Kernels 2.4.x for ISP ?
I'm just setting up new box - large mail server, dual Pentium, STL2 motherboard, Mylex AccleRAID. I'd like to try with kernel 2.4.x + ResierFS, but I worry a little about it's maturity. Is it really stable now ? Can anyone here write about his/her experiences with 2.4.x ? Does 2.4.x kernels support AcceleRAID 352 well ? Greetings -=Czaj-nik=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Kernels 2.4.x for ISP ?
On Fri, 18 May 2001, Tech Support wrote: > I can't comment on the ResierFS or AcceleRAID support but I can > comment on the stability of the 2.4.2 kernel. > We have been running 2.4.2 on our Web servers which also do DNS, mail, > mysql and FTP for 50 days now, without a problem. The 2.4.x kernel is > definitely faster then 2.2.x (no I didn't benchmark it :-) and it is > perfectly stable. > We have also run 2.4.3 on a test box and have subject that to some > ugly situations without harm. What configuration ? I want to run it on dual processor machine. > However, be aware of the security hole (security cavern!) in iptables, > so don't use the firewalling features. The problem is supposed to be > fixed in 2.4.5. Thanx. 2.4.5 fixes Adaptex AIC7xxx drivers also. > Also be sure to read the kernel source as you must upgrade modutils > and few other packages from woody to get 2.4.x to work. I know, I'm running 2.4.3 on my desktop comps. TIA -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Kernels 2.4.x for ISP ?
On Fri, 18 May 2001, Tech Support wrote: > Przemyslaw Wegrzyn wrote: > > What configuration ? > > I want to run it on dual processor machine. > Our web servers are single processor Pentium III on a Asus > motherboard. I also run 2.4.2 with SMP on a couple of Dell Poweredge > duel 1Ghz processor, 2Gb memory with aacRAID for a client. These servers > are running large Oracle 8i databases so I'm not sure their performance > translates to Web service but the SMP support is fine and stability has > been equal to my non-SMP machines. The kernels are patched to use the > aacRAID. What patch ? Isn't DAC960 module already included in kernel ? -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Kernels 2.4.x for ISP ?
On Fri, 18 May 2001, Tech Support wrote: > > What patch ? Isn't DAC960 module already included in kernel ? > Yes Mylex DAC960 support is in the kernel but this isn't the > controller that these machines used. Dell uses (or at least in these > machines) a controller made by Adaptec which is alternately called > PercRAID and aacRAID. The patch and info is at http://domsch.com/linux/ Aaaah, thanx a lot ! We have a choice Adaptec 2100S or Mylex AcceleRAID 170. We've decided to take the second - because of the kernel support. Adaptec seems to ignore linux - on their web page I've found drivers for 2.2.14 :| -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Kernels 2.4.x for ISP ?
On Fri, 18 May 2001, Krzysztof Mazurczyk wrote: > And third problem may be with dependencies of packets. Kernel suggests use > 2.4.x > of modutils, modutils depends on libc6 2.2.x and newer libc6 depends on > libdb2 2.7.7 > or similar. > But when I listalled libdb2 taken form woody lastly, syslogd didn't start. > Yesterday I send > question to debian-user list with this problem but I've no answer as yet. > It may be > nessesary to use 2.4.x kernel without modules. Well, there are packages available allowing use of 2.4.x with potato. I already use 2 potato hosts with 2.4.3 kernel. Works very fine, with no stability problems. But I wonder if the same kernel will be stale on different hardware - SMP, SCSI and so on. I've just subscribed linux-kernel mailinglist to see what's going on there... Greetings -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apache suEXEC problem
On Sat, 19 May 2001, Hector Castillo wrote: > It's impossible to move the pages to the original directory (there are > a lot users and a lot pages), and the suEXEC is necessary. > My question is: żdoes anybody know how to recompile the Debian source > of apache-ssl and which files would be necessary to change? Look at debian/rules of apache's source package. There's option to 'configure', specifing allowed dirs. -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Tomcat problem
I get the following exception starting tomcat. What can cause this error ? Tomcat starts and seems to work (at least exampels work properly) TIA -=Czaj-nick=- Starting tomcat. Check logs/tomcat.log for error messages org.apache.xerces.validators.datatype.InvalidDatatypeFacetException: org/apache/xerces/utils/regex/ParserForXMLSchema at org.apache.xerces.validators.datatype.DatatypeValidatorFactoryImpl.createDatatypeValidator(DatatypeValidatorFactoryImpl.java) at org.apache.xerces.validators.datatype.DatatypeValidatorFactoryImpl.createDatatypeValidator(DatatypeValidatorFactoryImpl.java) at org.apache.xerces.validators.datatype.DatatypeValidatorFactoryImpl.initializeDTDRegistry(DatatypeValidatorFactoryImpl.java) at org.apache.xerces.validators.common.XMLValidator.initDataTypeValidators(XMLValidator.java) at org.apache.xerces.validators.common.XMLValidator.setValidationEnabled(XMLValidator.java) at org.apache.xerces.framework.XMLParser.setValidation(XMLParser.java) at org.apache.xerces.framework.XMLParser.setFeature(XMLParser.java) at org.apache.xerces.parsers.SAXParser.setFeature(SAXParser.java) at org.apache.xerces.jaxp.SAXParserImpl.(SAXParserImpl.java) at org.apache.xerces.jaxp.SAXParserFactoryImpl.newSAXParser(SAXParserFactoryImpl.java) at org.apache.tomcat.util.xml.XmlMapper.readXml(XmlMapper.java) at org.apache.tomcat.context.WebXmlReader.processWebXmlFile(WebXmlReader.java) at org.apache.tomcat.context.WebXmlReader.contextInit(WebXmlReader.java) at org.apache.tomcat.core.ContextManager.initContext(ContextManager.java) at org.apache.tomcat.core.ContextManager.init(ContextManager.java) at org.apache.tomcat.startup.Tomcat.execute(Tomcat.java) at org.apache.tomcat.startup.Tomcat.main(Tomcat.java) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
kernel 2.2.19 limitations.
What is the maximum file size in 2.2.19 kernel ? Still 2GB ? Is it ext2 limitations , or ? Will ReiserFS have the same limit using 2.2.19 kernel ? -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: kernel 2.2.19 limitations.
On Sun, 27 May 2001, Peter Billson wrote: > Whoops... helps if I post *to* the list too! > > > Yes the limit is still the usual 2Gb. The limit is actually with ext2, i > > believe, although I'm not sure. > > The limit is in the kernel, not the ext2 file system, otherwise 2.4.x > wouldn't be able to support >2Gb file either. There are patches about > for adding LFS (large file system) support. I had compiled a 2.2.18 > kernel after patching with the LFS patch borrowed from RedHat's 6.2ee > (Enterprise Edition) source. > > But why not just run 2.4.x? Hmm, I'm building some so called "very important server". I'm not sure if 2.4.x is stable enough. It will run Apache, biiig PostgreSQL database, all on big RAID. Anyone has experience with 2.4.x + SMP (2 x PIII) + Mylex AcceleRAID ? After reading some newsgroups I believe 2.4.x kernels works pretty stable or does totaly weird things, all depends of the hardware configuration. Well, it seems some drivers are not yet stable enough... -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Ping - what the hell ?
Look at this: czajnik@earth:~$ ping 156.17.209.1 PING 156.17.209.1 (156.17.209.1): 56 data bytes 64 bytes from 156.17.209.1: icmp_seq=0 ttl=247 time=5427.7 ms 64 bytes from 156.17.209.1: icmp_seq=1 ttl=247 time=23.2 ms 64 bytes from 156.17.209.1: icmp_seq=2 ttl=247 time=429492829.5 ms ^^^ 64 bytes from 156.17.209.1: icmp_seq=3 ttl=247 time=429492907.1 ms 64 bytes from 156.17.209.1: icmp_seq=4 ttl=247 time=16.5 ms 64 bytes from 156.17.209.1: icmp_seq=5 ttl=247 time=21.4 ms What the hell can it be ?!? Every ping comes back in few milisecounds... We've recently added second CPU to this machine, can it (failed CPU) be the reason ? Overall stability is good... -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Ping - what the hell ?
On Wed, 30 May 2001, Ken Seefried wrote: > > I think that you can get this if you have an MP kernel compiled without > "Enhanced Real Time Clock" support. The default clock driver apparently > isn't MP-safe. Thx, I'll check it tomorrow :). -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Zmailer question.
Hmm, I use qmail almost everywhere, but friend asked me for some help with zmailer. Almost everything is done, but also we need ability to send mails to all our customers - all accounts in all virt domains. What's the easiest solution with zmailer ? -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Ping - what the hell ?
