers used their platform as a way to express agreement or disagreement with particular political positions for ages? Michael It is very rare in computer science to mix politics and work. Actually, this is true for most exact sciences. As for RMS, who is an advocate of freedom of rights, he has just agreed to accept money (travel costs) in exchange for giving away his freedom of speech. Tz'e Wolmad... /me don't see what the argument is about. I'll simply vote against RMS's and his politically inclined friends bs with my feet by _not_ going. End of story -- Orr Dunkelman, orr.dunkel...@gmail.com ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il -- Regards, Dan Yasny ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: UDP packets loss at Israeli ISPs during peak hours
It's probably due to their over commit, and it means your UDP packets are in queue until expired. Since Israeli ISPs has learned the trick, I tend to believe ICMP packets have high priority, so that no customer will be able to complain. When you can't complain, well, it means that the problem is with you. Ez 2011/7/3 Arie Skliarouk sklia...@gmail.com Hi, The company I work at uses openvpn extensively. We settled on UDP-based protocol as it is more effective than TCP based. Inter-Israeli VPN connection works perfectly all of the time, whereas international VPN has erratic behavior on at least one ISP. I suspect the ISP (XFone 018) dropping UDP packets occasionally during peak hours for following reasons: - ICMP ping to the internet-facing IP number of the VPN router works properly all of the time - over-VPN ping to some server has about 50% packet loss during peak hour (tested at 23:00) - on different ISP at the same time there was no packet loss - over-VPN ping on the same ISP worked perfectly in the morning hours Have anyone else noticed the same behavior? What is legal status of such network traffic policing? -- Arie ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: RMS, Hosts Must Support Boycott?
I have been reading quite a lot of messages about this topic. I feel (and this is my feeling only) that it has been talked enough. Mr. Stallman doesn't want to be here. He doesn't care about the other side of the story, and he is so much about his own religion, that he forgets that each coin has two sides. I say - Halas (it was in Arabic, too!). Let him be. A boycott is as effective as the boycottee cares. I don't care anymore. I won't go there, because, to my eyes, such a strong opinionism without even an attempt to see the other side, as Mr. Stallman shows both in his recent dealings with the local politics and local conflict between two nations who have nothing to do with him, and showed in the past regarding various concepts of how and what to code (some of them were enough to make a movement, to start something, but they are not necessarily the thing which is required to keep it going), and how things should be, shows me how irrelevant he is. Blocking your ears to alternate views is a bad thing. We should not encourage such actions, neither of strangers, nor of ourselves. We have the option to show our point of view, and that means, that if it were to me, the lecture, if at all, would have taken place somewhere near Gaza, so he experiences the rockets flying over our heads. Because there is no contravention about these locations. But this is nothing more than an imagination. Wasted time, just like this e-mail message. It doesn't matter. He doesn't care, he doesn't want to look at the other side of the coin, and by boycotting us, he exposes his own stubbornness. Halas. Don't waste your time with him. Etzion 2011/6/13 Tom Balazs tom123onl...@gmail.com I would have been happy to say this thread had run it's course, but this response from RMS and (his apparent associate / organizer) Kobi Snitz is too much. Mr. Snitz advises both myself and RMS that in order to make sure the talk in Israel meets the standards of the BDS-supporting organization which is hosting RMS then the talks in Israel must be both sponsored by a BDS supporting organization and held in a hall owned/run by a BDS supporting organization. Snitz's words were, in order to not violate the boycott the sponsors and the place should be dissenting israeli groups and institutions. It seems that Mr. Snitz is a mathematician, anarchist, and leader in the BDS (boycotts, divestment and sanctions) against Israel movement. If you'd like to read more about Mr. Snitz you can use the links below. (Responses from Kobi Snitz and from RMS are also below.) http://www.haaretz.com/print-edition/features/shin-bet-puts-israeli-anarchists-in-crosshairs-1.333140 - He was not the only anarchist the Jewish department dealt with that week. Five days earlier, Kobi Snitz was attending a conference when he received a call from an unidentified number. The caller told him, 'Shalom, this is Rona from the Shin Bet. I'm sure you've heard about me.' - She said she wanted to invite me for a friendly conversation and for us to exchange thoughts, said Snitz, 39, an anarchist activist and a mathematician. He asked whether he was being called in for an interrogation and when she said no, he said, no thanks. In 2009, Snitz served a 20-day sentence over an attempt a few years earlier to prevent the demolition of a house in Kharbatha, a village west of Ramallah. Two months ago, he was given another five-day sentence over a protest against the Second Lebanon War in 2006. http://www-users.math.umd.edu/~snitz/ http://israel-academia-monitor.com/index.php?type=large_advicadvice_id=7985page_data[id]=175cookie_lang=en http://www.radicalendar.org/calendar/all/all/display/85445/index.php?view=eventfulldate=2009-07-04 http://www.docstoc.com/docs/38551113/In-support-of-the-Palestinian-human-rights-community-call http://student.cs.ucc.ie/cs1064/jabowen/IPSC/php/authors.php?auid=4993 http://pipl.com/directory/name/Snitz/Kobi http://israel-academia-monitor.com/index.php?type=large_advicadvice_id=6410page_data[id]=175cookie_lang=enthe_session_id=0e356a3b68ae0e5dd29aefaf7ef56e77BLUEWEBSESSIONSID=21f827e2775005995748ebab68617a46 http://www.israel-academia-monitor.com/index.php?type=large_advicadvice_id=6189page_data[id]=162cookie_lang=hethe_session_id=cfbd3321f483be8d97e01a3d1db248bfBLUEWEBSESSIONSID=049669b2ba65f1c3fedf5a76c53d45aa --- fromKobi Snitz ksn...@gmail.com tor...@gnu.org dateMon, Jun 13, 2011 at 09:30 subjectRe: Finding a hall in Haifa mailed-bygmail.com signed-bygmail.com hide details 09:30 (1 hour ago) I am missing the earlier part of this correspondence but I can say that in order to not violate the boycott the sponsors and the place should be dissenting israeli groups and institutions. In tel aviv my first thought was the women's coalition for peace. I talked to their head last night and she said that they've never done anything about free software and she feels it is strange for them to do it. The upcoming boycott law
Re: An alternative to Skype
By logging in to their site, you can reactivate your credits. Ez 2011/6/10 Uri Even-Chen u...@speedy.net Hi people, I saw Richard Stallman's signature: Skype: No way! That's nonfree (freedom-denying) software. Use free telephony http://directory.fsf.org/category/tel/ I'm using Skype to communicate with my friends. I mostly use it for instant messaging (not phone calls). I did use it for phone calls in the past, but I don't have a microphone attached to my computer - so today I use regular phone (018). When I used Skype for phone calls, I paid Skype for Skype out credit, but Skype zeroed my credit after a few months of not using their service - which I consider stealing money from me. For all that reasons, I don't want to use Skype. But Skype is a monopoly - all my friends use it. Even my grandmother wants to start using it. What free alternatives are there to Skype? (free as in freedom of course). By the way, what do you think about Google talk? I use Google talk from my Gmail account (I didn't download the software). I'm using Windows XP on PC. And Google Chrome as my browser. Thanks, Uri Even-Chen Mobile Phone: +972-50-9007559 E-mail: u...@speedy.net Website: http://www.speedy.net/ ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Wanted - Linux sysadmin for Integration
Hi. Sorry for the cross posting. Wanted A small boutique integration-shop looks for a proficient Linux administator. He or she should answer these requirements, by this order: Well acquainted with Linux Knowledge of scripting – Shell, Perl or Python Understanding of storage – internal (LVM, Raid) and external (NAS, SAN, DAS) Acquaintance with databases (from the perspective of sysadmin) such as Oracle, MySQL Acquaintance with PC hardware, especially server-class Familiarity with various common network services Due to the nature of this position, driver license is mandatory. He or she are required to be independent, assertive, ready to learn (and a lot), knack for automated tasks and a corporate point of view. I am outing myself a bit, but you are most invited to search for my name, Etzion Bar-Noy or my nick name, ezaton, on the net. You are most invited to read my technical blog, at http://run.tournament.org.il Please send your CV to etz...@barnoy.co.il. You are most welcome asking for more info by e-mail. (and no. I did not find the time to set my web site just yet. Sorry for that) Thanks! Etzion ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Centos 5/5.5/6.0 and StorageWorks P2000
Probably. At worst, you can download the 'megaraid' driver and compile/run it. You should have boot driver floppy image in case you need it for boot and you can't find it. Ez 2011/5/25 Hetz Ben Hamo het...@gmail.com Here is the output which mentions the SAS: 07:00.0 Serial Attached SCSI controller: LSI Logic / Symbios Logic SAS2008 PCI-Express Fusion-MPT SAS-2 [Falcon] (rev 03) Subsystem: Hewlett-Packard Company Device 3371 Flags: bus master, fast devsel, latency 0, IRQ 10 I/O ports at 6000 [size=256] Memory at c7efc000 (64-bit, non-prefetchable) [size=16K] Memory at c7e8 (64-bit, non-prefetchable) [size=256K] [virtual] Expansion ROM at c200 [disabled] [size=512K] Capabilities: [50] Power Management version 3 Capabilities: [68] Express Endpoint, MSI 00 Capabilities: [d0] Vital Product Data ? Capabilities: [a8] Message Signalled Interrupts: Mask- 64bit+ Count=1/1 Enable- Capabilities: [c0] MSI-X: Enable- Mask- TabSize=15 Capabilities: [100] Advanced Error Reporting UESta: DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq- ACSVoil- UEMsk: DLP- SDES- TLP- FCP- CmpltTO- CmpltAbrt- UnxCmplt- RxOF- MalfTLP- ECRC- UnsupReq+ ACSVoil- UESvrt: DLP+ SDES+ TLP- FCP+ CmpltTO- CmpltAbrt- UnxCmplt- RxOF+ MalfTLP+ ECRC- UnsupReq- ACSVoil- CESta: RxErr- BadTLP- BadDLLP- Rollover- Timeout- NonFatalErr- CESta: RxErr- BadTLP- BadDLLP- Rollover- Timeout- NonFatalErr+ AERCap: First Error Pointer: 00, GenCap+ CGenEn- ChkCap+ ChkEn- Capabilities: [138] Power Budgeting ? IIRC, the centos 5.5 installer should recognize it without any issues, am I right? Hetz 2011/5/24 Baruch Even bar...@ev-en.org On May 24, 2011 1:57 PM, Hetz Ben Hamo het...@gmail.com wrote: Hi, I'm trying to assist someone. He has an IBM Server and he's planning to install CentOS 5.5 Part of the setup is a storage box (which he is not administrating) which is connected to the IBM with SAS cable. Are there any special instructions while doing the install to detect and use this SAS connected storage? (the box has currently OpenSuSE and when connecting the SAS cable, nothing showed up on /var/log/messages).. You are missing the sas controller driver probably. Try to do lspci to find the sas controller model and from there you'll need to find the driver. Baruch -- *חץ בן חמו חץ-ביז (הוסטינג) *השכרה ואירוח של שרתים פיזיים השכרת שרתים וירטואליים מקצועיים וגדולים במחירים *קטנים* בקרו באתרנו בכתובת hetz.biz http://www.hetz.biz/ ובבלוג שלנו: blog.hetz.biz טלפוןן: 078113/4/5, אימייל: sa...@hetz.biz מסנג'ר: sa...@hetz.biz - סקייפ: heunique ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Centos 5/5.5/6.0 and StorageWorks P2000
You should download and install the latest lsi logic sas driver. As Baruch said - check lspci for the exact model. Ez 2011/5/24 Baruch Even bar...@ev-en.org On May 24, 2011 1:57 PM, Hetz Ben Hamo het...@gmail.com wrote: Hi, I'm trying to assist someone. He has an IBM Server and he's planning to install CentOS 5.5 Part of the setup is a storage box (which he is not administrating) which is connected to the IBM with SAS cable. Are there any special instructions while doing the install to detect and use this SAS connected storage? (the box has currently OpenSuSE and when connecting the SAS cable, nothing showed up on /var/log/messages).. You are missing the sas controller driver probably. Try to do lspci to find the sas controller model and from there you'll need to find the driver. Baruch ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: auto-maximize a logical partition with ext3
DD is a lousy OS replication tool, and I would not have used it. Using simple (!!!) scripting to replicate systems, with any possible combination of sfdisk, LVM, mkfs.ext3, resize2fs would probably be better, not to mention - faster. Example: Boot; create a new partition layout on your new disk using sfdisk and an answer file (hint: read the man pages); format partitions (all full size, yes?); mount partitions; replicate using 'tar' over network, or over NFS share (better do it from a single tar file, or compressed tar file. It's faster to decompress than to compress); mount/tar remaining partitions; fix boot loader; reboot I have done so countless times in a set of automated tools. Easiest possible solution. DD seems faster on first look, but sets so many additional problems (amongst is the one you have described. Others deal with different hardware layout), and you copy lots of dead nothing. Ez On Sun, Apr 3, 2011 at 8:19 PM, Ira Abramov lists-linux...@ira.abramov.orgwrote: Quoting guy keren, from the post of Sun, 03 Apr: if it is the last partition, how about using expect (or similar) to automate 'fidk' on the device (probably need to set a loop device on the image file first), delete the partition and re-create it? as far as i remember, when you create a partition via fdisk's interactive prompt, by default it suggest to use all the available extra space. if it's a logical partition - you'll need to delete the underlying logical part as well, i assume. that solution has come to mind, and I have seen one case where it was used (to create, not resize). there's no need even for expect, you can simply redirect the input because the interaction is well known in advance, but it's a very ugly solution I hope not to have to implement for a dozen reasons. At the moment I'm taking all the information I can with greps and awk, dumping with SFDISK, making the corrections and loading back. c'est la vie. Happily, my next task is a bit more interesting. making vmbuilder work with logical partitions... -- Good omen Ira Abramov http://ira.abramov.org/email/ ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Checkpoint Endpoint Security VPN with linux
On Mon, Mar 21, 2011 at 10:02 AM, Shachar Shemesh shac...@shemesh.bizwrote: On 21/03/11 02:41, Etzion Bar-Noy wrote: It is common that the VPN provider policy *prevents* you from connecting to multiple networks (theirs and someone else's). The logic behind it is to prevent data leak, especially accidental, by combining somehow their network with someone else's. You have to connect to some network in order to get the VPN packets out. Your home LAN, Internet Cafe, whatever. True. So - this poses no problem to be dealt with. The common problem is that your local home network overlaps one of the organization's networks. Some of the VPN clients place themselves in the network interface stack, so they hijack the packets to their correct destination(s). That is the common reason (except for time and effort) that Linux clients are more rare. This operation is somewhat more complicated there, and would require root access. Hijacking the outgoing packets does not solve the routing conflict. When I send a packet to 172.27.245.17, you somehow need to know whether that is the 172.27.245.17 that is visible through the VPN, or the one visible locally. Hijacking ALL outgoing packets rarely makes sense. They avoid hijacking your default GW. Hijacking the network interface does allow you to route the ENCRYPTED packet without going into routing loops, and is the reason this is done. Still, you are hiding parts of the network if there is a conflict. You do, of course. Usually, the VPN clients hide the local network where a conflict exists. Ez Shachar -- Shachar Shemesh Lingnu Open Source Consulting Ltd. http://www.lingnu.com ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Checkpoint Endpoint Security VPN with linux
It is common that the VPN provider policy *prevents* you from connecting to multiple networks (theirs and someone else's). The logic behind it is to prevent data leak, especially accidental, by combining somehow their network with someone else's. So - this poses no problem to be dealt with. The common problem is that your local home network overlaps one of the organization's networks. Some of the VPN clients place themselves in the network interface stack, so they hijack the packets to their correct destination(s). That is the common reason (except for time and effort) that Linux clients are more rare. This operation is somewhat more complicated there, and would require root access. Fortigate VPN client (SSL VPN) does that. Juniper java SSL VPN does that. CP SSL VPN client (snx) does that. Theoretically, for the client-connect (office-connect, for previous versions of Checkpoint) you would be able to use some implementation of *SWAN (freeswan, openswan, something swan), but this is a very complicated setup, which requires noticeable effort on your side, and, on some cases, requires the VPN server owner to perform actions which negate the famous Etzion's principle of minimum effort, minimum energy, so it would probably never happen. Ez 2011/3/20 Elazar Leibovich elaz...@gmail.com On Sun, Mar 20, 2011 at 9:54 PM, Shachar Shemesh shac...@shemesh.bizwrote: On another side note, what does it do if I'm having a 192.168.4.* internal network? Then you are @!#*%!@#$@!)(!@#%@#! !@(%!@#()#!@$!@%#. Wow, I'm not sure I know any adjective that long in English ;-) VPN is designed to connect disparaged networks as if they are close together. Yeah, but as your probably know, VPN is used in practice to connect to your workstation from your laptop, and for this use case, you might want to connect to two VPNs which unfortunately share the same internal network address. I don't think that makes you an idiot. And VPN solution could offer NAT, in fact a shallow Google search[1] offers exactly the same solution. Is there something I'm missing here? [1] http://nimlabs.org/~nim/dirtynat.html ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Advice Needed - Bye Bye Nokia!