On Wed, 30 May 2001 [EMAIL PROTECTED] wrote: > Aren't you traveling several rfc1149 links? > > http://www.blug.linux.no/rfc1149/pinglogg.txt ?!? What do U mean ? isn't the ping time measured by storing system time the ICMP ECHO was sent, and comparng it to the system time the reply arrived ? I get it even when pinging my Cisco 1601 router, which is connected directly (I don't care 3com SSII 1100 now) -=Czaj-nick=- > On Wed, May 30, 2001 at 03:24:39PM +0200, Przemyslaw Wegrzyn wrote: > > > > Look at this: > > > > czajnik@earth:~$ ping 156.17.209.1 > > PING 156.17.209.1 (156.17.209.1): 56 data bytes > > 64 bytes from 156.17.209.1: icmp_seq=0 ttl=247 time=5427.7 ms > > 64 bytes from 156.17.209.1: icmp_seq=1 ttl=247 time=23.2 ms > > 64 bytes from 156.17.209.1: icmp_seq=2 ttl=247 time=429492829.5 ms > > ^^^ > > 64 bytes from 156.17.209.1: icmp_seq=3 ttl=247 time=429492907.1 ms > > 64 bytes from 156.17.209.1: icmp_seq=4 ttl=247 time=16.5 ms > > 64 bytes from 156.17.209.1: icmp_seq=5 ttl=247 time=21.4 ms > > > > What the hell can it be ?!? Every ping comes back in few milisecounds... > > > > We've recently added second CPU to this machine, can it (failed CPU) be > > the reason ? Overall stability is good... > > > > -=Czaj-nick=- > > > > > > > > -- > > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > -- > HoraPe > --- > Horacio J. Peńa > [EMAIL PROTECTED] > [EMAIL PROTECTED] > [EMAIL PROTECTED] > [EMAIL PROTECTED] > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Ping - what the hell ?
On Wed, 30 May 2001, Nathan E Norman wrote: > On Wed, May 30, 2001 at 08:47:38PM +0200, Przemyslaw Wegrzyn wrote: > > > > > > On Wed, 30 May 2001 [EMAIL PROTECTED] wrote: > > > > > Aren't you traveling several rfc1149 links? > > > > > > http://www.blug.linux.no/rfc1149/pinglogg.txt > > > > ?!? What do U mean ? > > It was a joke ... rfc 1149 is IP over avian carriers (birds) Ooooh, yeah, I didn't remember the rfc number, but I know this protocol : It's nice :) Anyway, my problem seems to be hardware: czajnik@earth:~$ more /proc/misc Segmentation fault czajnik@earth:~$ -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Zmailer question.
On Thu, 31 May 2001, Russell Coker wrote: > On Wednesday 30 May 2001 19:49, Przemyslaw Wegrzyn wrote: > > Hmm, I use qmail almost everywhere, but friend asked me for some help with > > zmailer. Almost everything is done, but also we need ability to send mails > > to all our customers - all accounts in all virt domains. What's the > > easiest solution with zmailer ? > > If you are using Maildir storage then my maildir-bulletin package might do > what you need. Nope. It makes no problem with Maildirs, but here it's still on mailboxes (will change soon). I see generaly (on every MTA) 2 ways - alias to all (but there's a risk of abuse), and second way - script simply sending to everyone in the prepared list. -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Ping - what the hell ?
On Sat, 2 Jun 2001, Craig Sanders wrote: > On Wed, May 30, 2001 at 09:41:54PM +0200, Przemyslaw Wegrzyn wrote: > > Anyway, my problem seems to be hardware: > > > > czajnik@earth:~$ more /proc/misc > > Segmentation fault > > czajnik@earth:~$ > > some possible causes: > > 1. bad memory - most likely. > > 2. bad swap partition (or bad disk controller causing the swap partition to > not work) > > 3. other bad hardware > > 4. bad libc6 or other library - not very likely. > It' solved, there were 2 reasons. Core dumps - hmmm, our admin borken the kernel by incorrectly patching it. Ping times - some stupid guy inserted two different CPUs PII 400 and 450. It's a miracle it was working all together... -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Ping - what the hell ?
On Sun, 3 Jun 2001, Chris Wagner wrote: > I'm sorry, but ROFLMAO!!! It's sad and (sometimes) funny, that I have to work with those people ;) We are just changing our admin. He was a real mistake :| Now it's all funny for me, but It costed me time, lots of time... > >It' solved, there were 2 reasons. > > Core dumps - hmmm, our admin borken the kernel by incorrectly patching > >it. > > Ping times - some stupid guy inserted two different CPUs PII 400 and 450. > > It's a miracle it was working all together... -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Webmail - considerations...
We are going to develop web-mail system, that's capable of handling relatively high loads. I know, there are many open source web-mail systems , but they doesn't satisfy me. Almost every falls into one of two cateogries: php based, using imap; perl cgi based, using IMAP or direct filesystem access... I'd like someone experienced with such systems help me with the following: 1. Almost every available webmail system uses the following way of handling (rreceiving, in this example) attachements: load the whole message body from IMAP server or message file, decode it and send to the client. The _whole_ attachement gets loaded into server's RAM. Isn't it waste of resources/ killing the server ? I think it should read/decode/send the attachement on a line-by-line (or part-by-part generaly) manner. Am I right ? 2. Which one is better - accessing maildirs directly, or using IMAP ? I can see that IMAP seems to be more scalable / universal... Maildirs probably can be much faster to work with directly, but probably less secure... Any other pros/contras ? 3. I'm going to develop the front-end using Apache::ASP or php, not decided yet, and access the mails through the middle-tier daemon. The question is - is it a good way to use persistent IMAP connections ? If so, there will be no overhead of authentication on every operation, but there can be many open IMAP connections to the local imap server (probably Courier-IMAP) at the same time. Which strategy is better ? 4. Are there any libraries similar to c-client, maybe some C++ ones ? 5. Does c-client library allow to retrieve the message body (attachements) partialy (see 1) ? I've seen some /tmp access in its source code - does it dowload whole message to /tmp ? Hmm, that's all for now... TIA -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Perl 5.6 & Potato
Is it possible to run perl 5.6 iwth Potato without upgrading 50% of the system ? I've tried to build debs from Woody, but they fails during ndbm tests... -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Webmail - considerations...
On Tue, 12 Jun 2001, Russell Coker wrote: > Another thing, do the webmail systems use directory notifications? With > directory notifications the application can be informed by the kernel > when new files are created. The IMAP protocol supports announcing new > mail to the client. If the mail store and the webmail system are on > separate machines then IMAP is the only way to avoid polling for new mail > (AFAIK directory notifications don't work over NFS). Thanx, I didn't know this feature... Acctually, I'm not so familiar with IMAP protocol yet... > So now you're planning to write your own webmail program? Yes, we need some nonusuall features... > I suggest that when a user logs in they get an IMAP connection. An > option to have the IMAP connection timeout and be closed before the > webmail system times out (webmail timeout on no activity should be long > like 30 minutes while IMAP timeout should be 10 minutes at most) would be > handy, but isn't required. But generally IMAP connections should stay > open for the duration of the session. That's as I see it, thanx.. All the actual logic will be implemented in the middle tier daemon(s). -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Webmail - considerations...
On Tue, 12 Jun 2001, Jeremy Lunn wrote: > On Tue, Jun 12, 2001 at 01:26:09PM +0200, Przemyslaw Wegrzyn wrote: > > Thanx, I didn't know this feature... Acctually, I'm not so familiar with > > IMAP protocol yet... > > Depending what language you use you won't need to know the IMAP > protocol. Things like perl and php have IMAP modules. Of course, I know... But our middle-tier will be developed using C++ , AFAIK. That's why I asked about c-client library... -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Webmail - considerations...
On Tue, 12 Jun 2001, Jeremy Lunn wrote: > On Tue, Jun 12, 2001 at 02:59:00PM +0200, Przemyslaw Wegrzyn wrote: > > Of course, I know... > > But our middle-tier will be developed using C++ , AFAIK. > > That's why I asked about c-client library... > > I would be pretty surprised if there is no c library for IMAP as well. Sure, but I'd like to check if there are any alternatives to c-client. IMHO it's poorly documented. Probably I need to dig its source code... -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Per-virtualhost bandwidth usage monitoring ?
How can I do it with Apache ? -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Zmailer issues
1. Curently I use zmailer zmailer-ssl 2.99.53pre1 (installed from deb), but I have problem with scheduler - it's dying almost everyday with segfault. Is this a known issue ? In the scheduler's log I've found: action='failed5.' what does it mean ? Also: scheduler: unlink(../queue/D/^<90>0$): No such file or directory There are many such lines, with random garbage after last / 2. I've tried to compile zmailer 2.99.55 (with patch1), but it fails. Are there any unofficial debs with version .54 or .55 ? -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: smbclient question.
> smbclient //server/share -U user%password -c "cd location\of\data;mget *" > > Then is says : > > Get file file1.txt? > > and this will then stop it from working in a script that is a cronjob. Any ideas on >how I can get > it to just get the files and not ask anything? > From man smbclient: prompt Toggle prompting for filenames during operation of the mget and mput commands. When toggled ON, the user will be prompted to con firm the transfer of each file during these com mands. When toggled OFF, all specified files will be transferred without prompting. Isn't it what you're looking for ? -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Apache - SIGSEGV but no core dumps.
Hello ! I have noticed the following lines in Apache's error.log: [Fri Jun 29 17:35:41 2001] [notice] child pid 18786 exit signal Segmentation fault (11) [Fri Jun 29 17:35:54 2001] [notice] child pid 20229 exit signal Segmentation fault (11) I've strace'd Apache - the child processes dies after finishing handling request and writing line to access log , so it isn't critical, but... I use Apache 1.3.19 + php4.0.5 from Woody in chrooted enviroment. My problem is actually I can't force Apache to drop core files. I've added ulimit -c unlimited to /etc/init.d/apache, and set CoreDumpDirectory to some world writeable dir. Nope - I can't get any core to do backtrace on it. Why ? -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apache - SIGSEGV but no core dumps.
On Sat, 30 Jun 2001, Russell Coker wrote: > > I use Apache 1.3.19 + php4.0.5 from Woody in chrooted enviroment. > > My problem is actually I can't force Apache to drop core files. > > I've added ulimit -c unlimited to /etc/init.d/apache, and set > > CoreDumpDirectory to some world writeable dir. > > Nope - I can't get any core to do backtrace on it. Why ? > > Well behaved daemons often set their own core limit to zero to avoid > wasting disk space in the case of errors. You may have to recompile > Apache with this option disabled, but that's no big deal. Debian > packages contain stripped binaries so you won't get that much information > from a core file unless you recompile with "-g" instead of "-O2". I don't think It's what Apache does - it has CoreDumpDirectory configuration directive, so I think it should drope some cores there ;) -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Apache load ballancing
I need to build a scalable webserver based on Apache. I'm not sure yet if the application requires persistent TCP connections. If so, it will require that after starting session all subsequent HTTP request in this session will come to the same machine in the cluster. This requires load ballancing in 7th layer, so LVS doesn't help much. The question is: are there any software level 7 ballancer available ? If not, which hardware box would you recommend ? -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apache load ballancing
On Tue, 10 Jul 2001, Eric Jennings wrote: > >I need to build a scalable webserver based on Apache. I'm not sure yet if > >the application requires persistent TCP connections. If so, it will > >require that after starting session all subsequent HTTP request in this > >session will come to the same machine in the cluster. This requires load > >ballancing in 7th layer, so LVS doesn't help much. > > 7th layer for sure? I'd imagine that would only need to land in the 4th Layer. > 7th layer would be custom HTTP requests to test things like an > application servers, CGI scripts, or PHP scripts (7th layer == > application layer). Nope - I need sessions opened on one of the servers to be continued on this one. This requires digging into TCP stream and extracting session IDs from cookies or URLs. And this is the 7th layer. -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apache load ballancing
On Tue, 10 Jul 2001, staf wagemakers wrote: > LVS supports persistent connections ( see the webpage for more information) > which works well in most cases. I don't need persistent HTTP connections - I need persistent connections from my web application. That's why I need each session to be continued on the same host where it started. > > The question is: are there any software level 7 ballancer available ? If > > not, which hardware box would you recommend ? > > If it is a java application you could use tomcat/mod_jk which supports > loabalancing. I know :) But this is going to be mod_perl application. Maybe we will give JSP a try (actually I'm experimennting with it), but I'm afraid about it's performance - I need to do some testing. -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apache load ballancing
On Tue, 10 Jul 2001, ARAKI Yasuhiro wrote: > Hello, > > > session will come to the same machine in the cluster. This requires load > > ballancing in 7th layer, so LVS doesn't help much. > > check libapache-mod-backhand. Someone already pointed me on it in my private mail . Looks _very_ interresting, I'm going to examine it this weekend probably (ugh, lots of other work right now :( ) It has the ability to preserv host-to-session assignment in very simple, nice way :) -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apache load ballancing
On Wed, 11 Jul 2001, staf wagemakers wrote: > > I don't need persistent HTTP connections - I need persistent connections > > from my web application. That's why I need each session to be continued on > > the same host where it started. > > That is exactly what LVS does, it even works with https, where it is > necessary that each session goes back to the same server. > > You can find more info about LVS & persistence at > http://www.linuxvirtualserver.org/persistence.html I'm just reading it right now, thanks for URL. I'm not sure if I understand it well - it assigns client to particular real server on client IP number basis ? If so, it can be quite unballanced sometimes... Imagine large masqerading networks. I need to check mod_backhand, do some benchmarks... Anybody has done any such benchmarks ? > or ask it at [EMAIL PROTECTED] if you dont believe me :-) Well , I need to check archives, I think it can help :) Thanx -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Apache bandwith...
What's the best way to limit particulat 's bandwidth usage ? I tried to use mod_throttle, but... I didn't check it deeply yet, but I guess it only can insert delay between subsequent request handling. Each request itself is handled at full bandwidth , isn't it ? So it can only limit _average_ bandwidth usage ? Is there any way to limit peak connection speed, not the average ? Say, no more than 10 connections 10KB/s each, any time... -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apache bandwidth limitation
On Fri, 13 Jul 2001, ARAKI Yasuhiro wrote: > Florian, > > I suggest you to check mod_throttle(http://www.snert.com/Software/mod_throttle/) or >mod_bandwidth(http://www.cohprog.com). > Both modules has not been debianized yet. Huh ? They are in apache packages ! czajnik@earth:~$ dpkg -L apache-common | grep band /usr/lib/apache/1.3/mod_bandwidth.so czajnik@earth:~$ dpkg -L apache-common | grep throt /usr/lib/apache/1.3/mod_throttle.so /usr/lib/apache/1.3/400mod_throttle.info czajnik@earth:~ I'm currently using mod_bandwidth, but... I'd like to find something that actually measures the bandwidth usage. With mod_bandwidth, if you set , say, 100kB/s limit, and there are 2 concurrent connections - 5kB/s (modem user) and the other that can operate at 100kB/s, then the first will work at 5kB, and the second at 50kB/s, becaues mod_bandwidth simply divides 100kB/s / 2 connections, and gives each connection 50kB/s limit. I'd like rather to give the second one 98kB/s. I've tried mod_throttle, but I can't get it work actually. -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Apache bandwidth limitation
On Fri, 13 Jul 2001, Gerard MacNeil wrote: > On Fri, 13 Jul 2001 12:30:46 +0200 (CEST), Przemyslaw Wegrzyn ><[EMAIL PROTECTED]> wrote: > > > > > > > On Fri, 13 Jul 2001, ARAKI Yasuhiro wrote: > > > > > Florian, > > > > > > I suggest you to check mod_throttle(http://www.snert.com/Software/mod_throttle/) >or mod_bandwidth(http://www.cohprog.com). > > > Both modules has not been debianized yet. > > > > Huh ? They are in apache packages ! > > > > > They are in woody/sid, not potato, iirc. Huh, right, I've forgotten I'm using woody ;) Packages in Potato are too old for me, even for server installs.. -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Strange thing about file descriptors...
I've found the following strange thing: black:~# lsof | wc -l 1826 black:~# more /proc/sys/fs/file-nr 409635138192 Why is the number of file handles in use (3513) so high, comparing to the size of lsof report ? What are all those file handles used for ? Any suggestions welcome.. -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: help with site+database
On Tue, 17 Jul 2001, Fabrizio Tivano wrote: > > Expecially if you have a dinamic PHP site, for me, MySQL and PHP, is the > better and faster solution. > If you take a look at http://www.mysql.com you can see all about. > ... another one, last day i see on a newspaper also NASA migrate to MySQL It all depends on funcionality you expect from your DB server. MySQL lacks some funcionality, look: "The FOREIGN KEY syntax in MySQL exists only for compatibility with other SQL vendors". Search the net for some feature comparision. -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Clustering mail servers - Cyrus or Courier ?
I'm looking for good solution for big, scalabale mailserver installation, for 500,000 accounts and more... As for now I think: - sharng NFS-mounted mail storage is not so good idea. - scaling can be done by partitioning mail storage between servers (with a little less avialabilty risk). However, AFAIK it can be done only with Cyrus with its IMAP Aggregator, or with qmail-ldap + Courier-IMAP... I started doing such installation with Courier-IMAP + qmail-ldap, and it's nice, expect one thing - quota support. I need _reliable_ quota system.. Tis forced me to look at Cyrus, with its good quoa support, however, I have no experience with Cyrus. Would you give some advices ? Any comparision benchmarks ? Greetings -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Clustering mail servers - Cyrus or Courier ?
On Mon, 6 Aug 2001, Jeff Waugh wrote: > > > > However, AFAIK it can be done only with Cyrus with its IMAP Aggregator, or > > with qmail-ldap + Courier-IMAP... > > You ought to check out Scalemail, which is being developed expressly for > this purpose. It is a combination of Courier POP/IMAP and postfix. Very > powerful combo. Hmmm, I can see it's in early stage of developement. Does postfix support ldap nativly ? -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Clustering mail servers - Cyrus or Courier ?