And did you downgrade your MFE to 2.9 from 3.0, as 3.0 doesn't work well with Google Apps? Ez 2011/2/20 Hetz Ben Hamo het...@gmail.com Hi Amichai, I'm in the exact position as you with the same phone, yet I'm not planning to switch for several reasons: - Battery life: Try to find any smart phone which can hold more then a day without charging. Good luck with that. I just found this saturday night that I forgot to charge my Nokia since Tuesday, and I used it extensivly: emails, talking, surfing on the net - and the battery wasn't drained fast. - Email and Calendar syncing - I don't know about you, but I'm using Google Calendar with my E72, and it pushes the email and calendar perfectly. - Yes, Nokia sold their soul to go with Microsoft, so? It's not that they're ditching all the E series today, it will take at least a year until they'll have some Windows Phone 7 only phones catalog. Meanwhile they will continue giving support to your (and mine) phones. - My girl friend has Galaxy S, it's great with sync, surfing the net and everything, but when it comes to battery life, her phone will die, while my phone will barely drop a single line in the battery indicator. All in all, it's really how you look at this situation: as a business man/freelancer or as a private person. I, as a hosting provider, needs my phone to work in god-knows what conditions, and I need it to work continuously without reaching for a charger, and my Nokia E72 does that much better then Galaxy S. Hetz 2011/2/20 Amichai Rotman amic...@iglu.org.il Hello all, I need you advice, before I make a horrible mistake - so please bear with me and don't flame ;-) I recently opened a small business and wanted a smartphone with a dedicated line. I ended up buying a Nokia E72 (about 2 months ago). It's an OK phone, but it has a few problems. Then Nokia made their announcement (as I am sure almost all of you know) about the future OS, so I got mad and even though I always used Nokia phnes (since my first Nokia 6120D back in the mid '90s) I am willing to switch... A friend of mine recommends the Samsung Galaxy S - a very good choice for my needs: Runs Android 2.2 and works perfectly with Push for mail and calendar sync (the Nokia isn't so great in that department...) I now have to find a way to cancel my Nokia E72 purchase and switch to the Galaxy. I also heard about some kind of a Big Brother issue with Android devices? Any of you know how to get out of the Nokia purchase? The Galaxy wasn't available at the time. Am I making a mistake? Is this a smart move? Is it really a good phone? Thanks! Amichai. ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il -- *חץ בן חמו חץ-ביז (הוסטינג) *השכרה ואירוח של שרתים פיזיים השכרת שרתים וירטואליים מקצועיים וגדולים במחירים *קטנים* בקרו באתרנו בכתובת hetz.biz http://www.hetz.biz/ ובבלוג שלנו: blog.hetz.biz טלפוןן: 078113/4/5, אימייל: sa...@hetz.biz מסנג'ר: sa...@hetz.biz - סקייפ: heunique ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Advice Needed - Bye Bye Nokia!
You remove the existing one and manually install the older 2.9 version, saved previously for that exact purpose. Ez On Sun, Feb 20, 2011 at 11:34 PM, Amichai Rotman amic...@iglu.org.ilwrote: How do I check what version I have? And how do I down grade? I tried to download it and install it, but got an error the existing version cannot be removed Amichai. On Sun, Feb 20, 2011 at 14:23, Etzion Bar-Noy eza...@tournament.org.ilwrote: And did you downgrade your MFE to 2.9 from 3.0, as 3.0 doesn't work well with Google Apps? Ez 2011/2/20 Hetz Ben Hamo het...@gmail.com Hi Amichai, I'm in the exact position as you with the same phone, yet I'm not planning to switch for several reasons: - Battery life: Try to find any smart phone which can hold more then a day without charging. Good luck with that. I just found this saturday night that I forgot to charge my Nokia since Tuesday, and I used it extensivly: emails, talking, surfing on the net - and the battery wasn't drained fast. - Email and Calendar syncing - I don't know about you, but I'm using Google Calendar with my E72, and it pushes the email and calendar perfectly. - Yes, Nokia sold their soul to go with Microsoft, so? It's not that they're ditching all the E series today, it will take at least a year until they'll have some Windows Phone 7 only phones catalog. Meanwhile they will continue giving support to your (and mine) phones. - My girl friend has Galaxy S, it's great with sync, surfing the net and everything, but when it comes to battery life, her phone will die, while my phone will barely drop a single line in the battery indicator. All in all, it's really how you look at this situation: as a business man/freelancer or as a private person. I, as a hosting provider, needs my phone to work in god-knows what conditions, and I need it to work continuously without reaching for a charger, and my Nokia E72 does that much better then Galaxy S. Hetz 2011/2/20 Amichai Rotman amic...@iglu.org.il Hello all, I need you advice, before I make a horrible mistake - so please bear with me and don't flame ;-) I recently opened a small business and wanted a smartphone with a dedicated line. I ended up buying a Nokia E72 (about 2 months ago). It's an OK phone, but it has a few problems. Then Nokia made their announcement (as I am sure almost all of you know) about the future OS, so I got mad and even though I always used Nokia phnes (since my first Nokia 6120D back in the mid '90s) I am willing to switch... A friend of mine recommends the Samsung Galaxy S - a very good choice for my needs: Runs Android 2.2 and works perfectly with Push for mail and calendar sync (the Nokia isn't so great in that department...) I now have to find a way to cancel my Nokia E72 purchase and switch to the Galaxy. I also heard about some kind of a Big Brother issue with Android devices? Any of you know how to get out of the Nokia purchase? The Galaxy wasn't available at the time. Am I making a mistake? Is this a smart move? Is it really a good phone? Thanks! Amichai. ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il -- *חץ בן חמו חץ-ביז (הוסטינג) *השכרה ואירוח של שרתים פיזיים השכרת שרתים וירטואליים מקצועיים וגדולים במחירים *קטנים* בקרו באתרנו בכתובת hetz.biz http://www.hetz.biz/ ובבלוג שלנו: blog.hetz.biz טלפוןן: 078113/4/5, אימייל: sa...@hetz.biz מסנג'ר: sa...@hetz.biz - סקייפ: heunique ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: total uptime between shutdowns
Your solution will work, of course, only on a tidy shutdown. For unplanned shutdown (aka - power failure) it will not work. Ez On Tue, Feb 1, 2011 at 10:18 AM, Omer Zak w...@zak.co.il wrote: On Tue, 2011-02-01 at 10:00 +0200, Tom Rosenfeld wrote: Hi Guys, Is there a tool that will show me the total uptime (availability) of a machine between reboots? That is, if the machine was up for 24 hours and then shutdown for an hour and then up for 23 hours I want an answer of 47. I need this for charge-back of departments using cloud computer which they sometimes turn off. 1. Maybe the cloud computers' provider has the appropriate tool? In this case, if you don't find the tool yourself, you'll have to tell us which cloud provider are you working with. 2. The 'uptime' command gives the uptime since last reboot. You may want to add 'uptime /var/log/my_uptimes_log.txt' to the shutdown script, and rotate process, using a custom Perl script, the /var/log/my_uptime_log.txt file each month to total the uptime in that month. --- Omer -- MCSE - acronym for Minesweeper Consultant Solitaire Expert. (Unknown) My own blog is at http://www.zak.co.il/tddpirate/ My opinions, as expressed in this E-mail message, are mine alone. They do not represent the official policy of any organization with which I may be affiliated in any way. WARNING TO SPAMMERS: at http://www.zak.co.il/spamwarning.html ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Help me understand connection to internet: Infrastructure and ISP
I have been pursuing an Internet problem for the last two weeks until its apparent solution last week. It was classical - Packet Loss. Hot blamed Bezeq Int. and did nothing about it, I have had to prove (using Bezeqint support personnel) that the problem was Hot's. It was very tiresome, and I consider sending them a bill for my hours under the title assistance with solving Internet connectivity problem. I don't think they will pay, but it could be interesting to see the response it might provoke there. I use Linux, and I use a dialer. Whoever claims you have to use MPLS (aka - no-dialer) for Linux is an idiot who doesn't understand the infrastructure. I use L2TP, as can be described here (works for modern distros as well): http://run.tournament.org.il/cables-connection-in-israel-for-linux/ I have written it :-) It allows for easier troubleshooting when it comes to support personnel, given you understand the actions required. BezeqInt were OK, regarding the waste of my time, as I had explained what I was doing, and gave the (true) impression I know what I'm doing. Hot, on the other side, had a great service. They had spent hours with me. It's a great service, where it can be measured on the level of service, but they are technically inadepts, who should be kept far away from computer equipment. I refer their technical teams, of course - Their infrastructure teams, especially, and their tech support teams. They are like baboons with keyboards, and most of their actions are based on random it succeeded before so let's do it again methods. However, since Bezeq don't supply NGN near my home, I am stuck with Hot for the next year, most likely. When using MPLS, try to use traceroute, and check if you get packet loss to your default GW. This is within Hot's domain, still. Onwards, it's Bezeq Int. You can use their speed tests, and other international speed tests to test the connectivity. You can check DNS, with their DNS server, or with free DNS servers on the net, such as 8.8.8.8 (google. Will supply service). Hope it works for you better in the future. Ez On Sat, Jan 29, 2011 at 1:33 PM, Dotan Cohen dotanco...@gmail.com wrote: Several times I have been caught in the situation in which the Infrastructure (Hot in this particular case) and the ISP (012 in this particular case) blame each other for the customer not having internet access. I need to understand this. After two weeks of my mother-in-law not having internet access, during which 012 blamed Hot and Hot blamed 012, I get there to take care of the issue. At Hot they tell me that since the modem lights are fine (4 lit, one flashing) the problem must be with 012. At 012 they tell me that because she connects without a Dialer (as all Linux customers do) they cannot trace the problem any further than a superficial check which checks out OK. 012 sees the modem, so they say the problem must be with Hot. Apparently, the ability to trace problems is the reason that Windows users must use a Dialer. Furthermore, at 012 they inform me that it is Hot, not 012, that is providing the internet outside because there is no dialer. They say that they gave Hot permission to connect me directly to the internet this way. So if Hot can provide connections out, then why do we need ISPs at all? I notice that no other nation in Europe has this Infrastructure/ISP dichotomy. Furthermore, the next time that there is a problem how can I check if the issue is with the infrastructure or with the ISP? Thanks! -- Dotan Cohen http://gibberish.co.il http://what-is-what.com ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: What to tell 13 year old kids about Linux and Open Source?
These kids are used to these four freedoms, illegal, but works fir them. You may want to stress that their current (probable) actions are illegal, and that the existing model of selling software is one which changes into selling services. Talk about the ability to avoid vendor-lock, as part of it (give them an example of a bug in the software and what they can do with it - commercial software vs OSS software). These kinds of things, I think. Ez On Mon, Jan 10, 2011 at 9:34 AM, Uri Even-Chen u...@speedy.net wrote: On Mon, Jan 10, 2011 at 08:35, Gabor Szabo szab...@gmail.com wrote: hi, in the school of my son they are interested in getting some 1 hour long presentations from parents about various interesting subject. I could talk hours and days about Linux and Open Source in general but I wonder what do you think. What would be interesting to 13 year old kids? I am not sure free speech is interesting to them and unfortunately they are probably used to that every software is free beer. Given the cheats and cracks. So what do you think? I would recommend speaking about the free as in freedom concept of Linux and open source software - the 4 freedoms a software user should have, and that copying a software and giving a copy to your friend, modifying software etc. are basic freedoms anyone should have. Then speak about how Linux evolved, history of Linux and the status of Linux today. When I first started to learn programming, in 1983, I didn't know nothing about free software, open source software etc. Only after participating in open source conferences in 2001 and 2002 and reading the book free as in freedom (about Richard Stallman) I realized the free as in freedom concept and its advantages. Even today I'm still using non-free software such as Windows and Microsoft Office, but today I understand more about it than 10 years ago. Uri Even-Chen Mobile Phone: +972-50-9007559 E-mail: u...@speedy.net Website: http://www.speedy.net/ ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Electronic Junk in Haifa
As far as I can recall, the junkyard near downtown, around where the bridge above the train line (the bridge which leads to road 22) is still functioning. Yankale's junkyard or some similar name. Ez On Mon, Dec 20, 2010 at 7:04 AM, geoffrey mendelson geoffreymendel...@gmail.com wrote: On Dec 19, 2010, at 11:32 PM, Stan Goodman wrote: Years ago, there was a junk shop in Haifa, near the wholesale vegetable market and not far from the old Turkish railway station. I know that it isn't there anymore; is there such a place anywhere in the vicinity where disused and unneeded electronic odds and ends are bought and sold? The only place I know of is YS Metronics which is in the Modiin Industrial park, which is really in Lod. http://www.ysmetronics.com/English/ (about page has a map) Geoff. -- Geoffrey S. Mendelson, N3OWJ/4X1GM Those who cannot remember the past are condemned to misquote it. ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Linux friendly scanner
Did you try to scan a document at that time, from the PC, not from the printer device? Ez On Sat, Nov 27, 2010 at 3:52 PM, Hetz Ben Hamo het...@gmail.com wrote: Actually it does. I had a problem with my color cartridge, and it showed me a message to replace the cartridge, which I couldn't do at the time, and the message wouldn't go away and wouldn't let me send faxes or do anything. Hetz 2010/11/27 Etzion Bar-Noy eza...@tournament.org.il HP products don't behave like this. Ez On Sat, Nov 27, 2010 at 12:45 AM, Steve G. word...@gmail.com wrote: Not true at all. I have a brother scanner-printer-fax, and when the ink runs out, you can not send faxes. They manufacturers sell the printer for next to nothing, then get you by forcing you to change the ink every 3 months at ridiculous prices. Every color that dries up bricks the machine FOR EVERY FUNCTION. Instead of starting, it gives an error message and quits. Z. 2010/11/26 Etzion Bar-Noy eza...@tournament.org.il No relationship between the scanner module and the printer module in the device. They both work indifferently. Ez On Fri, Nov 26, 2010 at 9:14 AM, Shlomo Solomon shlomo.solo...@gmail.com wrote: On Thursday, November 25, 2010, Hetz Ben Hamo wrote: All HP (and most others) scanners are working fine with Linux and sane-backend. Thank you and 4 others who suggested specific models. As I wrote, I didn't look at multifunction printers because I don't print much and the ink dries out. But, since three of the replies did refer to multifuntion devices, I have another question. What happens when the ink dries or clogs? Can I still use these devices to scan? The reason I ask is that after looking at prices, I realize that a multi-function is CHEAPER than a stand alone scanner. But, it would be useless to me if I have to keep feeding it expensive ink just to be able to scan. -- Shlomo Solomon http://the-solomons.net Sent by KMail 1.13.3 (KDE 4.4.3) on LINUX Mandriva 2010.1 ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il -- Check out my web site - www.words2u.net ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il -- *חץ בן חמו חץ-ביז (הוסטינג) *השכרה ואירוח של שרתים פיזיים השכרת שרתים וירטואליים מקצועיים וגדולים במחירים *קטנים* בקרו באתרנו בכתובת hetz.biz http://www.hetz.biz/ ובבלוג שלנו: blog.hetz.biz טלפוןן: 078113/4/5, אימייל: sa...@hetz.biz מסנג'ר: sa...@hetz.biz - סקייפ: heunique ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: rdesktop question ( using Linux in a windows environment.)
You could actually do this, also, on the Linux machine, when mouse-over the Windows terminal client. This should behave quite similarly. Meaning Linux mouse mover or Linux keyboard presser or the likes. Generate some keyboard/mouse move on your desktop. Ez 2010/11/27 shimi linux...@shimi.net On Sat, Nov 27, 2010 at 6:09 PM, Dan Shimshoni danshi...@gmail.comwrote: Hi, linux-il, This question might seem a bit not related to Linux; However, it is a real problem which I face daily as a result of that I am trying to use Linux in a windows environment. So I hope you can advice. The problem description: I am using Linux station in a company which is windows-oriented; namely, most desktops are windows. I also have a windows desktop. Most of the day I sit in an office which is in on another floor from where my windows desktop is. I need to access the windows machine quite often for many tasks (for example, reading mail, as the sysadmin did not agree to install install/configure a mail client on the linux machine because this is the policy here). So I use rdesktop from linux to windows. The trouble with this is that every 10 minutes or so, if I do not perform any action in windows, the rdesktop connection is disabled and I have to open a new session of rdesktop. And this is quite annoying ! And I don't have an administrative permission to my windows machine, so I cannot change it. So my question is: is there some workaround ? for example, some windows application which will be awakened every 5 minutes to let the windows machine think that there was some action on windows, and thus the rdesktop session will not finish ? Maybe something like this: http://rjlpranks.com/pranks/mouse_move/ will do the trick. Note: This is not to recommend or vouch for the specific program. I have no idea if it's Malware... There are other (mostly commercial, i.e. $$$) software packages that you can find on Google when you search for Windows Mouse Mover HTH, -- Shimi ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Linux friendly scanner
No relationship between the scanner module and the printer module in the device. They both work indifferently. Ez On Fri, Nov 26, 2010 at 9:14 AM, Shlomo Solomon shlomo.solo...@gmail.comwrote: On Thursday, November 25, 2010, Hetz Ben Hamo wrote: All HP (and most others) scanners are working fine with Linux and sane-backend. Thank you and 4 others who suggested specific models. As I wrote, I didn't look at multifunction printers because I don't print much and the ink dries out. But, since three of the replies did refer to multifuntion devices, I have another question. What happens when the ink dries or clogs? Can I still use these devices to scan? The reason I ask is that after looking at prices, I realize that a multi-function is CHEAPER than a stand alone scanner. But, it would be useless to me if I have to keep feeding it expensive ink just to be able to scan. -- Shlomo Solomon http://the-solomons.net Sent by KMail 1.13.3 (KDE 4.4.3) on LINUX Mandriva 2010.1 ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Linux friendly scanner
HP products don't behave like this. Ez On Sat, Nov 27, 2010 at 12:45 AM, Steve G. word...@gmail.com wrote: Not true at all. I have a brother scanner-printer-fax, and when the ink runs out, you can not send faxes. They manufacturers sell the printer for next to nothing, then get you by forcing you to change the ink every 3 months at ridiculous prices. Every color that dries up bricks the machine FOR EVERY FUNCTION. Instead of starting, it gives an error message and quits. Z. 2010/11/26 Etzion Bar-Noy eza...@tournament.org.il No relationship between the scanner module and the printer module in the device. They both work indifferently. Ez On Fri, Nov 26, 2010 at 9:14 AM, Shlomo Solomon shlomo.solo...@gmail.com wrote: On Thursday, November 25, 2010, Hetz Ben Hamo wrote: All HP (and most others) scanners are working fine with Linux and sane-backend. Thank you and 4 others who suggested specific models. As I wrote, I didn't look at multifunction printers because I don't print much and the ink dries out. But, since three of the replies did refer to multifuntion devices, I have another question. What happens when the ink dries or clogs? Can I still use these devices to scan? The reason I ask is that after looking at prices, I realize that a multi-function is CHEAPER than a stand alone scanner. But, it would be useless to me if I have to keep feeding it expensive ink just to be able to scan. -- Shlomo Solomon http://the-solomons.net Sent by KMail 1.13.3 (KDE 4.4.3) on LINUX Mandriva 2010.1 ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il -- Check out my web site - www.words2u.net ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Linux friendly scanner
Same goes for HP Photosmart C5283 All-in-One Ez On Thu, Nov 25, 2010 at 10:03 PM, Elazar Leibovich elaz...@gmail.comwrote: I'm having a multifunction printer Canon Pixma MP140, and was very surprised to find out that it works out of the box in Ubuntu 10.04 with the simple scan utility. On Thu, Nov 25, 2010 at 7:42 PM, Shlomo Solomon shlomo.solo...@gmail.com wrote: My Mustek 1200 UB died and I'm looking for a linux friendly scanner. The old one worked out-of-the-box in various versions of Mandriva and I hope to find something just as good. My criteria: - cheap - works with Linux (obviously) - good quality scans (good is fine - doesn't have to be proffesional level) BTW - I though of a multifuction printer, but I don't print much and ink-jets usually clog up if not used often enough and laser multi-functions are more than my budget. Any suggestions? -- Shlomo Solomon http://the-solomons.net Sent by KMail 1.13.3 (KDE 4.4.3) on LINUX Mandriva 2010.1 ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: wifi adapter problem
It did not accept your ESSID command. Look - your essid is empty. Try running that command again. Ez 2010/11/17 David Ronkin dron...@gmail.com Hi I bought Edimax EW-7711USN usb adapter (with linux support written on the box). Looks like the driver is found by my ubuntu 10.10 Maverick (on titan laptop) but still i cannot go to web (i have open wireless nw around) - getting the error below. I googled for some time but gave up meanwhile, Please kindly help! Here some logs: # lshw | grep wire *-firewire product: Firewire (IEEE 1394) configuration: driver=firewire_ohci latency=32 capabilities: ethernet physical wireless configuration: broadcast=yes driver=rt73usb driverversion=2.6.35-22-generic firmware=N/A link=no multicast=yes wireless=IEEE 802.11bg capabilities: ethernet physical wireless configuration: broadcast=yes multicast=yes wireless=Ralink STA #iwconfig wlan1 essid Homek #iwconfig wlan1 Ralink STA ESSID: Nickname:RT2870STA Mode:Auto Frequency=2.412 GHz Access Point: Not-Associated Bit Rate:1 Mb/s RTS thr:off Fragment thr:off Encryption key:off Link Quality=10/100 Signal level:0 dBm Noise level:-115 dBm Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0 Tx excessive retries:0 Invalid misc:0 Missed beacon:0 # dhclient wlan1 Internet Systems Consortium DHCP Client V3.1.3 Copyright 2004-2009 Internet Systems Consortium. All rights reserved. For info, please visit https://www.isc.org/software/dhcp/ Listening on LPF/wlan1/00:1f:1f:c7:16:53 Sending on LPF/wlan1/00:1f:1f:c7:16:53 Sending on Socket/fallback DHCPDISCOVER on wlan1 to 255.255.255.255 port 67 interval 3 DHCPDISCOVER on wlan1 to 255.255.255.255 port 67 interval 7 DHCPDISCOVER on wlan1 to 255.255.255.255 port 67 interval 14 DHCPDISCOVER on wlan1 to 255.255.255.255 port 67 interval 15 DHCPDISCOVER on wlan1 to 255.255.255.255 port 67 interval 9 DHCPDISCOVER on wlan1 to 255.255.255.255 port 67 interval 13 No DHCPOFFERS received. *No working leases in persistent database - sleeping.* Tnx! David -- בברכה, דוד רונקין ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: looking to buy ARM servers
Get yourself an old openmoko device. They are ARM based, well documented, with several simple options of Linux distros for them. Very weak, of course, but for ARM games (playing with your arm :-) ) they should be just fine. Ez On Tue, Nov 2, 2010 at 5:31 PM, Muli Ben-Yehuda m...@il.ibm.com wrote: On Tue, Nov 02, 2010 at 05:03:39PM +0200, Hetz Ben Hamo wrote: Hi, Why don't you take Atom processors? Because I want to run experiments on the ARM architecture. Cheers, Muli ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Pop toaster recommendations seeked
I have been managing a system based on ispman (well, created and managed), and it was a wonderful tool, full of features, and worked quite well. It was very complex, and requires substantial understanding of directories (LDAP). It was too complex (afterthought), but worked well. I would not recommend it for managing only several virtual mail domains. For larger tasks (hosting management) it could do just fine. Ez On Mon, Nov 1, 2010 at 6:55 AM, Yedidyah Bar-David linux...@didi.bardavid.org wrote: On Sun, Oct 31, 2010 at 10:23:34PM +0200, Ira Abramov wrote: Howdie folks. for years I would install qmail+vpopmail+qmailadmin to give a client a nice robust virtual-domains web manager with mailing lists and all they needed, but today I discovered, that after 2-3 years of code freeze and maybe drop of users, courier imap seems has broken the option of authenticating with vpopmail, which means I have a wonderful virtual mail management and delivery system, but no pop3 and imapd to use it with. maybe it's time to ditch this kit and find another. option 1: courier MTA and SQWebmail - not so happy about it. I don't know the MTA, the message store is proprietary and limiting, no central management via web of the users (other than PHPMyAdmin, not friendly enough for my client. options 2: Zimbra. seams like a bit of an overkill but I'm told it works well, has postfix for an engine, and rumors say it supports virtual domains well (though I could not be sure from the confusing admin manual). Option 3: ? Donno... can you recommend? I worked for several years with postfix+dovecot+postfixadmin. IIRC it was mostly based on this howto: http://bliki.rimuhosting.com/space/knowledgebase/linux/mail/postfixadmin+on+debian+sarge which is pretty dated, but postfixadmin itself (and the underlying tools, no doubt) is still maintained. It's pretty basic but working. You might consider trying one of virtual hosting packages out there. I then did and decided they were either too big or not mature enough (or both) and went with postfixadmin which did a rather small part of their common denominator but seemed mature. A partial list of the ones I then looked at: ispconfig web-cp dtc ispman vhcs gnuhh ravencore. Googling for most subsets of them will probably find others/reviews/etc. -- Didi ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Amazon free to new customers!
With defined limitations. You can use *their* kernel and initrd, or some others supplied, but you cannot supply your own. Also - adding custom OS selections is not trivial. If not already created by someone else (with his/hers S3 space to store that image), you might find yourself lacking the right image for your needs. That said, almost anything can be done using the existing, limited, images supplied by Amazon. Add to that the images supplied by users, and you will be just fine. Ez 2010/10/24 Tom Rosenfeld trosenf...@gmail.com On Sun, Oct 24, 2010 at 10:10 AM, geoffrey mendelson geoffreymendel...@gmail.com wrote: On Oct 24, 2010, at 9:51 AM, Tom Rosenfeld wrote: If you haven't already tried cloud computing, this is your chance to use free for a year! -tom Does this have to be a web server? Can one use it for an Asterisk or other SIP relay or a private HTTP or SOCKS proxy? Geoff -- Geoffrey S. Mendelson, N3OWJ/4X1GM To help restaurants, as part of the stimulus package, everyone must order dessert. As part of the socialized health plan, you are forbidden to eat it. :-) No, it does not have to be a web server. Actually, the point of Amazon is that you can install any OS and any software you like! -tom ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Amazon EC2 hosting,
A small note. I was led to understand (from a fried who uses EC2 and aggressively) that xlarge instances are (usually? Always? I think the later) alone on physical hardware. So you would prefer to use xlarge instance to prevent slowdowns. Ez 2010/10/10 Maxim Veksler ma...@vekslers.org On Sun, Oct 10, 2010 at 5:04 PM, Orna Agmon Ben-Yehuda ladyp...@gmail.com wrote: 2010/10/10 Ori Idan o...@helicontech.co.il 2010/10/10 Tom Rosenfeld trosenf...@gmail.com Hi, I just came across this thread from back in Aug about Amazon's cloud. I'd like to add that I have been a satisfied customer of Amazon for over a year, using their services for both consulting at at my current job where we use it to run our SaaS offering. The capabilities keep improving and the prices keep coming down. Their lowest end server is now just 2 cents an hour! There are some issues with the IO, but it is certainly adequate for all but high performance needs. We use 8 way stripped disks and get about 100 MBp/s sequential reads. If anyone wants more details, I'll be happy to share with you. -tom I am considering using EC2 for a web application. I am not sure how to calculate the payment per month. Do I pay only for the time someone makes a request? For example, I have a user who requests a certain report and it takes 1 second to load the report request form, then 20 seconds to produce the report and print it. I understand that I pay for 21 seconds? In addition to mistakes already corrected, there is another mistake of how long something takes. Amazon aim to provide a certain computation power unit, but benchmarks show that what is actually provided has high variability. For example, ping times to EC2 machines started rising significantly since Amazon announces the spot instances. See also: Some more input on EC2. Not all instances born alike. We recently ran a huge computation based on Hadoop and you can definitely see that some nodes perform faster (I/O was the bottleneck) then others. I too, when starting with EC2 made the mistake to of thinking that you only pay for as much CPU as you use. Wrong! OTOH, I was very happy to find out that with Google AppEngine this is actually the case: You pay for as much resources as you consume. And they DO count CPU Time vs. Amazon's instance is running time. Another note regarding EC2. Read bitbucket story about ec2 horrors http://blog.bitbucket.org/. Yet please don't get me wrong, generally EC2, S3, CloudFront, ELB and other Amazon's services work great - Our production farm (~40 servers is hosted there and we are relatively happy). Amazon's main issues are: I/O bandwidth is funny Occasionally peaks in connectivity time that lead to timeouts (between zones from the outside world). Not so fair hypervisor: We've seen occasions when an instance slows down for a couple of minutes. We assume (without being able to tell for sure) that some bigger instance type that happen to be hosted on the same physical server as we are got resource hungry and practically ate all our CPU time... Maxim. ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Awful Bandwidth from Most Sites on Bezeqint - What can I do about it?
Hi.
I am not quite happy myself, lately, either.
Browsing and downloading seems slow, as well.
wget
http://ftp.nluug.nl/pub/os/Linux/distr/Mandrivalinux/devel/cooker/SRPMS/main/release/gcc-4.5.1-1mnb2.src.rpm
--00:08:12--
http://ftp.nluug.nl/pub/os/Linux/distr/Mandrivalinux/devel/cooker/SRPMS/main/release/gcc-4.5.1-1mnb2.src.rpm
= `gcc-4.5.1-1mnb2.src.rpm'
Resolving ftp.nluug.nl... 192.87.102.42, 192.87.102.43
Connecting to ftp.nluug.nl|192.87.102.42|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 68,730,137 (66M) [application/x-rpm]
100%[] 68,730,137 113.99K/sETA
00:00
00:16:41 (132.27 KB/s) - `gcc-4.5.1-1mnb2.src.rpm' saved [68730137/68730137]
On a 12Mb/s connection. NTOP reported usage of 1.3-1.4Mb/s and that was the
only major use of the line at this time.
mirror.isoc.org.il resides on Israeli IP, as well as off.co.il (Netvision, I
think), so their speeds are excellent.
This is about midnight now. I can (and will) try later too.
Ez
On Wed, Oct 13, 2010 at 10:47 PM, Shlomi Fish shlo...@iglu.org.il wrote:
Hi all,
lately I've been getting awful bandwidth from most sites on my home
Bezeqint
connection from all hosts on my network - I've tried a Mandriva Cooker
desktop
computer connected via 100 Mbps Ethernet, a WinXP SP 3 machine also
connected
via Ethernet, and my Mandriva 2010.1 laptop with a Wifi connection, and
they
all yield the same bad bandwidth (about 40-70 KBytes/s where my maximum is
300
KBytes/s).
Note that I'm getting good connectivity from some sites in Israel - usually
from mirror.isoc.org.il and also from
http://off.co.il/apache/xmlgraphics/fop/
(an Israeli Apache mirror). But otherwise my downloading connection is bad.
Here's a log I've created:
{
* http://off.co.il/apache/xmlgraphics/fop/source/fop-1.0-src.zip
- gives speed of 300KB/s +.