On Mon, 6 Aug 2001, Kevin J. Menard, Jr. wrote: > Hey Przemyslaw, > > > Sunday, August 05, 2001, 10:10:13 AM, you wrote: > > > PW> However, AFAIK it can be done only with Cyrus with its IMAP Aggregator, or > PW> with qmail-ldap + Courier-IMAP... > > Perdition (http://www.ca.us.vergenet.net/linux/perdition/) should allow you > to do the same thing as Cyrus murder, on other mail systems. Yes, it's a good thing. However, yet I need some mailer that's capable to work as frontend, receiving mails and forwarding to the right backend server ( LMTP would be the best if talking about Cyrus). It should support LDAP database. Currently I done it with qmail-ldap + Courier, but I'm not fully satisified. It reqiured custom patching, and still it's a little experimental. I'm afraid it's not reliable enough for production site. -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Clustering mail servers - Cyrus or Courier ?
On Mon, 6 Aug 2001, Charl Matthee wrote: > If you are willing to spend the money I can recommend you use CommuniGate > Pro (Stalker Software, www.stalker.com). It is a fast and very scalable > mail server (It does SMTP, POP3, IMAP, etc.). > > We run a free webmail sevice and our original application was homegrown. > We ran into some scalability issues (becuase the software was never meant > to cope with 10+ users). CommuniGate Pro is the answer for us. Yeah, I've looked at their web page, software looks well. Can you tell anything about its design ? -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Clustering mail servers - Cyrus or Courier ?
On Tue, 7 Aug 2001, Jeff Waugh wrote: > > > > LMTP would be the best if talking about Cyrus). > > It should support LDAP database. > > Postfix supports both of these. It is an *awesome* MTA. > You've surprised me a little. I was always thinking postfix is just another MTA, nothing more... -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Clustering mail servers - Cyrus or Courier ?
On Sat, 11 Aug 2001, Russell Coker wrote: > On Sun, 5 Aug 2001 16:10, Przemyslaw Wegrzyn wrote: > > I'm looking for good solution for big, scalabale mailserver installation, > > for 500,000 accounts and more... > > > > As for now I think: > > - sharng NFS-mounted mail storage is not so good idea. > > Why not? If you use Maildir storage then it's NFS safe. If you use a NetAPP > Filer, a VA NAS device (or any machine running 2.2.x kernel with LVM, Ext3, > and appropriate NFS kernel-server patches as the VA NAS devices did), or some > similarly powerful NFS server then performance should be more than adequate. Hmmm, I'm really afraid about NFS, I've heared that Linux NFS implementation is broken, many times. Can anyone explain it a bit ? > For Ext3 you need to keep the number of entries in a directory down to <1000 > for best performance, but most users don't have that many messages in their > mail folder. So you just have to do suitable directory hashing to make sure > you don't have too many users in a directory (all mail storage software has > support for this). Yeah, I do such hushing on every bigger mailserver, it's a common solution. Currently I'm testing qmail-ldap + maildrop + Courier-IMAP. Works well, but it's not HA solution :| > Then you can have multiple mail servers running at the same time for > increased reliability. This is much better than the Perdition approach of > increasing the number of points of failure. Having an NFS file server as a > single point of failure is as good as it gets, making an NFS file server > reliable is much easier than making a mail server reliable... Running multiple mail servers over NFS will actually makie it HA-like system. But is it good solution for high load ? Using perdition-like solution to split mailspool between separate NFS servers, and using multiple mailservers for each NFS backend would probably build good scalable solution. -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
High performance filesystems ?
Is there any page comparing the performance/features of ReiserFS , JFS, XFS ? Which one would you suggest for maildirs on heavy loaded server ?
Re: High performance filesystems ?
On Tue, 6 Mar 2001, Russell Coker wrote: > On Monday 05 March 2001 18:33, Przemyslaw Wegrzyn wrote: > > Is there any page comparing the performance/features of ReiserFS , JFS, > > XFS ? > > > > Which one would you suggest for maildirs on heavy loaded server ? > > XFS and JFS are not released so I recommend not using them on important > servers (IE mail servers). Yes, I've already figured it out. It's all at beta-tests stage.. > ReiserFS had some fsck problems which should be fixed now, but I wouldn't > want to bet my business on it just yet. What problems exactly ? > I am likely to be installing a reasonable sized mail server in 2-3 months > time. I think it's reasonably likely that ReiserFSCK will have been proved > reliable enough by then. If not I'll get a NetApp. Do you mean NetApp filers ? > Also try out my Postal program for benchmarking mail servers. Thanx :)
Sybase ASE 11.0.3.3
Has anyone here experience with Sybase ASE 11.0.3.3 on Linux ? I'm wondering, if it is a good choice for production enviroment... It has a little strange license: "You are allowed to install and use the Software for free as long as you operate the Software at all times only with the Open Source - Linux and various BSD - operating systems running natively on your hardware system. " What does it mean _operate_the_Software_ ? ;) What about multi-tier applications, when connections to Sybase are made by linux-based-middle-tier only ? Is it allowed ? TIA PS - I know I should post this question directly to Sybase, but I'd like to know about your experiences with it...
Re: Sybase ASE 11.0.3.3 (fwd)
On Tue, 13 Mar 2001, J.H.M. Dassen (Ray) wrote: > On Tue, Mar 13, 2001 at 12:27:32 -0800, Eric Jennings wrote: > > I can't comment on the stability of the product within a production > > environment, but I was very pleased with the multi-level transaction > > system, > > >From what I read today, InterBase (which is free nowadays; see e.g. > http://www.ibphoenix.com/) has multi-level transactions. I'm reading it just now (little surprised), downloading sources on second console. Is it really free for production enviroments ? I'm thinking of two SQL servers right now - PostgreSQL or InterBase. I decided not to use Sybase for now (maybe I will for some future projects ?). I'll probably need BLOB support, which is a little weird in current release of PostgreSQL...
Re: Sybase ASE 11.0.3.3
On Tue, 13 Mar 2001, Eric Jennings wrote: > If only they'd open up the source completely, I'd move all of my web > app development immediately to it from PostgreSQL/MySQL. Me too. But I don't think it's possible in the near future. I would be very surprised, if they decided to do so... They don't even want to release free ODBC drivers :(
Re: Sybase ASE 11.0.3.3
On Tue, 13 Mar 2001, Eric Jennings wrote: > >From what I read today, InterBase (which is free nowadays; see e.g. > >http://www.ibphoenix.com/) has multi-level transactions. > > > Thanks for the info... I'll have to check that out. > > Anybody else have positive/negative experiences with InterBase? And, BTW, anybody already moved it into DEBs ?
Re: load balancer
On 15 Mar 2001, Fraser Campbell wrote: > Allen Ahoffman <[EMAIL PROTECTED]> writes: > > > Is there a distribution that will cheaply replace a load balancer? > > e.g. > > for web servers. > > LVS, ...? > > man ipmasqadm ... you'll see this: > >Redirect all web traffic to internals hostA and hostB, >where hostB will serve 2 times hostA connections. Forward >rules already masq internal hosts to outside (typical). > > ipchains -I input -p tcp -y -d yours.com/32 80 -m 1 > ipmasqadm mfw -I -m 1 -r hostA 80 -p 10 > ipmasqadm mfw -I -m 1 -r hostB 80 -p 20 > > Voila, load balancer ... any Linux distribution can do this. :)) And what about sessions ?