*
http://ftp.nluug.nl/pub/os/Linux/distr/Mandrivalinux/devel/cooker/SRPMS/main/release/gcc-4.5.1-1mnb2.src.rpm
- gives:
shlomif:~$ wget -c
http://ftp.nluug.nl/pub/os/Linux/distr/Mandrivalinux/devel/cooker/SRPMS/main/release/gcc-4.5.1-1mnb2.src.rpm
--2010-10-12 22:16:37--
http://ftp.nluug.nl/pub/os/Linux/distr/Mandrivalinux/devel/cooker/SRPMS/main/release/gcc-4.5.1-1mnb2.src.rpm
Resolving ftp.nluug.nl (ftp.nluug.nl)... 192.87.102.43, 192.87.102.42,
2001:610:1:80aa:192:87:102:42, ...
Connecting to ftp.nluug.nl (ftp.nluug.nl)|192.87.102.43|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 68730137 (66M) [application/x-rpm]
Saving to: `gcc-4.5.1-1mnb2.src.rpm'
1% [ ] 1,353,806 28.8K/s eta 14m
39s
^C
*
http://www.kernel.org/pub/linux/kernel/v2.6/testing/linux-2.6.36-rc7.tar.bz2
Resolving www.kernel.org (www.kernel.org)... 199.6.1.164, 130.239.17.4
Connecting to www.kernel.org (www.kernel.org)|199.6.1.164|:80...
connected.
HTTP request sent, awaiting response... 200 OK
Length: 70217056 (67M) [application/x-bzip2]
Saving to: `linux-2.6.36-rc7.tar.bz2'
2% [ ] 1,653,032 56.3K/s eta 16m
29s
^C
}
I'm getting good connectivity from these sites from some remote hosts that
I
can ssh into, and furthermore the download manager prozilla ( see:
http://sourceforge.net/projects/prozilla/ ) can open several streams like
that
and download individual files more quickly. I tried turning off the WinXP
machine, but the bad connectivity remained the same.
I tried calling the Bezeqint tech support, but the phone support guy was
clueless as usual and didn't help me solve this problem. I have a D-Link
router, which is pretty new (and might be the source of the problem).
Can anyone help me shed any light on this problem?
Regards,
Shlomi Fish
--
-
Shlomi Fish http://www.shlomifish.org/
Parody of The Fountainhead - http://shlom.in/towtf
rindolf She's a hot chick. But she smokes.
go|dfish She can smoke as long as she's smokin'.
Please reply to list if it's a mailing list post - http://shlom.in/reply .
___
Linux-il mailing list
Linux-il@cs.huji.ac.il
http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
___
Linux-il mailing list
Linux-il@cs.huji.ac.il
http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Awful Bandwidth from Most Sites on Bezeqint - What can I do about it?
P.S - I am on private NGN...
Ez
On Thu, Oct 14, 2010 at 12:23 AM, Etzion Bar-Noy
eza...@tournament.org.ilwrote:
Hi.
I am not quite happy myself, lately, either.
Browsing and downloading seems slow, as well.
wget
http://ftp.nluug.nl/pub/os/Linux/distr/Mandrivalinux/devel/cooker/SRPMS/main/release/gcc-4.5.1-1mnb2.src.rpm
--00:08:12--
http://ftp.nluug.nl/pub/os/Linux/distr/Mandrivalinux/devel/cooker/SRPMS/main/release/gcc-4.5.1-1mnb2.src.rpm
= `gcc-4.5.1-1mnb2.src.rpm'
Resolving ftp.nluug.nl... 192.87.102.42, 192.87.102.43
Connecting to ftp.nluug.nl|192.87.102.42|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 68,730,137 (66M) [application/x-rpm]
100%[] 68,730,137 113.99K/sETA
00:00
00:16:41 (132.27 KB/s) - `gcc-4.5.1-1mnb2.src.rpm' saved
[68730137/68730137]
On a 12Mb/s connection. NTOP reported usage of 1.3-1.4Mb/s and that was the
only major use of the line at this time.
mirror.isoc.org.il resides on Israeli IP, as well as off.co.il (Netvision,
I think), so their speeds are excellent.
This is about midnight now. I can (and will) try later too.
Ez
On Wed, Oct 13, 2010 at 10:47 PM, Shlomi Fish shlo...@iglu.org.il wrote:
Hi all,
lately I've been getting awful bandwidth from most sites on my home
Bezeqint
connection from all hosts on my network - I've tried a Mandriva Cooker
desktop
computer connected via 100 Mbps Ethernet, a WinXP SP 3 machine also
connected
via Ethernet, and my Mandriva 2010.1 laptop with a Wifi connection, and
they
all yield the same bad bandwidth (about 40-70 KBytes/s where my maximum is
300
KBytes/s).
Note that I'm getting good connectivity from some sites in Israel -
usually
from mirror.isoc.org.il and also from
http://off.co.il/apache/xmlgraphics/fop/
(an Israeli Apache mirror). But otherwise my downloading connection is
bad.
Here's a log I've created:
{
* http://off.co.il/apache/xmlgraphics/fop/source/fop-1.0-src.zip
- gives speed of 300KB/s +.
*
http://ftp.nluug.nl/pub/os/Linux/distr/Mandrivalinux/devel/cooker/SRPMS/main/release/gcc-4.5.1-1mnb2.src.rpm
- gives:
shlomif:~$ wget -c
http://ftp.nluug.nl/pub/os/Linux/distr/Mandrivalinux/devel/cooker/SRPMS/main/release/gcc-4.5.1-1mnb2.src.rpm
--2010-10-12http://ftp.nluug.nl/pub/os/Linux/distr/Mandrivalinux/devel/cooker/SRPMS/main/release/gcc-4.5.1-1mnb2.src.rpm--2010-10-1222:16:37--
http://ftp.nluug.nl/pub/os/Linux/distr/Mandrivalinux/devel/cooker/SRPMS/main/release/gcc-4.5.1-1mnb2.src.rpm
Resolving ftp.nluug.nl (ftp.nluug.nl)... 192.87.102.43, 192.87.102.42,
2001:610:1:80aa:192:87:102:42, ...
Connecting to ftp.nluug.nl (ftp.nluug.nl)|192.87.102.43|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 68730137 (66M) [application/x-rpm]
Saving to: `gcc-4.5.1-1mnb2.src.rpm'
1% [ ] 1,353,806 28.8K/s eta 14m
39s
^C
*
http://www.kernel.org/pub/linux/kernel/v2.6/testing/linux-2.6.36-rc7.tar.bz2
Resolving www.kernel.org (www.kernel.org)... 199.6.1.164, 130.239.17.4
Connecting to www.kernel.org (www.kernel.org)|199.6.1.164|:80...
connected.
HTTP request sent, awaiting response... 200 OK
Length: 70217056 (67M) [application/x-bzip2]
Saving to: `linux-2.6.36-rc7.tar.bz2'
2% [ ] 1,653,032 56.3K/s eta 16m
29s
^C
}
I'm getting good connectivity from these sites from some remote hosts that
I
can ssh into, and furthermore the download manager prozilla ( see:
http://sourceforge.net/projects/prozilla/ ) can open several streams like
that
and download individual files more quickly. I tried turning off the WinXP
machine, but the bad connectivity remained the same.
I tried calling the Bezeqint tech support, but the phone support guy was
clueless as usual and didn't help me solve this problem. I have a D-Link
router, which is pretty new (and might be the source of the problem).
Can anyone help me shed any light on this problem?
Regards,
Shlomi Fish
--
-
Shlomi Fish http://www.shlomifish.org/
Parody of The Fountainhead - http://shlom.in/towtf
rindolf She's a hot chick. But she smokes.
go|dfish She can smoke as long as she's smokin'.
Please reply to list if it's a mailing list post - http://shlom.in/reply.
___
Linux-il mailing list
Linux-il@cs.huji.ac.il
http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
___
Linux-il mailing list
Linux-il@cs.huji.ac.il
http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: How can I grab the filesystem from corrupted lvm (help recover data from lvm with one pv missing)
Bad. Depends on the amount of damage you have created, the following procedure would work: I believe the failure is in /dev/sda8, around 30GB, based on your fileserver lvm backup file. Upload an older one for me to be sure. pvcreate -u GBsXFQ-RdXS-iMhp-Phle-iqfM-5571-aJgQAa /dev/sda8 (if successful), try the following: in /etc/lvm/backup, find a file describing a good configuration, before you attempted to force-remove the PV. You can perform the following action then: vgcfgrestore -f the file you have found, and hopefully uploaded here fileserver If successful, run: vgchange -ay fileserver and you should be able to mount whatever LV you did not run over with zeros. Good luck Ez On Mon, Oct 11, 2010 at 11:20 PM, Oron Peled o...@actcom.co.il wrote: On Monday, 11 בOctober 2010 11:50:45 Boris shtrasman wrote: On Mon, Oct 11, 2010 at 11:45 AM, Yedidyah Bar-David Not gparted, gpart: http://www.brzitwa.de/mb/gpart/index.html Gpart is a tool which tries to guess the primary partition table of a PC-type hard disk in case the primary partition table in sector 0 is damaged, incorrect or deleted. If you fail, and still want to resurrect specific files, you can also try MagicRescue: http://www.itu.dk/people/jobr/magicrescue/ looks promising thank you And i was dding file by file :-( from the disk .. Just take notice that sequencial logical volume (partition) may be not sequencial on the physical volume (There's a logical extent to physical extent mapping). However, there is a good chance most/all of your partition is sequencial, especially if you created the volume group and the logical volumes when the disk was empty (e.g: during installation) without requiring a striped logical volume (it's not the default). Good luck, -- Oron Peled Voice: +972-4-8228492 o...@actcom.co.il http://users.actcom.co.il/~oron ... one of the main causes of the fall of the Roman Empire was that, lacking zero, they had no way to indicate successful termination of their C programs. -- Robert Firth ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: How can I grab the filesystem from corrupted lvm (help recover data from lvm with one pv missing)
Did you use pvcreate -u on the drive? Without it, it won't work. Also - you *must* use a backup configuration file with all the data in it. Post one of these online, and I would be able to assist further. Ez 2010/10/12 Boris shtrasman borissh1...@gmail.com /dev/sda8 had been wiped out totally The other restore fails : vgcfgrestore -f fileserver fileserver File descriptor 13 (socket:[7450]) leaked on vgcfgrestore invocation. Parent PID 3952: bash Couldn't find device with uuid fVkJmY-pdwD-bub8-URE8-lNc2-eyhr-IIxIez. Cannot restore Volume Group fileserver with 1 PVs marked as missing. Restore failed. Today ill try to create again an image of the drive (dd all the drive ) in order not make any more damage by restore attempts, Yesterday I found out that the harddrive going to die (errors from the kernel). 2010/10/12 Etzion Bar-Noy eza...@tournament.org.il Bad. Depends on the amount of damage you have created, the following procedure would work: I believe the failure is in /dev/sda8, around 30GB, based on your fileserver lvm backup file. Upload an older one for me to be sure. pvcreate -u GBsXFQ-RdXS-iMhp-Phle-iqfM-5571-aJgQAa /dev/sda8 (if successful), try the following: in /etc/lvm/backup, find a file describing a good configuration, before you attempted to force-remove the PV. You can perform the following action then: vgcfgrestore -f the file you have found, and hopefully uploaded here fileserver If successful, run: vgchange -ay fileserver and you should be able to mount whatever LV you did not run over with zeros. Good luck Ez On Mon, Oct 11, 2010 at 11:20 PM, Oron Peled o...@actcom.co.il wrote: On Monday, 11 בOctober 2010 11:50:45 Boris shtrasman wrote: On Mon, Oct 11, 2010 at 11:45 AM, Yedidyah Bar-David Not gparted, gpart: http://www.brzitwa.de/mb/gpart/index.html Gpart is a tool which tries to guess the primary partition table of a PC-type hard disk in case the primary partition table in sector 0 is damaged, incorrect or deleted. If you fail, and still want to resurrect specific files, you can also try MagicRescue: http://www.itu.dk/people/jobr/magicrescue/ looks promising thank you And i was dding file by file :-( from the disk .. Just take notice that sequencial logical volume (partition) may be not sequencial on the physical volume (There's a logical extent to physical extent mapping). However, there is a good chance most/all of your partition is sequencial, especially if you created the volume group and the logical volumes when the disk was empty (e.g: during installation) without requiring a striped logical volume (it's not the default). Good luck, -- Oron Peled Voice: +972-4-8228492 o...@actcom.co.il http://users.actcom.co.il/~oronhttp://users.actcom.co.il/%7Eoron ... one of the main causes of the fall of the Roman Empire was that, lacking zero, they had no way to indicate successful termination of their C programs. -- Robert Firth ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il -- -- -- Boris Shtrasman |Gnu/Linux Software developer | | IM : bori...@jabber.org | | URL : myrtfm.blogspot.com| ___ ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: seeking cross-platform professional backup recommendation
I don't know about Bacula and Windows, but the requirement was that the server runs on Windows, right? ArcServ is a lousy product. Easier to backup. Harder to restore. Symantec product is OK. You will not have the ability to restore from bare-metal without some major work, however, you can create (if you have enough time) a unified bootable USB disk which can run Linux OS with backup agent to achieve bare-metal. BE is a better and more reliable. I would strongly recommend on it, from the portfolio of non-OSS solutions. Ez On Mon, Oct 11, 2010 at 5:39 PM, Avi Rozen avi.ro...@mobileye.com wrote: Ira Abramov wrote: Bacula and AMANDA are probably not options since they won't use shadow copies thus can't back up databases swiftly and restore them partially and intelligently like propriatary tools do. I'm in no position to compare it with other backup solutions, but I do know that Bacula can use the volume shadow copy service on windows [1], since version 1.37.x (current version is 5.0.x). Cheers, Avi [1] http://www.bacula.org/5.0.x-manuals/en/main/main/Windows_Version_Bacula.html#SECTION00376 This mail was sent via Mail-SeCure system. ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: CPU RAM in a storage box
Linux LVM2 has been around for several years now. It can take and use snapshots, and I do it for the last three or so years on *production* sites. There are limitations, such as space utilization and performance, but the most significant one is that LVM snapshots are nowhere near NetApp snapshots. Check BTRFS or ZFS for such behavior. Ez 2010/9/10 Erez D erez0...@gmail.com here is my 3 cents: In two companies i worked for, i designed similiar servers. 1. i wanted to have as less as a down time, and i didn't want to buy another server just to sit and wait for a failure, so i decided it should work on any pc with any raid controller - i decided to do the raid in software, so it will be hardware independant - in case of hardware failure - just get any pc with any raid controller ( the disks may even be connected via usb if you do not have a raid controller) and everything is up in minutes. 2. I wanted to have snapshots like in a net-app - making snapshots every hour or so. using the most up to date redhat distributions of that time - when creating a snapshot, once in a while the server got locked (dead). i tried it again when LVM2 was officially released, and got similiar results. so as they say once bitten, twice shy - I will take a lot of proof LVM is stable enough for snapshots, before i try it again. 3. linux caches the disk data before it writes it to disk. this is very good for performance, but may create a big data loss in case of power failure... cheers, erez. 2010/9/9 Hetz Ben Hamo het...@gmail.com Hi people, I'm planning to add some big storage solution to my VPS businesshttp://hetz.biz. I did some checking and calculated the costs, and figured out that if I want to have a decent 12TB solution NAS box, it would be best if I would roll my own. (12 TB before all the RAID stuff, after that it would lot less). All other solutions are very expensive (example: IBM EXP 3000 costs here 6K nis without a single hard disk). I'm planning to use hardware based RAID card, minimal Linux distribution and have some offers like iSCSI, NFS, CIFS - the usual suspects. My question is: since I'll use hardware RAID card, which processor and how much RAM should I put in such a machine? Xeon is overkill IIRC. What do you suggest? Thanks and Shana Tova Hetz -- my blog (hebrew): http://benhamo.org Skype: heunique MSN: hetz-b...@benhamo.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: CPU RAM in a storage box
Nope. I did not say production. I said that if you want NetApp-style snapshots, use one of the mentioned file systems. I never said I actually used them for production, nor for anything at all (I don't) Ez On Fri, Sep 10, 2010 at 2:52 PM, Amos Shapira amos.shap...@gmail.comwrote: 2010/9/10 Etzion Bar-Noy eza...@tournament.org.il: Linux LVM2 has been around for several years now. It can take and use snapshots, and I do it for the last three or so years on production sites. There are limitations, such as space utilization and performance, but the most significant one is that LVM snapshots are nowhere near NetApp snapshots. Check BTRFS or ZFS for such behavior. Ez I use LVM2 for a few years too but BTRFS or ZFS on Linux for production?? --Amos ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: CPU RAM in a storage box
This is a joke, right? You want someone to host your system, which, by design, will not be rack-mountable, and would be large, due to the amount of disks you are to place there. It is possible, but extremely expensive to host a non-1-U server nowadays. Who would give it to you? An industrial-grade, 2U system could host, today, about 6 3.5 SATA disks. A 3 U can do much more, with up to 12-14 disks, depending on the system. And RAM is extremely important. Since you will not invest in an industial-class RAID controller (3ware, LSI-Logic, Adaptec, Intel, etc) which will cost several hundreds of dollars, as I see it, you would want to compensate for the high write latency with a large amount of RAM and fully buffered writes (not secure, but good enough). Especially with 7200RPM SATA drives with low seek speed. NFS shares, in async mode would give great performance, provided you give the system enough RAM. Then your RAM will actually become the disk write cache. Ez 2010/9/9 Hetz Ben Hamo het...@gmail.com Hi, 2010/9/9 geoffrey mendelson geoffreymendel...@gmail.com On Sep 9, 2010, at 6:35 PM, Hetz Ben Hamo wrote: I'm planning to add some big storage solution to my VPS business. I did some checking and calculated the costs, and figured out that if I want to have a decent 12TB solution NAS box, it would be best if I would roll my own. (12 TB before all the RAID stuff, after that it would lot less). All other solutions are very expensive (example: IBM EXP 3000 costs here 6K nis without a single hard disk). The question you should be asking yourself, IMHO, is what can I buy that will be as reliable as a commerical, industrial grade server? Not looking for industrial grade one. I'm planning to use hardware based RAID card, minimal Linux distribution and have some offers like iSCSI, NFS, CIFS - the usual suspects. My question is: since I'll use hardware RAID card, which processor and how much RAM should I put in such a machine? Xeon is overkill IIRC. For example, a system which costs under 900 NIS would do the job. You can get them from Ivory or KSP. They have a dual core ATOM processor, one PCI slot and one DDR2 memory slot. The power supply is not very big, but it will power a bunch of 5400 rpm green disks. This storage will be mainly used for backups. If someone wants to do a colocation to my rack, I want to give him a bonus, something that you can't find today with my competitors: I want to give him 50-100GB for storage. You'll get an NFS/CIFS/iSCSI and you mount it to your machine and use it for your backup/rsync/whatever. By comparison, when you colocate a server to Netvision's farm, you get ... 5GB backup space.. yippee.. How well will it work? How long will it last? Will it be fast enough? Fast doesn't matter much when you're doing backups or storing some temporary stuff, does it really matter when it take 20 seconds instead of 10 when you're doing rsync? I don't think so.. And the killer question, how much will it cost to replace, in the value of downtime, your time to replace it, bad will among your customers, etc? Really depends. I'm not planning to fully use all the disks, some will be disconnected or out of the RAID, perhaps I'll put a redundant PSU. Hetz Geoff. -- Geoffrey S. Mendelson, N3OWJ/4X1GM To help restaurants, as part of the stimulus package, everyone must order dessert. As part of the socialized health plan, you are forbidden to eat it. :-) -- my blog (hebrew): http://benhamo.org Skype: heunique MSN: hetz-b...@benhamo.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: CPU RAM in a storage box
On top of rpath Linux, which is the root of all evil. Also - although I have implemented quite a few of these, OpenFiler suffers from various bugs and shortcomings. Still - in the free like beer area - it is good enough for most purposes. Ez On Fri, Sep 10, 2010 at 1:08 AM, Amos Shapira amos.shap...@gmail.comwrote: 2010/9/10 Hetz Ben Hamo het...@gmail.com Hi people, I'm planning to add some big storage solution to my VPS business. I did some checking and calculated the costs, and figured out that if I want to have a decent 12TB solution NAS box, it would be best if I would roll my own. (12 TB before all the RAID stuff, after that it would lot less). All other solutions are very expensive (example: IBM EXP 3000 costs here 6K nis without a single hard disk). I'm planning to use hardware based RAID card, minimal Linux distribution and have some offers like iSCSI, NFS, CIFS - the usual suspects. How about looking at openfiler (http://www.openfiler.com/), consider whether you want to use it for that solution and see what kind of hardware they recommend? I have never used it but considered it before and as far as I remember it's an integration of SAN management interface on top of CentOS. Cheers, --Amos ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: CPU RAM in a storage box
I happen to have an IBM SAN storage at home, so I am familiar with IBM line of storage products. The EXP3000 is an expansion to IBM storage, which can perform for itself (JBoD), however - it contains no CPU, or RAID abilities internally. You will have to connect it to an additional server (1U, I assume) which will include the RAID hardware, if you decide to use it, and OS. I have not seen any commodity ATOM 1U systems. I know someone in my or IT Experts forum in Tapuz has asked about such a solution, and it is not simple to find. ATOMs are limited with the max amount of RAM. Desktop-class CPUs are also rather rare in the 1-2U markets, and are far from trivial to obtain, and worse - get service for. Ez On Fri, Sep 10, 2010 at 12:27 AM, Hetz Ben Hamo het...@gmail.com wrote: Hi, 2010/9/9 Etzion Bar-Noy eza...@tournament.org.il This is a joke, right? You want someone to host your system, which, by design, will not be rack-mountable, and would be large, due to the amount of disks you are to place there. It is possible, but extremely expensive to host a non-1-U server nowadays. Who would give it to you? Huh? Of course it will be rack-mountable. I'm planning to put it on a 2U or 3U chassis. I'm also not looking for someone to host my hardware, I already got a rack in Netvision today. An industrial-grade, 2U system could host, today, about 6 3.5 SATA disks. A 3 U can do much more, with up to 12-14 disks, depending on the system. IBM EXP 3000 can host 12 3.5 hard drives on a 2U chassis. See thisftp://public.dhe.ibm.com/common/ssi/ecm/en/tsd03025usen/TSD03025USEN.PDFPDF file. There are other 2U cases which can host 12 3.5 disks. And RAM is extremely important. Since you will not invest in an industial-class RAID controller (3ware, LSI-Logic, Adaptec, Intel, etc) which will cost several hundreds of dollars, as I see it, you would want to compensate for the high write latency with a large amount of RAM and fully buffered writes (not secure, but good enough). Especially with 7200RPM SATA drives with low seek speed. NFS shares, in async mode would give great performance, provided you give the system enough RAM. Then your RAM will actually become the disk write cache. Since this machine will not be using any Xeon with it's expensive RAM, I could put some gigs of RAM quite easily. Thanks for your points. Hetz Ez 2010/9/9 Hetz Ben Hamo het...@gmail.com Hi, 2010/9/9 geoffrey mendelson geoffreymendel...@gmail.com On Sep 9, 2010, at 6:35 PM, Hetz Ben Hamo wrote: I'm planning to add some big storage solution to my VPS business. I did some checking and calculated the costs, and figured out that if I want to have a decent 12TB solution NAS box, it would be best if I would roll my own. (12 TB before all the RAID stuff, after that it would lot less). All other solutions are very expensive (example: IBM EXP 3000 costs here 6K nis without a single hard disk). The question you should be asking yourself, IMHO, is what can I buy that will be as reliable as a commerical, industrial grade server? Not looking for industrial grade one. I'm planning to use hardware based RAID card, minimal Linux distribution and have some offers like iSCSI, NFS, CIFS - the usual suspects. My question is: since I'll use hardware RAID card, which processor and how much RAM should I put in such a machine? Xeon is overkill IIRC. For example, a system which costs under 900 NIS would do the job. You can get them from Ivory or KSP. They have a dual core ATOM processor, one PCI slot and one DDR2 memory slot. The power supply is not very big, but it will power a bunch of 5400 rpm green disks. This storage will be mainly used for backups. If someone wants to do a colocation to my rack, I want to give him a bonus, something that you can't find today with my competitors: I want to give him 50-100GB for storage. You'll get an NFS/CIFS/iSCSI and you mount it to your machine and use it for your backup/rsync/whatever. By comparison, when you colocate a server to Netvision's farm, you get ... 5GB backup space.. yippee.. How well will it work? How long will it last? Will it be fast enough? Fast doesn't matter much when you're doing backups or storing some temporary stuff, does it really matter when it take 20 seconds instead of 10 when you're doing rsync? I don't think so.. And the killer question, how much will it cost to replace, in the value of downtime, your time to replace it, bad will among your customers, etc? Really depends. I'm not planning to fully use all the disks, some will be disconnected or out of the RAID, perhaps I'll put a redundant PSU. Hetz Geoff. -- Geoffrey S. Mendelson, N3OWJ/4X1GM To help restaurants, as part of the stimulus package, everyone must order dessert. As part of the socialized health plan, you are forbidden to eat it. :-) -- my blog (hebrew): http://benhamo.org Skype: heunique MSN: hetz-b
Re: question about a firewall
Hi. I am in your shoes. I maintain several Linux systems hosted in Netvision (currently) for the last few years. For the last 7 years or so, I have been using iptables to protect my systems from intrusion. I have been using denyhosts to prevent unauthorized SSH logins, and prevented direct root login, or blocked all/some except my home fixed address and some other well-trusted addresses. This setup has proven itself to be effective and reliable, with zero intrusions (I stopped logging them after a while, because it's not that interesting, after all. The amount of random port scans are huge). Assuming you understand iptables, and you know how to handle it right, there is no problem with that solution. None that I have noticed. Ez 2010/9/3 Hetz Ben Hamo het...@gmail.com Hi people, As I setup my VPS/dedicated hosting here in Israel, I have been asked by the hosting company (Netvision) to either buy and bring a firewall or rent from them since the bandwidth I bought exceeds what is allowed under their firewall. They're offering Cisco 1383 (or 1838, I don't remember exactly which model). As a person who really loves Linux, I thought to myself: Why do I need to buy/rent some proprietary Cisco solution? Can't Linux handle the firewall task well? I'm sure Cisco/Checkpoint solutions are great, but yet... So here's my question: If you were in my shoes, would you take a cisco or apply some Linux solution? If you say Linux solution, what kind of solution? Could you name an app/module/whatever that can do a good protection against the usual suspect and protect against stuff like DDoS attack? I prefer the Linux solution because then I can run other services on this machine (small mail server, nagios, etc..) Suggestions? Thanks, Hetz -- my blog (hebrew): http://benhamo.org Skype: heunique MSN: hetz-b...@benhamo.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: question about a firewall
Your Cisco won't protect you against these either. There are specific DDoS protection systems, which you are not going to try and afford. Unless your servers are about gambling, porn or something very hot, you will not likely be the target of DDoS attack. I haven't been myself, for the last 7 years or so. Ez On Sat, Sep 4, 2010 at 12:42 AM, Hetz Ben Hamo het...@gmail.com wrote: I'm doing this thing right now. The only issue I worry about is attacks like DDoS. Hetz 2010/9/4 Etzion Bar-Noy eza...@tournament.org.il Hi. I am in your shoes. I maintain several Linux systems hosted in Netvision (currently) for the last few years. For the last 7 years or so, I have been using iptables to protect my systems from intrusion. I have been using denyhosts to prevent unauthorized SSH logins, and prevented direct root login, or blocked all/some except my home fixed address and some other well-trusted addresses. This setup has proven itself to be effective and reliable, with zero intrusions (I stopped logging them after a while, because it's not that interesting, after all. The amount of random port scans are huge). Assuming you understand iptables, and you know how to handle it right, there is no problem with that solution. None that I have noticed. Ez 2010/9/3 Hetz Ben Hamo het...@gmail.com Hi people, As I setup my VPS/dedicated hosting here in Israel, I have been asked by the hosting company (Netvision) to either buy and bring a firewall or rent from them since the bandwidth I bought exceeds what is allowed under their firewall. They're offering Cisco 1383 (or 1838, I don't remember exactly which model). As a person who really loves Linux, I thought to myself: Why do I need to buy/rent some proprietary Cisco solution? Can't Linux handle the firewall task well? I'm sure Cisco/Checkpoint solutions are great, but yet... So here's my question: If you were in my shoes, would you take a cisco or apply some Linux solution? If you say Linux solution, what kind of solution? Could you name an app/module/whatever that can do a good protection against the usual suspect and protect against stuff like DDoS attack? I prefer the Linux solution because then I can run other services on this machine (small mail server, nagios, etc..) Suggestions? Thanks, Hetz -- my blog (hebrew): http://benhamo.org Skype: heunique MSN: hetz-b...@benhamo.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il -- my blog (hebrew): http://benhamo.org Skype: heunique MSN: hetz-b...@benhamo.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: question about a firewall
Well. It's either I never felt it, or just never caused anything I could have felt. Ez On Sat, Sep 4, 2010 at 1:18 AM, Hetz Ben Hamo het...@gmail.com wrote: I'm not writing in my blogs about any of the issues that you mentioned, nor do I host any such content, yet I had the honour of being DDoS attacked. Hetz 2010/9/4 Etzion Bar-Noy eza...@tournament.org.il Your Cisco won't protect you against these either. There are specific DDoS protection systems, which you are not going to try and afford. Unless your servers are about gambling, porn or something very hot, you will not likely be the target of DDoS attack. I haven't been myself, for the last 7 years or so. Ez On Sat, Sep 4, 2010 at 12:42 AM, Hetz Ben Hamo het...@gmail.com wrote: I'm doing this thing right now. The only issue I worry about is attacks like DDoS. Hetz 2010/9/4 Etzion Bar-Noy eza...@tournament.org.il Hi. I am in your shoes. I maintain several Linux systems hosted in Netvision (currently) for the last few years. For the last 7 years or so, I have been using iptables to protect my systems from intrusion. I have been using denyhosts to prevent unauthorized SSH logins, and prevented direct root login, or blocked all/some except my home fixed address and some other well-trusted addresses. This setup has proven itself to be effective and reliable, with zero intrusions (I stopped logging them after a while, because it's not that interesting, after all. The amount of random port scans are huge). Assuming you understand iptables, and you know how to handle it right, there is no problem with that solution. None that I have noticed. Ez 2010/9/3 Hetz Ben Hamo het...@gmail.com Hi people, As I setup my VPS/dedicated hosting here in Israel, I have been asked by the hosting company (Netvision) to either buy and bring a firewall or rent from them since the bandwidth I bought exceeds what is allowed under their firewall. They're offering Cisco 1383 (or 1838, I don't remember exactly which model). As a person who really loves Linux, I thought to myself: Why do I need to buy/rent some proprietary Cisco solution? Can't Linux handle the firewall task well? I'm sure Cisco/Checkpoint solutions are great, but yet... So here's my question: If you were in my shoes, would you take a cisco or apply some Linux solution? If you say Linux solution, what kind of solution? Could you name an app/module/whatever that can do a good protection against the usual suspect and protect against stuff like DDoS attack? I prefer the Linux solution because then I can run other services on this machine (small mail server, nagios, etc..) Suggestions? Thanks, Hetz -- my blog (hebrew): http://benhamo.org Skype: heunique MSN: hetz-b...@benhamo.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il -- my blog (hebrew): http://benhamo.org Skype: heunique MSN: hetz-b...@benhamo.org -- my blog (hebrew): http://benhamo.org Skype: heunique MSN: hetz-b...@benhamo.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: official way to load aoe module?