Re: load balancer
On Fri, 16 Mar 2001, Marcel Hicking wrote: > Przemyslaw Wegrzyn <[EMAIL PROTECTED]> 16 Mar 2001, at > 10:17: > > On 15 Mar 2001, Fraser Campbell wrote: > > > Allen Ahoffman <[EMAIL PROTECTED]> writes: > > > > Is there a distribution that will cheaply replace a load > balancer? > > > > e.g. for web servers. LVS, ...? > > > > > > man ipmasqadm ... you'll see this: > > > > > >Redirect all web traffic to internals hostA and hostB, > > >where hostB will serve 2 times hostA connections. Forward > > >rules already masq internal hosts to outside (typical). > > > > > > ipchains -I input -p tcp -y -d yours.com/32 80 -m 1 > > > ipmasqadm mfw -I -m 1 -r hostA 80 -p 10 ipmasqadm mfw > > > -I -m 1 -r hostB 80 -p 20 > > > > > > Voila, load balancer ... any Linux distribution can do this. > > > > :)) And what about sessions ? > > As session data is usually stored in a database, Not always.. > have the sql server on a separate machine (you > would anyway if you have enough traffic to need > a load balancer, wouldn't you?) Generaly, right... Greetings
Re: record a mac address of a networkcard
On Fri, 16 Mar 2001, Simon Tennant wrote: > I run arpwatch. It watches for new arp requests and mails me the > following: > > Subject: new station (w192.z064003150.sjc-ca.dsl.cnc.net) >Date: Thu, 08 Mar 2001 16:49:59 -0800 >From: [EMAIL PROTECTED] (Arpwatch) > To: [EMAIL PROTECTED] > > hostname: w192.z064003150.sjc-ca.dsl.cnc.net > ip address: 64.3.150.192 > ethernet address: 0:10:67:0:3b:92 > ethernet vendor: Redback Networks, Inc. >timestamp: Thursday, March 8, 2001 16:49:59 -0800 > At our campus we're usuing static ARP mapping, with table generated from our users database. But still one can easily change his MAC :(
Re: webmail
On Tue, 17 Apr 2001, John Ackermann N8UR wrote: > > I want to recommend you IMP. You can set IMAP server on firewall and > > IMP/Horde. > > I've been playing with SilkyMail (http://www.cyrusoft.com/silkymail), which > is an enhanced version of IMP. It's under the GPL, but Cyrusoft are > offering commercial support packages for it. I'm very impressed, and > (under the right circumstances, which mine weren't) it's a snap to build as > all the tools (apache, php, etc.) are included in the tarball and a single > make command builds the whole environment. Then you untar the php files > into the htdocs directory, run a config script, and you're in business. > How about attachments ? Some time ago I've checked several webmail programs. And all of theme were doing one thing that make me cry - _load_ _whole_attachment_to_RAM_ before reciving / sending. I think it should be done by processing attachments part-by-part. -=Czaj-nick=-
Lilo problem
I have problems booting my system from new HD. I've got: - ABIT KT7 - Seagate connected to first channel (as /dev/hda) - CD-ROM as /dev/hdc - new disk (DTLA 307030) connected to HPT370 as /dev/hde I want to boot my system form /dev/hde, ( I want to be able to take /dev/hda away sometimes, and still be able to boot my system, of course after rerunning lilo). If new disk is the only one being seen by BIOS ( Seagate is disabled in bios setup) it works fine - new disk is 0x80 in BIOS. On the other hand when 2 disk are available, LILO stops after "LI". Of course I rerun lilo on every change. I guess it tries to access not the right drive. What will be the bios drive number of my new disk, when both are turned on ? 0x81 ? TIA -=Czaj-nick=-
Re: Lilo problem
On Wed, 25 Apr 2001, Russell Coker wrote: > On Friday 20 April 2001 11:40, Przemyslaw Wegrzyn wrote: > > I have problems booting my system from new HD. > > > > I've got: > > - ABIT KT7 > > - Seagate connected to first channel (as /dev/hda) > > - CD-ROM as /dev/hdc > > - new disk (DTLA 307030) connected to HPT370 as /dev/hde > > > > I want to boot my system form /dev/hde, ( I want to be able to take > > /dev/hda away sometimes, and still be able to boot my system, of course > > after rerunning lilo). > > > > If new disk is the only one being seen by BIOS ( Seagate is disabled in > > bios setup) it works fine - new disk is 0x80 in BIOS. > > On the other hand when 2 disk are available, LILO stops after "LI". > > Of course I rerun lilo on every change. I guess it tries to access not the > > right drive. > > What will be the bios drive number of my new disk, when both are turned on > > ? 0x81 ? > > 0x81 sounds reasonable, but really it could be anything. The only thing to > do is to try telling LILO to use 0x81 and see if that makes it boot! It's already solved (by me :) ) - nobody proposed to put there 0x80. It works ! Actually, if I setup my BIOS to boot from RAID first , dev/hde becomes 0x80 and /dev/hda becomes 0x81 ! I guess (not checked) that drive selected to boot from always becomes 0x80 - that's why DOS knows how to load itself - it loads itself from 0x80 always ! If U boot from first drive there's no problem , because lilo figures it out as 0x80. disk = /dev/hde bios = 0x80 Voilla ! Well, now it's pretty trivial for me ;) but I've spent much time figuring it out. Maybe someone should include it in some faq ? Greetings -=Czaj-nick=-
Kernels 2.4.x for ISP ?
I'm just setting up new box - large mail server, dual Pentium, STL2 motherboard, Mylex AccleRAID. I'd like to try with kernel 2.4.x + ResierFS, but I worry a little about it's maturity. Is it really stable now ? Can anyone here write about his/her experiences with 2.4.x ? Does 2.4.x kernels support AcceleRAID 352 well ? Greetings -=Czaj-nik=-
Re: Kernels 2.4.x for ISP ?
On Fri, 18 May 2001, Tech Support wrote: > I can't comment on the ResierFS or AcceleRAID support but I can > comment on the stability of the 2.4.2 kernel. > We have been running 2.4.2 on our Web servers which also do DNS, mail, > mysql and FTP for 50 days now, without a problem. The 2.4.x kernel is > definitely faster then 2.2.x (no I didn't benchmark it :-) and it is > perfectly stable. > We have also run 2.4.3 on a test box and have subject that to some > ugly situations without harm. What configuration ? I want to run it on dual processor machine. > However, be aware of the security hole (security cavern!) in iptables, > so don't use the firewalling features. The problem is supposed to be > fixed in 2.4.5. Thanx. 2.4.5 fixes Adaptex AIC7xxx drivers also. > Also be sure to read the kernel source as you must upgrade modutils > and few other packages from woody to get 2.4.x to work. I know, I'm running 2.4.3 on my desktop comps. TIA -=Czaj-nick=-
Re: Kernels 2.4.x for ISP ?
On Fri, 18 May 2001, Tech Support wrote: > Przemyslaw Wegrzyn wrote: > > What configuration ? > > I want to run it on dual processor machine. > Our web servers are single processor Pentium III on a Asus > motherboard. I also run 2.4.2 with SMP on a couple of Dell Poweredge > duel 1Ghz processor, 2Gb memory with aacRAID for a client. These servers > are running large Oracle 8i databases so I'm not sure their performance > translates to Web service but the SMP support is fine and stability has > been equal to my non-SMP machines. The kernels are patched to use the > aacRAID. What patch ? Isn't DAC960 module already included in kernel ? -=Czaj-nick=-
Re: Kernels 2.4.x for ISP ?
On Fri, 18 May 2001, Tech Support wrote: > > What patch ? Isn't DAC960 module already included in kernel ? > Yes Mylex DAC960 support is in the kernel but this isn't the > controller that these machines used. Dell uses (or at least in these > machines) a controller made by Adaptec which is alternately called > PercRAID and aacRAID. The patch and info is at http://domsch.com/linux/ Aaaah, thanx a lot ! We have a choice Adaptec 2100S or Mylex AcceleRAID 170. We've decided to take the second - because of the kernel support. Adaptec seems to ignore linux - on their web page I've found drivers for 2.2.14 :|
Re: Kernels 2.4.x for ISP ?
On Fri, 18 May 2001, Krzysztof Mazurczyk wrote: > And third problem may be with dependencies of packets. Kernel suggests use > 2.4.x > of modutils, modutils depends on libc6 2.2.x and newer libc6 depends on > libdb2 2.7.7 > or similar. > But when I listalled libdb2 taken form woody lastly, syslogd didn't start. > Yesterday I send > question to debian-user list with this problem but I've no answer as yet. > It may be > nessesary to use 2.4.x kernel without modules. Well, there are packages available allowing use of 2.4.x with potato. I already use 2 potato hosts with 2.4.3 kernel. Works very fine, with no stability problems. But I wonder if the same kernel will be stale on different hardware - SMP, SCSI and so on. I've just subscribed linux-kernel mailinglist to see what's going on there... Greetings -=Czaj-nick=-
Re: Apache suEXEC problem
On Sat, 19 May 2001, Hector Castillo wrote: > It's impossible to move the pages to the original directory (there are > a lot users and a lot pages), and the suEXEC is necessary. > My question is: żdoes anybody know how to recompile the Debian source > of apache-ssl and which files would be necessary to change? Look at debian/rules of apache's source package. There's option to 'configure', specifing allowed dirs. -=Czaj-nick=-
Tomcat problem
I get the following exception starting tomcat. What can cause this error ? Tomcat starts and seems to work (at least exampels work properly) TIA -=Czaj-nick=- Starting tomcat. Check logs/tomcat.log for error messages org.apache.xerces.validators.datatype.InvalidDatatypeFacetException: org/apache/xerces/utils/regex/ParserForXMLSchema at org.apache.xerces.validators.datatype.DatatypeValidatorFactoryImpl.createDatatypeValidator(DatatypeValidatorFactoryImpl.java) at org.apache.xerces.validators.datatype.DatatypeValidatorFactoryImpl.createDatatypeValidator(DatatypeValidatorFactoryImpl.java) at org.apache.xerces.validators.datatype.DatatypeValidatorFactoryImpl.initializeDTDRegistry(DatatypeValidatorFactoryImpl.java) at org.apache.xerces.validators.common.XMLValidator.initDataTypeValidators(XMLValidator.java) at org.apache.xerces.validators.common.XMLValidator.setValidationEnabled(XMLValidator.java) at org.apache.xerces.framework.XMLParser.setValidation(XMLParser.java) at org.apache.xerces.framework.XMLParser.setFeature(XMLParser.java) at org.apache.xerces.parsers.SAXParser.setFeature(SAXParser.java) at org.apache.xerces.jaxp.SAXParserImpl.(SAXParserImpl.java) at org.apache.xerces.jaxp.SAXParserFactoryImpl.newSAXParser(SAXParserFactoryImpl.java) at org.apache.tomcat.util.xml.XmlMapper.readXml(XmlMapper.java) at org.apache.tomcat.context.WebXmlReader.processWebXmlFile(WebXmlReader.java) at org.apache.tomcat.context.WebXmlReader.contextInit(WebXmlReader.java) at org.apache.tomcat.core.ContextManager.initContext(ContextManager.java) at org.apache.tomcat.core.ContextManager.init(ContextManager.java) at org.apache.tomcat.startup.Tomcat.execute(Tomcat.java) at org.apache.tomcat.startup.Tomcat.main(Tomcat.java)
kernel 2.2.19 limitations.