RHCS is a good solution. I have just finished teaching it to one of Cisco's development teams today. If you understand its abilities and limitations, you can live happily with it. OCFS2 solves the mount/disk management issues with HA clusters. However, it does not solve application and access locking. For this there is a cluster. Ez On Mon, Aug 23, 2010 at 1:56 PM, Amos Shapira amos.shap...@gmail.comwrote: On 23 August 2010 14:47, Etzion Bar-Noy eza...@tournament.org.il wrote: Adding LUNs does not require a reboot. Removing ones do. However, if you let the cluster software manage all disk mount operations, as it should, multiple-mounts will never happen, so no need for any special masking, but only letting the cluster manage the mounts. Indeed. Just today we had a long conference call with the hosting provider's SAN expert (they provide managed shared SAN service, very useful to cut costs) and that point was raised. We ended up doing what you describe - use a none-clustered file system and hope that RHCS does its job. Nope. I have been doing it for several years now, and I can quote the procedure, if you like. Download OCFS2 package for your kernel from here: http://oss.oracle.com/projects/ocfs2/files/RedHat/RHEL5/x86_64/1.4.7-1/ (I assumed it's x86_64 platform. Also - RHEL5) Download OCFS2 tools from here: http:// http://oss.oracle.com/projects/ocfs2-tools/files/RedHat/RHEL5/x86_64/1.4.4-1/(no need for the devel and debug packages) Install all three RPMs on all nodes. Run, on each node, the command: /etc/init.d/o2cb configure Answer the questions. Run on one node the GUI command (for first-timers) ocfs2console Create a new clsuter, add all nodes with their names and IPs (if you have multiple networks, select a non-busy one) Press on propagate cluster configuration (or something similar, I don't recall exactly), and you will have to insert the root password of each node for scp command there. Following that, your cluster is ready. Create a new partition, and mount it. You will have to configure it to automatically mount via the /etc/fstab, of course. Thanks. Much appreciated. GUI and interactive tools are not an option, except maybe for first time learning-the-ropes. We configure everything automatically using puppet so we'll have to find a way to generate the files with that if we use it. Complexity. It increases the complexity of the RHCS configuration. It requires RH cluster, even if you do not. Does this mean that OCFS doesn't require RHCS? (not that it matters much now - we use RHCS for other reasons already and feel that a large part of the learning curve is behind us). Cheers, --Amos ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: official way to load aoe module?
Because it is. Not in a way you will suffer physical damage. Your legs will be fine, and so will be your hands. Your data, on the other hand, will probably be very unhealthy... Anyhow, RHCS, as a clustering infrastructure, should allow you to solve this problem with minimal chance of human error. Ez On Sun, Aug 22, 2010 at 4:22 PM, Oleg Goldshmidt p...@goldshmidt.org wrote: On Sun, Aug 22, 2010 at 2:25 PM, Amos Shapira amos.shap...@gmail.com wrote: We are a little concerned about the situation of two guests mounting the ext3 and starting to manipulate the sqlite files on it in parallel. I think you should be *very* concerned about the situation where 2 guests mount an ext3 partition and start to manipulate files *sequentially*. It looks like you *are* concerned (rightly), since you wrote that only one client *mounts* the partition at a time. Another option was to allow all guests to mount the file system read/write but carefully configure each guest to own different files or directories of sqlite files on the FS. What if one starts, e.g., creating files or appending content to existing files (and allocating new blocks, etc., in the process)? The other clients won't be aware of it. I admit I have not thought long and hard about it, but it sounds dangerous to me. -- Oleg Goldshmidt | o...@goldshmidt.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: official way to load aoe module?
Inline On Sun, Aug 22, 2010 at 4:41 PM, Amos Shapira amos.shap...@gmail.comwrote: On 22 August 2010 23:22, Oleg Goldshmidt p...@goldshmidt.org wrote: On Sun, Aug 22, 2010 at 2:25 PM, Amos Shapira amos.shap...@gmail.com wrote: We are a little concerned about the situation of two guests mounting the ext3 and starting to manipulate the sqlite files on it in parallel. I think you should be *very* concerned about the situation where 2 guests mount an ext3 partition and start to manipulate files *sequentially*. It looks like you *are* concerned (rightly), since you wrote that only one client *mounts* the partition at a time. Yes. But apart from hoping that RHCS does its job right, there is nothing preventing other guests from mounting the same partition in parallel. Of course there is - LUN masking. Its purpose is exactly this. You expose the LUN *only* to the nodes which require direct access to it. Another option was to allow all guests to mount the file system read/write but carefully configure each guest to own different files or directories of sqlite files on the FS. What if one starts, e.g., creating files or appending content to existing files (and allocating new blocks, etc., in the process)? The other clients won't be aware of it. That's why we looked at cluster-aware file systems in form of GFS but decided the performance hit is too great to go with it. A brief look at OCFS installation steps gave an impression that it isn't trivial or well supported on CentOS 5. Incorrect impression. Less than 5 minutes work, being extra slow, with coffee included. Simple enough? I admit I have not thought long and hard about it, but it sounds dangerous to me. It is. As was pointed out earlier in this thread - a large part of the file system is about how the file system module caches information in memory and synchronises it on the disk. If it's not a cluster-aware file system then parallel mounting is equivalent to opening the LV or device by an application and randomly starting to write data on it. True. But cluster-aware FS should be considered carefully. For some purposes, they ease management. For some others, they complicate it. GFS has always been misunderstood by me. It has little benefits, and major drawbacks. Can't see any reason to use it, from the existing variety of clustered FS around. Ez Cheers, --Amos ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: official way to load aoe module?
Again, inline. On Mon, Aug 23, 2010 at 2:46 AM, Amos Shapira amos.shap...@gmail.comwrote: On 23 August 2010 04:42, Etzion Bar-Noy eza...@tournament.org.il wrote: Inline On Sun, Aug 22, 2010 at 4:41 PM, Amos Shapira amos.shap...@gmail.com wrote: Yes. But apart from hoping that RHCS does its job right, there is nothing preventing other guests from mounting the same partition in parallel. Of course there is - LUN masking. Its purpose is exactly this. You expose the LUN *only* to the nodes which require direct access to it. The issue is that we want all our internal servers to be able to mount any of LUN's, as the shards of the multi-FS-multi-sqlite migrate among them. There is no way to dynamically mask and unmask LUN's, as it is now adding a new LUN requires a server reboot. Adding LUNs does not require a reboot. Removing ones do. However, if you let the cluster software manage all disk mount operations, as it should, multiple-mounts will never happen, so no need for any special masking, but only letting the cluster manage the mounts. That's why we looked at cluster-aware file systems in form of GFS but decided the performance hit is too great to go with it. A brief look at OCFS installation steps gave an impression that it isn't trivial or well supported on CentOS 5. Incorrect impression. Less than 5 minutes work, being extra slow, with coffee included. Simple enough? Thanks for the info. Do you have a good pointer to follow? Nope. I have been doing it for several years now, and I can quote the procedure, if you like. Download OCFS2 package for your kernel from here: http://oss.oracle.com/projects/ocfs2/files/RedHat/RHEL5/x86_64/1.4.7-1/ (I assumed it's x86_64 platform. Also - RHEL5) Download OCFS2 tools from here: http:// http://oss.oracle.com/projects/ocfs2-tools/files/RedHat/RHEL5/x86_64/1.4.4-1/(no need for the devel and debug packages) Install all three RPMs on all nodes. Run, on each node, the command: /etc/init.d/o2cb configure Answer the questions. Run on one node the GUI command (for first-timers) ocfs2console Create a new clsuter, add all nodes with their names and IPs (if you have multiple networks, select a non-busy one) Press on propagate cluster configuration (or something similar, I don't recall exactly), and you will have to insert the root password of each node for scp command there. Following that, your cluster is ready. Create a new partition, and mount it. You will have to configure it to automatically mount via the /etc/fstab, of course. Have you compared OCFS performance to GFS and ext3? Nope. I never bothered. I avoid GFS, as said, as much as possible, due to its complexity, and I avoid clustered FS unless I must, so ext3 is not an option then... It is. As was pointed out earlier in this thread - a large part of the file system is about how the file system module caches information in memory and synchronises it on the disk. If it's not a cluster-aware file system then parallel mounting is equivalent to opening the LV or device by an application and randomly starting to write data on it. True. But cluster-aware FS should be considered carefully. For some purposes, they ease management. For some others, they complicate it. GFS has always been misunderstood by me. It has little benefits, and major drawbacks. Can't see any reason to use it, from the existing variety of clustered FS around. What drawbacks and compared to what? We've noticed the performance problem when we tried it. What are other *practical* alternatives? GFS comes built-in and supported in RHEL/CentOS and the RHCS, that's why it was the almost natural choice in our mind. Complexity. It increases the complexity of the RHCS configuration. It requires RH cluster, even if you do not. It uses the cluster fencing methods, which is somewhat good, but somewhat bad. I can't exactly point there. Its implementation is somewhat ugly and over-complex to my liking. Ez Thanks, --Amos ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: official way to load aoe module?
Oops. Reply all is better. Insert into /etc/modprobe.conf the line: alias scsi_hostadapter2 aoe Rebuild your initrd using mkinitrd, and it will be available on startup. Ez On Sat, Aug 21, 2010 at 11:23 AM, Shlomi Fish shlo...@iglu.org.il wrote: On Saturday 21 August 2010 01:20:26 Hetz Ben Hamo wrote: Hi, I'm trying to write a small mini-guide about AoE. So far, so good.. The problem I'm facing is simple: what is the official way to autoload a module such aoe in distributions like CentOS/RedHat? I can put a simple modprobe line in rc.local but that doesn't look like a clean way.. Just for the record, for people wondering what AoE is: http://en.wikipedia.org/wiki/ATA_over_Ethernet (That's what seems most likely - it's also Age of Empires and other things). Regards, Shlomi Fish -- - Shlomi Fish http://www.shlomifish.org/ Funny Anti-Terrorism Story - http://shlom.in/enemy God considered inflicting XSLT as the tenth plague of Egypt, but then decided against it because he thought it would be too evil. Please reply to list if it's a mailing list post - http://shlom.in/reply . ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: official way to load aoe module?
Indeed. The easiest to implement, amongst the free clustered filesystems is OCFS2 by Oracle. Two or three RPMs, a short configuration phase, and you're fine. Ez On Sat, Aug 21, 2010 at 4:20 PM, Oleg Goldshmidt p...@goldshmidt.org wrote: Hetz Ben Hamo het...@gmail.com writes: I just wonder about one thing which I haven't found a solution (yet) for it: I mounted the aoe on 2 machines and I see the shared partition from the server. So far, so good. But any change that I do on any machine is not being seen on the other one. Anyone knows anything about this? AFAIK, AoE is a block level thing. What filesystem are you using? Unless it is a distributed (clustered, whatever) filesystem like AFS or GPFS you simply cannot do it. If you try this with extN or CIFS/samba or anything normal all you'll get is a bunch of corrupted files. A filesystem does not live on disk only, it is represented as a bunch of data structures in your computer's memory. You cannot mount a disk on two independent computers simultaneously and expect consistency, unless the filesystem supports it. -- Oleg Goldshmidt | p...@goldshmidt.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: official way to load aoe module?
I think OCFS2 is slightly better. Listen - if you don't need clustered filesystem, avoid it at any cost. However, if you do need it, then A/P cluster is not enough. Ez On Sun, Aug 22, 2010 at 12:01 AM, Amos Shapira amos.shap...@gmail.comwrote: 2010/8/22 Etzion Bar-Noy eza...@tournament.org.il Indeed. The easiest to implement, amongst the free clustered filesystems is OCFS2 by Oracle. Two or three RPMs, a short configuration phase, and you're fine. How is its performance? GFS comes as part of RHEL/CentOS base, so nothing special needs to be done to work with it. We tested it on a Fiber-Channel EMC SAN device, compared it to plain ext3 and xfs (yes we know that GFS is the only clustered one) and found that the performance hit too high to ignore. We ended up going back to ext3 and making sure only one guest mounts the filesystem at a time (that's what we need for our application anyway). --Amos ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: You develop in Linux and are looking for work, and are requested to provide CV as a .doc file - what would you do?
And I presume you edit your CV at least once a week. In a market where there the employers can pick, you want to play by the rule. Where you, as an employee can pick, they, mostly, will play by your rules. When you're out looking for a job, your rules are dictated by those who hire. When you are being contacted by head hunters or via friends, you can dictate the rules, again, to a level. maintaining your CV is zero work. I have invested more time in reading this thread (and replying to it) than I have invested in my CV for the last four or five years. Ez On Mon, Aug 16, 2010 at 1:25 PM, Muli Ben-Yehuda m...@il.ibm.com wrote: On Mon, Aug 16, 2010 at 08:06:54PM +1000, Amos Shapira wrote: On 16 August 2010 17:41, Oleg Goldshmidt p...@goldshmidt.org wrote: My CV is written in LaTeX as well. I convert it to HTML using latex2html and do some relatively minor manual tweaks to the result. I go through this effort specifically for people who do not have a PDF reader. All this effort to prove that you don't succumb to the fluff of using an office suite? Not answering for Oleg, but since I do the same thing -- indeed, my CV .tex template is based on Oleg's -- all this effort because writing and modifying text documents with complex layout in LaTeX is *less* effort for me than doing the equivalent in your favorite office suite. YMMV. Cheers, Muli ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: You develop in Linux and are looking for work, and are requested to provide CV as a .doc file - what would you do?
Why? Can't your OpenOffice create doc files? It you are to make a war about it, fine, but you might miss some of the better jobs because the subcontractor's HR are not technical people, and if your CV is a little more difficult to open in their own software or whatever, your loss. Next. If you do it because it doesn't look good (it might look different as DOC created by OO), then know they don't care. They feed the doc into their software and look for keywords, pretty or not. So... What is your motive? What is your war? Ez On Sun, Aug 15, 2010 at 11:28 PM, Omer Zak w...@zak.co.il wrote: I am now looking for work, and am undergoing the usual drill of sending E-mail messages in response to job/project ads and referrals. My mode of operation is to E-mail the URL to my CV in my Website. Light and sweet E-mail and the receipient's E-mail client has the convenient affordance of letting the receipient see my CV by clicking on the URL in the E-mail message. I encountered an interesting phenomenon. Some of those companies (both placement and project subcontract work outfits) look for a Linux software developer AND expect you to E-mail them a MS-Word .doc file. My current rule of thumb is to accommodate the job placement companies - they just matchmake according to keywords, and they have too much work to educate their workers about the foibles of Linux developers. But I would expect the subcontractors to have a clue about Linux developers/users (no MS-Word, in other words). If YOU were looking for work now, how would YOU deal with such companies? --- Omer -- MS-Windows is the Pal-Kal of the PC world. My own blog is at http://www.zak.co.il/tddpirate/ My opinions, as expressed in this E-mail message, are mine alone. They do not represent the official policy of any organization with which I may be affiliated in any way. WARNING TO SPAMMERS: at http://www.zak.co.il/spamwarning.html ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: You develop in Linux and are looking for work, and are requested to provide CV as a .doc file - what would you do?
On Mon, Aug 16, 2010 at 12:44 AM, Omer Zak w...@zak.co.il wrote: On Mon, 2010-08-16 at 00:13 +0300, Etzion Bar-Noy wrote: Why? Can't your OpenOffice create doc files? Actually, whenever I am due to send a .doc file, I send a .rtf, figuring that it is a more reliable way to preserve the document styling. RTF would probably mean: a. formatting would be ruined anyhow, as MS Word cannot handle it very well (and in many cases, such is the case with OO) b. You still avoid the automated filtering systems, and require human time. Human time means you will almost always be at the bottom of the list. It you are to make a war about it, fine, but you might miss some of the better jobs because the subcontractor's HR are not technical people, and if your CV is a little more difficult to open in their own software or whatever, your loss. Next. If you do it because it doesn't look good (it might look different as DOC created by OO), then know they don't care. They feed the doc into their software and look for keywords, pretty or not. So... What is your motive? What is your war? I am not fighting any wars. It is a filtering criteria, and my question is really whether it is a valid filtering criteria. Are you positive that good work (i.e. no WTF stuff, project manager with a clue, high caliber co-workers, good technical challenge, adequate opportunity for professional development, you name it) can hide behind HR which insists upon getting .doc CVs? Yes. I am positive. And not only because of the HR, but even because me, as an employer, would not have hired you if you have acted up just to make a point where it is so insignificant. Work places are no place for Holy Wars (TM). They are focused on product, on time tables, on money. Hiring someone who has the ability to perform a task but won't, just because it's not aligned with his agenda, is, well, not common. The employer does not wish to have fights. (s)he wants a working person, who would think, but obey. Your method of trying to work around the rules (hiring rules, that is) tells an employer things about you. Now, that said, I am one who sends his CV in PDF formats, because I *dont' care* about HR responses. I hardly ever get to them. If this is your case, there is no point in the question, is there? If this is not your case, you should, to my opinion, play by the rules, even if you don't like them. This is such a small effort, after all. Ez --- Omer ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Amazon EC2 hosting, was: Dedicated hosting in the US
Amazon has experienced a set of performance problems recently. Their system is overly complicated, and, except for specific usage, I would recommend people to avoid. Other cloud services exist, but still - the cloud is a buzzword which, translated to simple language is you might have performance issues, and you have no means whatsoever of finding their source. You can guess, though. This sums up most of the cloud utilizations I have seen so far, which is, also, a funny way of saying virtualization you do not maintain. Cheers. Ez On Thu, Aug 12, 2010 at 9:56 AM, Gadi Cohen dra...@wastelands.net wrote: On 12/08/2010 00:00, Maxim Veksler wrote: Are you familiar with EC2 cc1.4xlarge instance type ? http://aws.amazon.com/ec2/instance-types/ Not really... I did have a look around there a while back, but it just seemed overly complicated. I like paying a flat rate and knowing exactly what I get, especially since it's more than enough power for me for the next little while. Also with 16 cores can I quite easily divide my work into nice little VMs (xen-based), which I believe (but am not sure) that I could migrate to the cloud at a later stage. It says bandwidth is free until November... then what? Or more importantly... assuming you are in fact a proud owner of an xlarge instance, please share your use case and experiences with us :) Gadi -- Gadi Cohen aka Kinslayer dra...@wastelands.net www.wastelands.net Freelance admin/coding/design HABONIM DROR linux/fantasy enthusiast KeyID 0x93F26EF5: 256A 1FC7 AA2B 6A8F 1D9B 6A5A 4403 F34B 93F2 6EF5 / / ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Amazon EC2 hosting, was: Dedicated hosting in the US
You could find tons of I/O related issues with Amazon by a simple google. Check out this: http://blog.dt.org/index.php/2010/06/amazon-ec2-io-performance-local-emphemeral-disks-vs-raid0-striped-ebs-volumes/ This is one attempt of a solution. Clouds (SaaS) as it is now do not solve any problem. They introduce new problems without solving the storage performance issues, which are the most common problems. Clouds are excellent for non-I/O intensive systems, especially if you have the ability to fully distribute your application, and make it fully redundant. Ez On Thu, Aug 12, 2010 at 5:11 PM, Amos Shapira amos.shap...@gmail.comwrote: 2010/8/12 Etzion Bar-Noy eza...@tournament.org.il Amazon has experienced a set of performance problems recently. Their system is overly complicated, and, except for specific usage, I would recommend people to avoid. Have you got reference to this claim? We keep looking at cloud options for our servers and so far haven't found anything which could guarantee disk IO SLA for us, enough to relay our heavily disk-bound C++ programs on it. We also used Amazon hosts for small testing projects and noticed that they were noticebly more expensive than we first expected, even for very small virtual servers. All in all - I can't understand how SaaS companies manage to maintain user experience on top of Amazon and the other clouds. I'm talking about use-case examples from cloud services companies like rightscale etc. Other cloud services exist, but still - the cloud is a buzzword which, translated to simple language is you might have performance issues, and you have no means whatsoever of finding their source. You can guess, though. This sums up most of the cloud utilizations I have seen so far, which is, also, a funny way of saying virtualization you do not maintain. So far that's exactly our experience too. I'll just appreciate a pointer to specific concrete examples so I can shoot it over to the less technical people in the company who keep bugging me about using the cloud to minimise costs and overheads (we currently rent a couple of dozens of traditional dedicated 2xQuad-core servers and run our own set of xen guests on them, at a cost of quite a few tens of thousands of dollars per month). Cheers, --Amos ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Server Disk Problem
As I have said - if he could mount it, he can read the superblock. Ez On Tue, Jul 27, 2010 at 3:50 AM, Amos Shapira amos.shap...@gmail.comwrote: 2010/7/27 Etzion Bar-Noy eza...@tournament.org.il Are you checking the correct device? If fsck.ext3 can't find superblock, the system could not have mounted the device to begin with. Please post the results of lvm lvs cat /etc/fstab (if available) Thanks Ez Here is a method using testdisk to find the backup superblocks: http://www.cgsecurity.org/wiki/Advanced_Find_ext2_ext3_Backup_SuperBlock (never tried this). --Amos 2010/7/26 Daniel Feiglin dilog...@inter.net.il Hello folks! I am trying to assist in the following situation: The user has a 1u IBM Pizza server. It was configured as one partition (ext3) and loaded with Centos something-or-other, with the partition set up as a single LVM volume (yes, including the root directory). One fine day, after a reboot, it ran fsck which conked out after checking about 12% of the disk advising to run fsck manually. At that point, we logged in as root, and ran fsck -n /dev/whatever to see what's happening. The latter yielded the dreaded corrupt super block, try the next one (8193). That (as I kind of expected) didn't work either. From being root I can see the various directories an even cd to them. I am aware that it means little if their contents are corrupted. Question: 1. Are there any recovery tools for this kind of situation? 2. Is my only choice, to install another pre-partitioned hard disk, log in with (say) a live CD, mount the corrupt disk and try to manually copy my data directories? Regards, Daniel ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Alternative for getline() function in AIX 5.3
IBM supply a set of GNU utilities, including GCC-related software (and GCC, as well, if I recall correctly, but an old one) in an additional CD supplied with AIX. This is called something around Utilities for Linux or some other lie. So you do not need to force gcc to compile under AIX, but only use it. Same goes, probably, for many additional libraries. P.S - they install in RPM format :-) Ez On Tue, Jul 27, 2010 at 11:32 PM, Oleg Goldshmidt p...@goldshmidt.orgwrote: Omer Zak w...@zak.co.il writes: What I mean is to use whatever preprocessor symbol that is defined when building using gcc, but undefined on AIX; or vice versa. I see. I suspect you may need both __GNUC__ and _GNU_SOURCE to be defined for getline() to work in the context though: the first will react to compilation with gcc and the second will include the right bits from glibc. I have not checked, but it is not clear to me if one implies the other, I suspect not. In any case, the explicit requirement is that the resulting program should be portable (at least between Liux and AIX), and, I suppose, the fact that gets() is insecure should be an additional consideration. Yet another consideration is that newer incarnaions of POSIX removed gets() from requirements, AFAIK, neither Linux nor AIX actually ripped it out yet, but it may happen in the future (on Linux sooner than on AIX). So, between 1. develop a non-portable Linux/gcc branch based on getline() and choose at compile time with the preprocessor (keeping AIX insecure with the existing, gets()-based code), 2. develop a getline() branch for Linux and rewrite AIX with the secure fgets(), 3. develop a portable and secure fgets()-based implementation for both AIX and Linux, I'd vote for door #3. There is option 4 suggested by Nadav (rip getline() from glibc and use on AIX). Depending on circumstances, it may be fairly easy or completely unfeasible (it is, no doubt, feasible technically, it may be unfeasible operatonally) or anything in between. -- Oleg Goldshmidt | p...@goldshmidt.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Server Disk Problem
Are you checking the correct device? If fsck.ext3 can't find superblock, the system could not have mounted the device to begin with. Please post the results of lvm lvs cat /etc/fstab (if available) Thanks Ez 2010/7/26 Daniel Feiglin dilog...@inter.net.il Hello folks! I am trying to assist in the following situation: The user has a 1u IBM Pizza server. It was configured as one partition (ext3) and loaded with Centos something-or-other, with the partition set up as a single LVM volume (yes, including the root directory). One fine day, after a reboot, it ran fsck which conked out after checking about 12% of the disk advising to run fsck manually. At that point, we logged in as root, and ran fsck -n /dev/whatever to see what's happening. The latter yielded the dreaded corrupt super block, try the next one (8193). That (as I kind of expected) didn't work either. From being root I can see the various directories an even cd to them. I am aware that it means little if their contents are corrupted. Question: 1. Are there any recovery tools for this kind of situation? 2. Is my only choice, to install another pre-partitioned hard disk, log in with (say) a live CD, mount the corrupt disk and try to manually copy my data directories? Regards, Daniel ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: linux beivrit
Hi. I love those with the flare in their eyes, and their self-righteousness. The majority of the crowd, in Israel especially, will not be able to read ODT. Since PDF is good enough format, those who did not bother to read your document were, well, so filled with their hatred that they forgot the goal - reach maximum newbies, both on Linux and on Windows. Since probably all of this mailing list members use OpenOffice, and since this program has been known to open DOC format for the last, say, seven years (with variable levels of accuracy), the DOC and PDF formats would be fine for all your customers. The self-righteousness are probably not your crowd. However, try to avoid DOCX, as, although it is being opened by my own OpenOffice (and I would bet most of the others), it is rather new, and not well documented to begin with, so the DOC format, which would account just the same for your purposes, will not miss the new Linux users making their first move with Linux and OpenOffice. Ez On Mon, Jul 5, 2010 at 3:16 PM, Raz razi...@gmail.com wrote: Hey linux il and others In http://sos-linux.svn.sourceforge.net/viewvc/sos-linux/LinuxHebrew/ you'll find a small book for linux beginners in pdf format and doc format. anyone willing to send his reviews and remarks please send it to me to this email. please edit it in word and use track changes. access: Simply download file one at a time through the web interface. In case you wish to add an entire section, you'll be acknowledged. raz ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Common problems with Ubuntu
This seems to me like issue with the video card. Probably firmware. Ez 2010/5/11 Orna Agmon Ben-Yehuda ladyp...@gmail.com Hi Elazar, Another problem I have been experiencing for the past 3 major Ubuntu distributions (8.*, 9.*, 10.04, 64 bit OS on a 64 bit dual core) is that the X becomes extremely slow after a major operation (such as running heavy-memory Matlab scripts, or even an ad with sound on walla's weather page). It gives me the feeling that even once the application is long gone, the memory is still not really freed. I tried using google-chrome instead of firefox (which causes this problem itself sometimes), but it did not help. Even logging out does not solve the issue, not even ordering a reboot - I have to shut down and restart manually when this happens. I can no longer proudly claim that Reboot is only due to electricity outages, and I now consider going back to Debian, in which I do not recall such problems. Orna. On Mon, May 10, 2010 at 5:46 PM, geoffrey mendelson geoffreymendel...@gmail.com wrote: On May 10, 2010, at 5:05 PM, Elazar Leibovich wrote: I remeber a few times where users of this mailing list were arguing that ubuntu is a very problematic distribution. I'm evaluating a distribution for developer desktop. Ubuntu seems fitting mainly due to the hardware detection and the ease of configuration. Also, it has up to date versions of many desktop packages. I'll be happy to know which problems did you have with the Ubuntu distribution. Googling with Ubuntu problems etc, did not help me find any informative list of problems. You need to go to the UBUNTU site and look at their problem databases. They are very good at tracking problems, less good at fixing them. The problems I think you will encounter are: 1. They have a very strict release cycle with deadlines. Problems found after the freeze date for a distribution are not fixed until after the distribution. This meant in 9.04 IDE optical drives did not work, ATOM processors did not boot and a lot of minor bugs. The ATOM problem was fixed with the netbook respin, but AFAIK a new boot disk of the regular version was never released. 2. They take about a month after a release to to fix things and then often break them. For example, I have a system where gnome stopped working, and I have tried reinstalling gnome, deleting prefs, etc and it still does not work. It's too involved to reinstall from scratch. 3, They moved things around and are not like any UNIX or Linux based distro. While it's debian based, they forked off a long time ago, and debian packages often won't work, nor will any of the administration things you know. 4. They set things up the way they want them and it's darned near impossible to make them work properly if it is not what they wanted. Ask anyone with a Mac running MacOS 10.5 or 10.6 who wanted to use netatalk. 5. Long term support is a relative term. Fixes that you would think are applied are not carried back. Only the obvious critical ones. 6. Packages are not updated. Many of them are never updated, some are updated daily. I'm still faced with the same bugs in the UBUNTU version of Asterisk that were there since the original one that came with the release. In short a great desktop system for simple users, not a good one for someone to maintain or do anything beyond it. Geoff. -- geoffrey mendelson N3OWJ/4X1GM Jerusalem Israel geoffreymendel...@gmail.com New word I coined 12/13/09, Sub-Wikipedia adj, describing knowledge or understanding, as in he has a sub-wikipedia understanding of the situation. i.e possessing less facts or information than can be found in the Wikipedia. ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il -- Orna Agmon Ben-Yehuda. http://ladypine.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: WAN connection through a Linux machine
Oops - and now with reply-all... Hi. You should run both these commands (I will not disclose how you make it apply after-reboot for now) 1. echo 1 /proc/sys/net/ipv4/ip_forward 2. iptables -t nat -A POSTROUTING -o ppp0 -j MASQUERADE Don't forget to set correct DNS on your host B Ez On Tue, Apr 20, 2010 at 8:29 AM, Dan Shimshoni danshi...@gmail.com wrote: Hello, First, I don't know much about iptables. Second, I have a Linux machine (A) with two nics, which is connected to the Internet via Bezeq ADSL router. This machines runs a pppoe connection to the intenet, so the connection is done via ppp0. On eth1 I have an inner IP address which I set manually (192.168.0.10). I have a second Linux machine (B), nearby. This machine is connected directly to machine A via ethernet cable. On eth0 I have an inner IP address which I set manually (192.168.0.11). So I can ping from 192.168.0.11 to 192.168.0.10. Sometimes I want to connect from machine B to the Internet via machine A. I suppose this is possible by setting some iptables rule on machine A, and setting it to work in forwarding mode. Could this done by setting one iptables rule on machine A? What is the proper rule to achieve it? DS ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: High Availability
The answer to the original post is It's not HA, it's MA, however... Middle-Availability is not high-availability, of course, but it's a good start. HA is a very nice buzz word. HA should protect you against what? software failure? OS failure? hardware failure? network failure? storage failure? site failure? Internet failure? Each of these add some small (or huge) amount of money to the total solution. You can protect yourself from everything you would ever consider, but I presume a fully continent-distributed HA solution with zero loss is not cheap, and probably, more than you need. To give an example you could relate to - if you're going to buy a car, you will not buy yourself (probably) a Farari. Not because it's not a good car - it's excellent car, probably way better than what you would have imagined yourself driving in - it's excellent, but you need good enough. I tend to divide HA solutions into several groups, each has its own price tag and level of protection. They should match the requirements, and more than that - the customer should be aware of the protection he gets. Don't expect it to do miracles, just know what's expected of it. HA is much like backup, or like insurance. You want it there, but you don't want to use it, and again - you would not buy the one protecting your teeth with 4M$. The original question level of protection is against these: Application failure OS failure Application misconfiguration It does not protect against network failure (maybe against link failure if it's over bonding), or storage failure (I presume it's local storage). However, it's a cheap solution and easy to understand. The pros of this specific solution is that the infrastructure is already there, Investing the required amount of money, and he can split it into two separated machines, with shared storage device, and gain some hardware failure protection, without any logical change to the infrastructure or the design of the current systems. This is a solution he can grow with to everything he wants. This is a good solution, based on the expected expenses, and the fact that most companies cannot invest the capital required for enterprise-class HA solutions. Not everyone can afford EMC DMX4 with SRDF to another EMC, the leased line between them, the geo-cluster software and the distributed and fully redundant network devices involved. Compromises are common, and as long as the cluster functionality is tested, and the behavior is expected and documented, this can be called MA, which is poor-people's-HA. I say - if you know what you're doing - this is good enough. Ez P.S - Amos, sorry for sending this directly to you. The list behavior dictates that reply would send to the person sent the mail, and not the list. Sorry. Ez On Thu, Apr 15, 2010 at 3:12 PM, Amos Shapira amos.shap...@gmail.comwrote: On 15 April 2010 17:50, guy keren c...@actcom.co.il wrote: and - test test test. many people fail to test their highly-available setup, and as a result, think they are 'highly available when they are not. Great point! Earlier at my current position we failed to test that the fail-over works on one of our servers after an upgrade and got beaten the next time the primary went south. Since then it's our standard practice to: 1. update the current stand-by 2. switch over (effectively updates the service to the new version). 3. Wait a while and see that everything is fine (the current stand-by still ready with the old proven version). 4. Update the current secondary. 5. Switch back to make sure HA works right. (6. For good measure - switch again). We also do all of this through an elaborate set of scripts around xen, kickstart and puppet (keeping the xen guest images in LV's, came handy when we had to roll-back) so all this deployment procedure is completely automated and repeatable so we exercise not just the in-house built software and its configuration setup but also the deployment procedure itself. We reached a stage where a single operations engineer upgrades our entire production system of about 50 virtual servers across 18 physical servers in three mornings of running automatic scripts (i.e. no need for manual configuration changes). (We stick to mornings because of a rule not to schedule production changes in the afternoons or before a weekend, unless absolutely necessary). --Amos ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Switch between virtual consoles CtrlAltFx problem.