What is the maximum file size in 2.2.19 kernel ? Still 2GB ? Is it ext2 limitations , or ? Will ReiserFS have the same limit using 2.2.19 kernel ? -=Czaj-nick=-
Re: kernel 2.2.19 limitations.
On Sun, 27 May 2001, Peter Billson wrote: > Whoops... helps if I post *to* the list too! > > > Yes the limit is still the usual 2Gb. The limit is actually with ext2, i > > believe, although I'm not sure. > > The limit is in the kernel, not the ext2 file system, otherwise 2.4.x > wouldn't be able to support >2Gb file either. There are patches about > for adding LFS (large file system) support. I had compiled a 2.2.18 > kernel after patching with the LFS patch borrowed from RedHat's 6.2ee > (Enterprise Edition) source. > > But why not just run 2.4.x? Hmm, I'm building some so called "very important server". I'm not sure if 2.4.x is stable enough. It will run Apache, biiig PostgreSQL database, all on big RAID. Anyone has experience with 2.4.x + SMP (2 x PIII) + Mylex AcceleRAID ? After reading some newsgroups I believe 2.4.x kernels works pretty stable or does totaly weird things, all depends of the hardware configuration. Well, it seems some drivers are not yet stable enough... -=Czaj-nick=-
Ping - what the hell ?
Look at this: [EMAIL PROTECTED]:~$ ping 156.17.209.1 PING 156.17.209.1 (156.17.209.1): 56 data bytes 64 bytes from 156.17.209.1: icmp_seq=0 ttl=247 time=5427.7 ms 64 bytes from 156.17.209.1: icmp_seq=1 ttl=247 time=23.2 ms 64 bytes from 156.17.209.1: icmp_seq=2 ttl=247 time=429492829.5 ms ^^^ 64 bytes from 156.17.209.1: icmp_seq=3 ttl=247 time=429492907.1 ms 64 bytes from 156.17.209.1: icmp_seq=4 ttl=247 time=16.5 ms 64 bytes from 156.17.209.1: icmp_seq=5 ttl=247 time=21.4 ms What the hell can it be ?!? Every ping comes back in few milisecounds... We've recently added second CPU to this machine, can it (failed CPU) be the reason ? Overall stability is good... -=Czaj-nick=-
Re: Ping - what the hell ?
On Wed, 30 May 2001, Ken Seefried wrote: > > I think that you can get this if you have an MP kernel compiled without > "Enhanced Real Time Clock" support. The default clock driver apparently > isn't MP-safe. Thx, I'll check it tomorrow :).
Zmailer question.
Hmm, I use qmail almost everywhere, but friend asked me for some help with zmailer. Almost everything is done, but also we need ability to send mails to all our customers - all accounts in all virt domains. What's the easiest solution with zmailer ? -=Czaj-nick=-
Re: Ping - what the hell ?
On Wed, 30 May 2001 [EMAIL PROTECTED] wrote: > Aren't you traveling several rfc1149 links? > > http://www.blug.linux.no/rfc1149/pinglogg.txt ?!? What do U mean ? isn't the ping time measured by storing system time the ICMP ECHO was sent, and comparng it to the system time the reply arrived ? I get it even when pinging my Cisco 1601 router, which is connected directly (I don't care 3com SSII 1100 now) -=Czaj-nick=- > On Wed, May 30, 2001 at 03:24:39PM +0200, Przemyslaw Wegrzyn wrote: > > > > Look at this: > > > > [EMAIL PROTECTED]:~$ ping 156.17.209.1 > > PING 156.17.209.1 (156.17.209.1): 56 data bytes > > 64 bytes from 156.17.209.1: icmp_seq=0 ttl=247 time=5427.7 ms > > 64 bytes from 156.17.209.1: icmp_seq=1 ttl=247 time=23.2 ms > > 64 bytes from 156.17.209.1: icmp_seq=2 ttl=247 time=429492829.5 ms > > ^^^ > > 64 bytes from 156.17.209.1: icmp_seq=3 ttl=247 time=429492907.1 ms > > 64 bytes from 156.17.209.1: icmp_seq=4 ttl=247 time=16.5 ms > > 64 bytes from 156.17.209.1: icmp_seq=5 ttl=247 time=21.4 ms > > > > What the hell can it be ?!? Every ping comes back in few milisecounds... > > > > We've recently added second CPU to this machine, can it (failed CPU) be > > the reason ? Overall stability is good... > > > > -=Czaj-nick=- > > > > > > > > -- > > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] > > -- > HoraPe > --- > Horacio J. Peńa > [EMAIL PROTECTED] > [EMAIL PROTECTED] > [EMAIL PROTECTED] > [EMAIL PROTECTED] > > > -- > To UNSUBSCRIBE, email to [EMAIL PROTECTED] > with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED] >
Re: Ping - what the hell ?
On Wed, 30 May 2001, Nathan E Norman wrote: > On Wed, May 30, 2001 at 08:47:38PM +0200, Przemyslaw Wegrzyn wrote: > > > > > > On Wed, 30 May 2001 [EMAIL PROTECTED] wrote: > > > > > Aren't you traveling several rfc1149 links? > > > > > > http://www.blug.linux.no/rfc1149/pinglogg.txt > > > > ?!? What do U mean ? > > It was a joke ... rfc 1149 is IP over avian carriers (birds) Ooooh, yeah, I didn't remember the rfc number, but I know this protocol : It's nice :) Anyway, my problem seems to be hardware: [EMAIL PROTECTED]:~$ more /proc/misc Segmentation fault [EMAIL PROTECTED]:~$ -=Czaj-nick=-
Re: Zmailer question.
On Thu, 31 May 2001, Russell Coker wrote: > On Wednesday 30 May 2001 19:49, Przemyslaw Wegrzyn wrote: > > Hmm, I use qmail almost everywhere, but friend asked me for some help with > > zmailer. Almost everything is done, but also we need ability to send mails > > to all our customers - all accounts in all virt domains. What's the > > easiest solution with zmailer ? > > If you are using Maildir storage then my maildir-bulletin package might do > what you need. Nope. It makes no problem with Maildirs, but here it's still on mailboxes (will change soon). I see generaly (on every MTA) 2 ways - alias to all (but there's a risk of abuse), and second way - script simply sending to everyone in the prepared list. -=Czaj-nick=-
Re: Ping - what the hell ?
On Sat, 2 Jun 2001, Craig Sanders wrote: > On Wed, May 30, 2001 at 09:41:54PM +0200, Przemyslaw Wegrzyn wrote: > > Anyway, my problem seems to be hardware: > > > > [EMAIL PROTECTED]:~$ more /proc/misc > > Segmentation fault > > [EMAIL PROTECTED]:~$ > > some possible causes: > > 1. bad memory - most likely. > > 2. bad swap partition (or bad disk controller causing the swap partition to > not work) > > 3. other bad hardware > > 4. bad libc6 or other library - not very likely. > It' solved, there were 2 reasons. Core dumps - hmmm, our admin borken the kernel by incorrectly patching it. Ping times - some stupid guy inserted two different CPUs PII 400 and 450. It's a miracle it was working all together... -=Czaj-nick=-
Re: Ping - what the hell ?
On Sun, 3 Jun 2001, Chris Wagner wrote: > I'm sorry, but ROFLMAO!!! It's sad and (sometimes) funny, that I have to work with those people ;) We are just changing our admin. He was a real mistake :| Now it's all funny for me, but It costed me time, lots of time... > >It' solved, there were 2 reasons. > > Core dumps - hmmm, our admin borken the kernel by incorrectly patching > >it. > > Ping times - some stupid guy inserted two different CPUs PII 400 and 450. > > It's a miracle it was working all together... -=Czaj-nick=-
Webmail - considerations...
We are going to develop web-mail system, that's capable of handling relatively high loads. I know, there are many open source web-mail systems , but they doesn't satisfy me. Almost every falls into one of two cateogries: php based, using imap; perl cgi based, using IMAP or direct filesystem access... I'd like someone experienced with such systems help me with the following: 1. Almost every available webmail system uses the following way of handling (rreceiving, in this example) attachements: load the whole message body from IMAP server or message file, decode it and send to the client. The _whole_ attachement gets loaded into server's RAM. Isn't it waste of resources/ killing the server ? I think it should read/decode/send the attachement on a line-by-line (or part-by-part generaly) manner. Am I right ? 2. Which one is better - accessing maildirs directly, or using IMAP ? I can see that IMAP seems to be more scalable / universal... Maildirs probably can be much faster to work with directly, but probably less secure... Any other pros/contras ? 3. I'm going to develop the front-end using Apache::ASP or php, not decided yet, and access the mails through the middle-tier daemon. The question is - is it a good way to use persistent IMAP connections ? If so, there will be no overhead of authentication on every operation, but there can be many open IMAP connections to the local imap server (probably Courier-IMAP) at the same time. Which strategy is better ? 4. Are there any libraries similar to c-client, maybe some C++ ones ? 5. Does c-client library allow to retrieve the message body (attachements) partialy (see 1) ? I've seen some /tmp access in its source code - does it dowload whole message to /tmp ? Hmm, that's all for now... TIA -=Czaj-nick=-
Perl 5.6 & Potato
Is it possible to run perl 5.6 iwth Potato without upgrading 50% of the system ? I've tried to build debs from Woody, but they fails during ndbm tests... -=Czaj-nick=-
Re: Webmail - considerations...