Use Alt+F1-7 when in console mode. Don't use Ctrl+Alt+F1-7. Ez 2010/3/23 David Harel harel...@gmail.com Hi, Using Ubuntu 9.10 (karmic) with plain gnome desktop. When I switch to text only console using the CtrlAltFx keys I can't travel back to X11 mode using CtrlAltF7 or switch to any other virtual console. I use in my ~/.bashrc the command: setxkbmap -option grp:lswitch,grp:alt_shift_toggle,grp_led:scroll us,il Maybe that is the cause. Can I bind different keys for console switching? -- Thanks. David Harel, == Home office +972 77 7657645 Cellular: +972 54 4534502 Snail Mail: Amuka D.N Merom Hagalil 13802 Israel Email: harel...@ergolight-sw.com ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Xen and storage
On Mon, Mar 15, 2010 at 10:41 AM, Amos Shapira amos.shap...@gmail.comwrote: 2010/3/12 Hetz Ben Hamo het...@gmail.com: Hi, I have taken 3 machines for a project: 2 machines will act as Xen servers and one machine will act as storage. The storage box is just a machine with few hard disks connected with a RAID controller. What I would like to do is create few Xen VM's with the fastest possible I/O in terms of storage. I have few options: 1. I can create an LVM on the storage machine, create few Logical Volumes and export them as NFS to the Xen servers and configure each VM to some file images. Problem is, that file I/O with Xen is slower compared working with LVM's. 2. I can create an LVM on the storage machine, create few Logical Volumes, and export those as iSCSI devices. I'm not sure whats the performance of Xen with iSCSI devices exported from the storage box. 3. I can create few partitions on the storage machine, export them as iSCSI devices and do LVM on the Xen servers. Problem: I don't know how much the penalty doing LVM on the Xen machines. My question: What is the best option? Thanks, Hetz I don't have practical experience with hosting Xen images on SAN but when I researched the market for a SAN-based configuration of our production network (currently 20 Xen hosts hosting about 10 Xen guests each, doing DRBD between pairs of Xen guests and linux-ha for HA), at least one or two of the options I checked mentioned that if I store the Xen images on the SAN then it will require much higher bandwidth to it than if I use it just for plain data. Why? Where does the secret IO arrive from? Ez Based on this input, I'd recommend that you'll look at having the images on internal server disks and try to achieve HA at the xen guest level, then compare the performance with iSCSI hosted xen images. Hope this helps. --Amos ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Xen and storage
Assuming your farm is a production one, suspend to disk is rather rare, and by design, would probably not be a day-to-day process of the system. Anyhow - suspend to disk is an operation which is being performed on NFS SR as well, just the same (create file which contains memory dump of the VM). NFS has other advantages, like LUN alignment and thin provisioning (assuming you did not purchase the foundation for Citrix XenServer package). Also - for real-life production systems I have seen that network communication over iSCSI, for about 50 VMs on 5 physical servers would not exceed the 200Mb/s at peak times. Of course - specific applications can (and will) stress the storage and network, however, many common storage devices cannot maintain a high rate of random IO (common to DBs, like Oracle, MySQL, Exchange, MSSQL, etc). The disks would commonly be the bottleneck, and not the network/FCS transport. Don't believe me? Check your virtual farm. See what throughput you get for your DRBD/central storage links. Ez On Mon, Mar 15, 2010 at 10:58 PM, Amos Shapira amos.shap...@gmail.comwrote: 2010/3/16 Etzion Bar-Noy eza...@tournament.org.il: On Mon, Mar 15, 2010 at 10:41 AM, Amos Shapira amos.shap...@gmail.com wrote: 2010/3/12 Hetz Ben Hamo het...@gmail.com: Hi, I have taken 3 machines for a project: 2 machines will act as Xen servers and one machine will act as storage. The storage box is just a machine with few hard disks connected with a RAID controller. What I would like to do is create few Xen VM's with the fastest possible I/O in terms of storage. I have few options: 1. I can create an LVM on the storage machine, create few Logical Volumes and export them as NFS to the Xen servers and configure each VM to some file images. Problem is, that file I/O with Xen is slower compared working with LVM's. 2. I can create an LVM on the storage machine, create few Logical Volumes, and export those as iSCSI devices. I'm not sure whats the performance of Xen with iSCSI devices exported from the storage box. 3. I can create few partitions on the storage machine, export them as iSCSI devices and do LVM on the Xen servers. Problem: I don't know how much the penalty doing LVM on the Xen machines. My question: What is the best option? Thanks, Hetz I don't have practical experience with hosting Xen images on SAN but when I researched the market for a SAN-based configuration of our production network (currently 20 Xen hosts hosting about 10 Xen guests each, doing DRBD between pairs of Xen guests and linux-ha for HA), at least one or two of the options I checked mentioned that if I store the Xen images on the SAN then it will require much higher bandwidth to it than if I use it just for plain data. Why? Where does the secret IO arrive from? I haven't dug into this but I figured it was around reading the program files from the storage to the iSCSI client which actually runs the Xen image and storing the Xen guest's state if you use xen's suspend to disk stuff. --Amos ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Where to learn Linux?
Kids will always be kids, and there will always be competition. I think competition is good. Otherwise - go be a Cobol developer. No competition there, you know. Not in five years time, when everyone who knows something will retire. Linux is a dynamic and evolving profession. It gets more and more into the core systems of companies' infrastructure. It becomes, well, a commodity OS. There is always a shortage of excellent people. Moderate, you'll find anywhere you look. Good you'll find some. Excellent - hardly ever. These people get nice jobs, and get a whole selection with their price tag, be it high (well, reasonable, but high). Be a Linux sysadmin. Be a good Linux sysadmin, think like you should in a corporate, but have the spark to rethink old concepts and force others around you to think of them (don't nag, though), to examine them. See that corporates (in Israel, in particular) are very conservative, and very slow-moving, and think of how, with OSS technologies, you can make things better, more efficient, more manageable, but still - make it fully documented, make it fitting to a corporate and fully usable to the person who would be there after you. Be excellent, and I tell you - I would be happy to employ you myself. Ez 2010/3/14 Hetz Ben Hamo het...@gmail.com Hi Dotan, My recommendation is simple: don't go there. Why? the market is full, salleries went down, and you'll find yourself competing against kids. My suggestion: be a Linux developer. Hetz 2010/3/14 Dotan Cohen dotanco...@gmail.com I have been using Linux as an end user for several years, but now I think that I might like to make a career out of *nix administration. Where are some good places to get a certificate from? Is an online certificate as good as an offline course? What online certificates are honourable? What real-world courses in Israel are recommended? Thanks! -- Dotan Cohen http://bido.com http://what-is-what.com Please CC me if you want to be sure that I read your message. I do not read all list mail. ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il -- my blog (hebrew): http://benhamo.org Skype: heunique MSN: hetz-b...@benhamo.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: Where to learn Linux?
Actually, you should be in another profession to make lots of money. You won't make your millions with computers, you know. Programming and syadmin are two different disciplines, and none is better than the other. It should be first what you love doing, otherwise, you will never be excellent in it. Ez 2010/3/15 Hetz Ben Hamo het...@gmail.com Oh, I'll tell you more.. Want to make the biggest sallery? be a kernel programmer. Try to fix or write a driver and make sure it's committed to the GIT (your code will be examined). Once it's in the GIT, it will be worth a lot for you. Learn some linux embedding skills, the GNU chaintool, Xorg, etc - any of these will get you some credits that worth more then recommendations. Hetz 2010/3/14 Dotan Cohen dotanco...@gmail.com My suggestion: be a Linux developer. Tell me more. Right now I have people coming to me asking me to admin their servers. By Linux developer do you mean to develop Linux software or the kernel? -- Dotan Cohen http://bido.com http://what-is-what.com Please CC me if you want to be sure that I read your message. I do not read all list mail. -- my blog (hebrew): http://benhamo.org Skype: heunique MSN: hetz-b...@benhamo.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: better platform for virtualization
PV drivers were released by Oracle, who run their own virtualization platform based on XenCommunity. KVM is wasteful and requires VT support even for Linux machines. Not only that, but its virtualized hardware is legacy old hardware supplied by QEMU. The leading virtualization solutions currently in the market are Vmware ESXi, which, for a single server is free and very nice (although you better make sure your hardware is in their support matrix), Microsoft Hyper-V, which is not Linux-friendly, if you care about it, and Citrix XenServer. You can compare these three on the internet. Less common, but aggressively in use (with the same performance profile) are OracleVM, RHEL Xen platform (with Oracle PV drivers for Windows) - Xen Community based. On the level of the low performance you would find Vmware Server, and Sun VirtualBox. KVM was designed, and is focused on VDI - desktop virtualization, being the focus of Kumranet in the past. RedHat cannot maintain two virtualization platforms. Assuming you aim at Windows virtualization, and assuming you want the assurance of enterprise-class solution, I would recommend any of the top three, with my favorite Citrix XenServer (check the pros and cons of each and make your own decisions in that matter). Other solutions are not complete, or will have near-zero support or knowledge base on the net. To make things clear - I earn my keep by performing various system and IT architecture-related integration operations, amongst are virtualization design and implementations. I have several customers with large XenServer farms, running mission-critical, 24/7/365. The longer one has his farm running since about March, containing about 40-50 VMs, including the company's Exchange server (~400 users), several Oracle DB environments, MSSQL, ADS, TS and more. These are memory-hungry applications, and the total memory allocated there (usually memory is the immediate bottleneck, followed by disk IO performance) is about 150GB ram, in total, with shared storage and the entire shabang. Other customers of mine use smaller environments, running one or two servers, with several tenths of vms on them. They were unable to reach anywhere near this capacity (amount of VMs, performance of every single VM) using VMware Server, of course. Make your own pick. I can only recommend to use the enterprise class tools, especially if they can come for free, and/or you could buy support. Ez On Tue, Jan 19, 2010 at 10:11 PM, Oleg Goldshmidt p...@goldshmidt.orgwrote: Jonathan Ben Avraham y...@tkos.co.il writes: Hi Gilad, Why do you recommend KVM over XEN? Have you fiddled with both? Are there particular problems with XEN? Apart from the fact that XEN is paravirtualization technology and running a mission-critical Windows DomU is possible mostly in theory? Disclaimer: I have not touched Xen over a couple of years (when Windows guests were possible on KVM, at least in principle, and not possible on Xen). I checked the current docs out of curiousity and phrases like PV drivers are being developed and you need to disable driver signature checking on (every!) reboot [original emphasis] don't inspire much confidence. Other points Gilad made (KVM being much less intrusive and already in the vanilla kernel and provided by RedHat) are very much valid. To the OP: Xen is not for you. I have no first-hand experience (beyond a tiny bit of tinkering) with KVM. I have quite a bit of production experience with VMware. I am surprised that most of the postings focus on the VMware Server (previously known as GSX). IIRC the OP mentioned crucial servers but did *not* say $0 was a requirement. I'd go with ESX for mission-crtical stuff. For a serious installation I would not keep data (or system images, for that matter, but YMMV) on directly attached disks. -- Oleg Goldshmidt | p...@goldshmidt.org ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il ___ Linux-il mailing list Linux-il@cs.huji.ac.il http://mailman.cs.huji.ac.il/mailman/listinfo/linux-il
Re: better platform for virtualization
True indeed. XenCommunity is a fine option, which I have found to be good. I have been running a bunch of servers on it, from a single VM on a physical server (to achieve the management benefits with the very minimalistic loss of Xen) to several tenths of VMs on a server in several farms abroad. I have there about 16 servers in each farm, running several tenths of mission critical VMs. The setup is elegant - any new physical server added to the farm is being automatically installed and defined to be able to run VMs, hands-free. This is something easier to manage when using RHEL/Centos, compared to other products, indeed. Still - the mass use one of the three leading solutions. You would see them in the fortune500, and on many other sites. Notice, again, that KVM is VDI-focused, and as the battle in the server virtualization rages between the leading commercial vendors, the VDI market is somewhat quieter. It's about prestige and ego, but everyone wants to virtualize servers, as desktops are less glorious and require harder work, on most cases. RHEL has made a step towards VDI, with a very clear view of the future, and KVM is their tool. Ez On Wed, Jan 20, 2010 at 3:46 AM, Amos Shapira amos.shap...@gmail.comwrote: 2010/1/20 Etzion Bar-Noy eza...@tournament.org.il: PV drivers were released by Oracle, who run their own virtualization platform based on XenCommunity. Just wondering - are these required to be installed separately when trying to run Windows on CentOS? We generally managed to do that when we tried (got stuck on none of our licenses being accepted). KVM is wasteful and requires VT support even for Linux machines. Not only that, but its virtualized hardware is legacy old hardware supplied by QEMU. The leading virtualization solutions currently in the market are Vmware ESXi, which, for a single server is free and very nice (although you better make sure your hardware is in their support matrix), Microsoft Hyper-V, which is not Linux-friendly, if you care about it, and Citrix XenServer. You can compare these three on the internet. Less common, but aggressively in use (with the same performance profile) are OracleVM, RHEL Xen platform (with Oracle PV drivers for Windows) - Xen Community based. On the level of the low performance you would find Vmware Server, and Sun VirtualBox. KVM was designed, and is focused on VDI - desktop virtualization, being the focus of Kumranet in the past. RedHat cannot maintain two virtualization platforms. I expect this targeting will change quickly since RH plan to replace replace Xen by KVM (in 5.5 or 6.0?) Assuming you aim at Windows virtualization, and assuming you want the assurance of enterprise-class solution, I would recommend any of the top three, with my favorite Citrix XenServer (check the pros and cons of each and make your own decisions in that matter). Other solutions are not complete, or will have near-zero support or knowledge base on the net. To make things clear - I earn my keep by performing various system and IT architecture-related integration operations, amongst are virtualization design and implementations. I have several customers with large XenServer farms, running mission-critical, 24/7/365. The longer one has his farm running since about March, containing about 40-50 VMs, including the company's Exchange server (~400 users), several Oracle DB environments, MSSQL, ADS, TS and more. These are memory-hungry applications, and the total memory allocated there (usually memory is the immediate bottleneck, followed by disk IO performance) is about 150GB ram, in total, with shared storage and the entire shabang. I do not have the tools (or wish) to dispute your recommendations above, they sound reasonable and well baked, but if we are about to demonstrate what can be achieved with the options then let me describe what I got running on CentOS 5 (and the old Xen 3.0 which comes with it): 18 physical 2xQuad-core servers with at least 64Gb RAM (some have 80Gb but we decided to stay clear form 8Gb DIMM's) in production. 4 more servers in test env, also with 64Gb RAM each. To do the calculation for you - this is upwards of 1408 Gb of RAM. A total of 70+ virtual guests. ~1.5Tb disk space, on 12 spindles, in each server. That system have been up for over two years now. Only unplanned down time we had were either due to hardware issues or our own mistakes, not the platform's fault. We use DRBD to replicate disks, linux-ha for heartbeat and fail-over, LVS for Virtual IP load balancing. Our up time on the worst part of the system (an old customer portal we are about to replace) is %99.93 in 11 months (5:20 hours of down time since February 2009, that's the beginning of the monitoring data of the current monitoring tools), up time on better parts (properly load balanced) is up to %99.98 since around the same time (1:25 hours down since March 2009