On Tue, 12 Jun 2001, Russell Coker wrote: > Another thing, do the webmail systems use directory notifications? With > directory notifications the application can be informed by the kernel > when new files are created. The IMAP protocol supports announcing new > mail to the client. If the mail store and the webmail system are on > separate machines then IMAP is the only way to avoid polling for new mail > (AFAIK directory notifications don't work over NFS). Thanx, I didn't know this feature... Acctually, I'm not so familiar with IMAP protocol yet... > So now you're planning to write your own webmail program? Yes, we need some nonusuall features... > I suggest that when a user logs in they get an IMAP connection. An > option to have the IMAP connection timeout and be closed before the > webmail system times out (webmail timeout on no activity should be long > like 30 minutes while IMAP timeout should be 10 minutes at most) would be > handy, but isn't required. But generally IMAP connections should stay > open for the duration of the session. That's as I see it, thanx.. All the actual logic will be implemented in the middle tier daemon(s). -=Czaj-nick=-
Re: Webmail - considerations...
On Tue, 12 Jun 2001, Jeremy Lunn wrote: > On Tue, Jun 12, 2001 at 01:26:09PM +0200, Przemyslaw Wegrzyn wrote: > > Thanx, I didn't know this feature... Acctually, I'm not so familiar with > > IMAP protocol yet... > > Depending what language you use you won't need to know the IMAP > protocol. Things like perl and php have IMAP modules. Of course, I know... But our middle-tier will be developed using C++ , AFAIK. That's why I asked about c-client library... -=Czaj-nick=-
Re: Webmail - considerations...
On Tue, 12 Jun 2001, Jeremy Lunn wrote: > On Tue, Jun 12, 2001 at 02:59:00PM +0200, Przemyslaw Wegrzyn wrote: > > Of course, I know... > > But our middle-tier will be developed using C++ , AFAIK. > > That's why I asked about c-client library... > > I would be pretty surprised if there is no c library for IMAP as well. Sure, but I'd like to check if there are any alternatives to c-client. IMHO it's poorly documented. Probably I need to dig its source code... -=Czaj-nick=-
Per-virtualhost bandwidth usage monitoring ?
How can I do it with Apache ? -=Czaj-nick=-
Zmailer issues
1. Curently I use zmailer zmailer-ssl 2.99.53pre1 (installed from deb), but I have problem with scheduler - it's dying almost everyday with segfault. Is this a known issue ? In the scheduler's log I've found: action='failed5.' what does it mean ? Also: scheduler: unlink(../queue/D/^<90>0$): No such file or directory There are many such lines, with random garbage after last / 2. I've tried to compile zmailer 2.99.55 (with patch1), but it fails. Are there any unofficial debs with version .54 or .55 ? -=Czaj-nick=-
Re: smbclient question.
> smbclient //server/share -U user%password -c "cd location\of\data;mget *" > > Then is says : > > Get file file1.txt? > > and this will then stop it from working in a script that is a cronjob. Any > ideas on how I can get > it to just get the files and not ask anything? > From man smbclient: prompt Toggle prompting for filenames during operation of the mget and mput commands. When toggled ON, the user will be prompted to con firm the transfer of each file during these com mands. When toggled OFF, all specified files will be transferred without prompting. Isn't it what you're looking for ? -=Czaj-nick=-
Apache - SIGSEGV but no core dumps.
Hello ! I have noticed the following lines in Apache's error.log: [Fri Jun 29 17:35:41 2001] [notice] child pid 18786 exit signal Segmentation fault (11) [Fri Jun 29 17:35:54 2001] [notice] child pid 20229 exit signal Segmentation fault (11) I've strace'd Apache - the child processes dies after finishing handling request and writing line to access log , so it isn't critical, but... I use Apache 1.3.19 + php4.0.5 from Woody in chrooted enviroment. My problem is actually I can't force Apache to drop core files. I've added ulimit -c unlimited to /etc/init.d/apache, and set CoreDumpDirectory to some world writeable dir. Nope - I can't get any core to do backtrace on it. Why ? -=Czaj-nick=-
Re: Apache - SIGSEGV but no core dumps.
On Sat, 30 Jun 2001, Russell Coker wrote: > > I use Apache 1.3.19 + php4.0.5 from Woody in chrooted enviroment. > > My problem is actually I can't force Apache to drop core files. > > I've added ulimit -c unlimited to /etc/init.d/apache, and set > > CoreDumpDirectory to some world writeable dir. > > Nope - I can't get any core to do backtrace on it. Why ? > > Well behaved daemons often set their own core limit to zero to avoid > wasting disk space in the case of errors. You may have to recompile > Apache with this option disabled, but that's no big deal. Debian > packages contain stripped binaries so you won't get that much information > from a core file unless you recompile with "-g" instead of "-O2". I don't think It's what Apache does - it has CoreDumpDirectory configuration directive, so I think it should drope some cores there ;) -=Czaj-nick=-
Apache load ballancing
I need to build a scalable webserver based on Apache. I'm not sure yet if the application requires persistent TCP connections. If so, it will require that after starting session all subsequent HTTP request in this session will come to the same machine in the cluster. This requires load ballancing in 7th layer, so LVS doesn't help much. The question is: are there any software level 7 ballancer available ? If not, which hardware box would you recommend ? -=Czaj-nick=-
Re: Apache load ballancing
On Tue, 10 Jul 2001, Eric Jennings wrote: > >I need to build a scalable webserver based on Apache. I'm not sure yet if > >the application requires persistent TCP connections. If so, it will > >require that after starting session all subsequent HTTP request in this > >session will come to the same machine in the cluster. This requires load > >ballancing in 7th layer, so LVS doesn't help much. > > 7th layer for sure? I'd imagine that would only need to land in the 4th > Layer. > 7th layer would be custom HTTP requests to test things like an > application servers, CGI scripts, or PHP scripts (7th layer == > application layer). Nope - I need sessions opened on one of the servers to be continued on this one. This requires digging into TCP stream and extracting session IDs from cookies or URLs. And this is the 7th layer. -=Czaj-nick=-
Re: Apache load ballancing
On Tue, 10 Jul 2001, staf wagemakers wrote: > LVS supports persistent connections ( see the webpage for more information) > which works well in most cases. I don't need persistent HTTP connections - I need persistent connections from my web application. That's why I need each session to be continued on the same host where it started. > > The question is: are there any software level 7 ballancer available ? If > > not, which hardware box would you recommend ? > > If it is a java application you could use tomcat/mod_jk which supports > loabalancing. I know :) But this is going to be mod_perl application. Maybe we will give JSP a try (actually I'm experimennting with it), but I'm afraid about it's performance - I need to do some testing. -=Czaj-nick=-
Re: Apache load ballancing
On Tue, 10 Jul 2001, ARAKI Yasuhiro wrote: > Hello, > > > session will come to the same machine in the cluster. This requires load > > ballancing in 7th layer, so LVS doesn't help much. > > check libapache-mod-backhand. Someone already pointed me on it in my private mail . Looks _very_ interresting, I'm going to examine it this weekend probably (ugh, lots of other work right now :( ) It has the ability to preserv host-to-session assignment in very simple, nice way :) -=Czaj-nick=-
Re: Apache load ballancing
On Wed, 11 Jul 2001, staf wagemakers wrote: > > I don't need persistent HTTP connections - I need persistent connections > > from my web application. That's why I need each session to be continued on > > the same host where it started. > > That is exactly what LVS does, it even works with https, where it is > necessary that each session goes back to the same server. > > You can find more info about LVS & persistence at > http://www.linuxvirtualserver.org/persistence.html I'm just reading it right now, thanks for URL. I'm not sure if I understand it well - it assigns client to particular real server on client IP number basis ? If so, it can be quite unballanced sometimes... Imagine large masqerading networks. I need to check mod_backhand, do some benchmarks... Anybody has done any such benchmarks ? > or ask it at [EMAIL PROTECTED] if you dont believe me :-) Well , I need to check archives, I think it can help :) Thanx -=Czaj-nick=-
Apache bandwith...
What's the best way to limit particulat 's bandwidth usage ? I tried to use mod_throttle, but... I didn't check it deeply yet, but I guess it only can insert delay between subsequent request handling. Each request itself is handled at full bandwidth , isn't it ? So it can only limit _average_ bandwidth usage ? Is there any way to limit peak connection speed, not the average ? Say, no more than 10 connections 10KB/s each, any time... -=Czaj-nick=-
High performance filesystems ?
Is there any page comparing the performance/features of ReiserFS , JFS, XFS ? Which one would you suggest for maildirs on heavy loaded server ? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: High performance filesystems ?
On Tue, 6 Mar 2001, Russell Coker wrote: > On Monday 05 March 2001 18:33, Przemyslaw Wegrzyn wrote: > > Is there any page comparing the performance/features of ReiserFS , JFS, > > XFS ? > > > > Which one would you suggest for maildirs on heavy loaded server ? > > XFS and JFS are not released so I recommend not using them on important > servers (IE mail servers). Yes, I've already figured it out. It's all at beta-tests stage.. > ReiserFS had some fsck problems which should be fixed now, but I wouldn't > want to bet my business on it just yet. What problems exactly ? > I am likely to be installing a reasonable sized mail server in 2-3 months > time. I think it's reasonably likely that ReiserFSCK will have been proved > reliable enough by then. If not I'll get a NetApp. Do you mean NetApp filers ? > Also try out my Postal program for benchmarking mail servers. Thanx :) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Sybase ASE 11.0.3.3
Has anyone here experience with Sybase ASE 11.0.3.3 on Linux ? I'm wondering, if it is a good choice for production enviroment... It has a little strange license: "You are allowed to install and use the Software for free as long as you operate the Software at all times only with the Open Source - Linux and various BSD - operating systems running natively on your hardware system. " What does it mean _operate_the_Software_ ? ;) What about multi-tier applications, when connections to Sybase are made by linux-based-middle-tier only ? Is it allowed ? TIA PS - I know I should post this question directly to Sybase, but I'd like to know about your experiences with it... -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Sybase ASE 11.0.3.3 (fwd)
On Tue, 13 Mar 2001, J.H.M. Dassen (Ray) wrote: > On Tue, Mar 13, 2001 at 12:27:32 -0800, Eric Jennings wrote: > > I can't comment on the stability of the product within a production > > environment, but I was very pleased with the multi-level transaction > > system, > > >From what I read today, InterBase (which is free nowadays; see e.g. > http://www.ibphoenix.com/) has multi-level transactions. I'm reading it just now (little surprised), downloading sources on second console. Is it really free for production enviroments ? I'm thinking of two SQL servers right now - PostgreSQL or InterBase. I decided not to use Sybase for now (maybe I will for some future projects ?). I'll probably need BLOB support, which is a little weird in current release of PostgreSQL... -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Sybase ASE 11.0.3.3
On Tue, 13 Mar 2001, Eric Jennings wrote: > If only they'd open up the source completely, I'd move all of my web > app development immediately to it from PostgreSQL/MySQL. Me too. But I don't think it's possible in the near future. I would be very surprised, if they decided to do so... They don't even want to release free ODBC drivers :( -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Sybase ASE 11.0.3.3
On Tue, 13 Mar 2001, Eric Jennings wrote: > >From what I read today, InterBase (which is free nowadays; see e.g. > >http://www.ibphoenix.com/) has multi-level transactions. > > > Thanks for the info... I'll have to check that out. > > Anybody else have positive/negative experiences with InterBase? And, BTW, anybody already moved it into DEBs ? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: load balancer
On 15 Mar 2001, Fraser Campbell wrote: > Allen Ahoffman <[EMAIL PROTECTED]> writes: > > > Is there a distribution that will cheaply replace a load balancer? > > e.g. > > for web servers. > > LVS, ...? > > man ipmasqadm ... you'll see this: > >Redirect all web traffic to internals hostA and hostB, >where hostB will serve 2 times hostA connections. Forward >rules already masq internal hosts to outside (typical). > > ipchains -I input -p tcp -y -d yours.com/32 80 -m 1 > ipmasqadm mfw -I -m 1 -r hostA 80 -p 10 > ipmasqadm mfw -I -m 1 -r hostB 80 -p 20 > > Voila, load balancer ... any Linux distribution can do this. :)) And what about sessions ? -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: load balancer
On Fri, 16 Mar 2001, Marcel Hicking wrote: > Przemyslaw Wegrzyn <[EMAIL PROTECTED]> 16 Mar 2001, at > 10:17: > > On 15 Mar 2001, Fraser Campbell wrote: > > > Allen Ahoffman <[EMAIL PROTECTED]> writes: > > > > Is there a distribution that will cheaply replace a load > balancer? > > > > e.g. for web servers. LVS, ...? > > > > > > man ipmasqadm ... you'll see this: > > > > > >Redirect all web traffic to internals hostA and hostB, > > >where hostB will serve 2 times hostA connections. Forward > > >rules already masq internal hosts to outside (typical). > > > > > > ipchains -I input -p tcp -y -d yours.com/32 80 -m 1 > > > ipmasqadm mfw -I -m 1 -r hostA 80 -p 10 ipmasqadm mfw > > > -I -m 1 -r hostB 80 -p 20 > > > > > > Voila, load balancer ... any Linux distribution can do this. > > > > :)) And what about sessions ? > > As session data is usually stored in a database, Not always.. > have the sql server on a separate machine (you > would anyway if you have enough traffic to need > a load balancer, wouldn't you?) Generaly, right... Greetings -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: record a mac address of a networkcard
On Fri, 16 Mar 2001, Simon Tennant wrote: > I run arpwatch. It watches for new arp requests and mails me the > following: > > Subject: new station (w192.z064003150.sjc-ca.dsl.cnc.net) >Date: Thu, 08 Mar 2001 16:49:59 -0800 >From: [EMAIL PROTECTED] (Arpwatch) > To: [EMAIL PROTECTED] > > hostname: w192.z064003150.sjc-ca.dsl.cnc.net > ip address: 64.3.150.192 > ethernet address: 0:10:67:0:3b:92 > ethernet vendor: Redback Networks, Inc. >timestamp: Thursday, March 8, 2001 16:49:59 -0800 > At our campus we're usuing static ARP mapping, with table generated from our users database. But still one can easily change his MAC :( -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: webmail
On Tue, 17 Apr 2001, John Ackermann N8UR wrote: > > I want to recommend you IMP. You can set IMAP server on firewall and > > IMP/Horde. > > I've been playing with SilkyMail (http://www.cyrusoft.com/silkymail), which > is an enhanced version of IMP. It's under the GPL, but Cyrusoft are > offering commercial support packages for it. I'm very impressed, and > (under the right circumstances, which mine weren't) it's a snap to build as > all the tools (apache, php, etc.) are included in the tarball and a single > make command builds the whole environment. Then you untar the php files > into the htdocs directory, run a config script, and you're in business. > How about attachments ? Some time ago I've checked several webmail programs. And all of theme were doing one thing that make me cry - _load_ _whole_attachment_to_RAM_ before reciving / sending. I think it should be done by processing attachments part-by-part. -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Lilo problem
I have problems booting my system from new HD. I've got: - ABIT KT7 - Seagate connected to first channel (as /dev/hda) - CD-ROM as /dev/hdc - new disk (DTLA 307030) connected to HPT370 as /dev/hde I want to boot my system form /dev/hde, ( I want to be able to take /dev/hda away sometimes, and still be able to boot my system, of course after rerunning lilo). If new disk is the only one being seen by BIOS ( Seagate is disabled in bios setup) it works fine - new disk is 0x80 in BIOS. On the other hand when 2 disk are available, LILO stops after "LI". Of course I rerun lilo on every change. I guess it tries to access not the right drive. What will be the bios drive number of my new disk, when both are turned on ? 0x81 ? TIA -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Lilo problem
On Wed, 25 Apr 2001, Russell Coker wrote: > On Friday 20 April 2001 11:40, Przemyslaw Wegrzyn wrote: > > I have problems booting my system from new HD. > > > > I've got: > > - ABIT KT7 > > - Seagate connected to first channel (as /dev/hda) > > - CD-ROM as /dev/hdc > > - new disk (DTLA 307030) connected to HPT370 as /dev/hde > > > > I want to boot my system form /dev/hde, ( I want to be able to take > > /dev/hda away sometimes, and still be able to boot my system, of course > > after rerunning lilo). > > > > If new disk is the only one being seen by BIOS ( Seagate is disabled in > > bios setup) it works fine - new disk is 0x80 in BIOS. > > On the other hand when 2 disk are available, LILO stops after "LI". > > Of course I rerun lilo on every change. I guess it tries to access not the > > right drive. > > What will be the bios drive number of my new disk, when both are turned on > > ? 0x81 ? > > 0x81 sounds reasonable, but really it could be anything. The only thing to > do is to try telling LILO to use 0x81 and see if that makes it boot! It's already solved (by me :) ) - nobody proposed to put there 0x80. It works ! Actually, if I setup my BIOS to boot from RAID first , dev/hde becomes 0x80 and /dev/hda becomes 0x81 ! I guess (not checked) that drive selected to boot from always becomes 0x80 - that's why DOS knows how to load itself - it loads itself from 0x80 always ! If U boot from first drive there's no problem , because lilo figures it out as 0x80. disk = /dev/hde bios = 0x80 Voilla ! Well, now it's pretty trivial for me ;) but I've spent much time figuring it out. Maybe someone should include it in some faq ? Greetings -=Czaj-nick=- -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
