toasters

toasters@lists.teaparty.net

13515 discussions

Apache/Netapp problem + solution (fwd)
by David R. Van Sandt 05 Feb '98

05 Feb '98

What version of ONTAP are you using on the NetApp? ---------------------------------------------------------------- David R. Van Sandt, dv(a)corp.earthlink.net, 626/296-5137 Earthlink Network, Lead System Administrator "It's Your Internet" ---------------------------------------------------------------- ---------- Forwarded message ---------- Date: Wed, 4 Feb 1998 13:39:14 -0500 (EST) From: rasmus(a)bellglobal.com To: toasters(a)mathworks.com Subject: Apache/Netapp problem + solution Apache 1.2.5 on a Solaris 2.5.1 Sparc box connected to a Netapp 630 filer. Problem symptom: Web server established a connection, but would not respond to requests. Looking at the parent httpd process we see that it is in a normal "wait for something to do" loop: 0.000000 sigsuspend([] <unfinished ...> 0.415977 --- SIGALRM (Alarm Clock) --- 0.000214 <... sigsuspend resumed> ) = -1 EINTR (Interrupted system call) <0.415636> 0.000272 setcontext({uc_sigmask=[ALRM], ...}) = ? <0.000113> 0.000525 alarm(0) = 0 <0.000083> 0.000389 sigprocmask(SIG_UNBLOCK, [ALRM], NULL) = 0 <0.000087> 0.000475 sigaction(SIGALRM, {SIG_DFL}, NULL) = 0 <0.000086> 0.000464 waitid(P_ALL, 0, {si_signo=0, si_code=SI_USER, si_pid=0, si_uid=0, ...}, WNOHANG|WEXITED|WTRAPPED) = 0 <0.000091> 0.000499 alarm(0) = 0 <0.000084> 0.000349 sigaction(SIGALRM, {0xef5b8cfc, [], 0}, {SIG_DFL}) = 0 <0.000094> 0.000570 sigprocmask(SIG_BLOCK, [ALRM], []) = 0 <0.000090> 0.000577 alarm(1) = 0 <0.000086> 0.000351 sigsuspend([] <unfinished ...> 0.995342 --- SIGALRM (Alarm Clock) --- 0.000184 <... sigsuspend resumed> ) = -1 EINTR (Interrupted system call) <0.995330> 0.000248 setcontext({uc_sigmask=[ALRM], ...}) = ? <0.000107> 0.000517 alarm(0) = 0 <0.000083> Looking at the spawned child httpd's we see: 0.000000 fcntl(21, F_SETLKW, {type=F_WRLCK, whence=SEEK_SET, start=0, len=0} And they got there by: ef5b6fd8 fcntl (15, 7, 5cfcc) ef5b6fd8 _libc_fcntl (15, 7, 5cfcc, 0, 0, 0) + 8 ef78aba8 s_fcntl (15, 7, 5cfcc, ef611938, 34, effff944) + 164 00019138 accept_mutex_on (62d08, 1, 0, effffa78, effffa88, 2) + 18 0001a874 child_main (6c170, 5d000, 5d000, 5cc00, 5d000, 5b000) + 1f8 0001ad1c make_child (61880, 2, 61400, 11, 0, 5cfc9) + d4 0001b3a0 standalone_main (effffb78, 5cf48, 5f400, 5cc00, 75, 5f7b0) + 2a4 0001b920 main (3, effffc94, effffca4, 5f3f4, 1, 0) + 2a0 00017b48 _start (0, 0, 0, 0, 0, 0) + 5c Ok, so what are we locked on? Parent process says: Current rlimit: 92 file descriptors 0: S_IFCHR mode:0620 dev:32,0 ino:162863 uid:100 gid:7 rdev:24,2 O_RDWR 1: S_IFCHR mode:0620 dev:32,0 ino:162863 uid:100 gid:7 rdev:24,2 O_RDWR 2: S_IFCHR mode:0620 dev:32,0 ino:162863 uid:100 gid:7 rdev:24,2 O_RDWR 4: 0xd000 mode:0444 dev:164,0 ino:8952 uid:0 gid:0 size:0 O_RDONLY close-on-exec 15: S_IFCHR mode:0000 dev:32,0 ino:26208 uid:0 gid:0 rdev:42,8491 O_RDWR 16: S_IFREG mode:0644 dev:162,1 ino:2924135 uid:0 gid:1 size:13387 O_WRONLY|O_APPEND 17: S_IFCHR mode:0000 dev:32,0 ino:6664 uid:0 gid:0 rdev:42,8911 O_RDWR 18: S_IFREG mode:0644 dev:162,1 ino:4842016 uid:0 gid:1 size:61789 O_WRONLY|O_APPEND 19: S_IFREG mode:0644 dev:162,1 ino:4842022 uid:0 gid:1 size:4551030 O_WRONLY|O_APPEND 20: S_IFREG mode:0644 dev:162,1 ino:4842019 uid:0 gid:1 size:29387 O_WRONLY|O_APPEND 21: S_IFREG mode:0644 dev:162,1 ino:4842021 uid:0 gid:102 size:0 O_WRONLY advisory write lock set by system 0x7FFF process 2135 Inode 4842021 is: 4842021 -rw-r--r-- 1 root devel 0 Feb 3 15:34 logs/.nfsCAC So, for some reason the accept queue mutex mechanism is no longer working. There are no lock manager error messages from the NetApp, but this server which has been running fine for months will suddenly no longer lock. Other servers sitting right next to it with the same configuration lock fine. Would be nice to determine why it stopped working, but the solution is to add the the following line to the Apache httpd.conf file: LockFile /tmp/accept.lock That will keep the lockfile on the local disk. This is probably not a bad idea for anybody running Apache off of a Netapp to do. It may save you some future headaches. -Rasmus

2 1

Apache/Netapp problem + solution
by rasmus＠bellglobal.com 04 Feb '98

04 Feb '98

Apache 1.2.5 on a Solaris 2.5.1 Sparc box connected to a Netapp 630 filer. Problem symptom: Web server established a connection, but would not respond to requests. Looking at the parent httpd process we see that it is in a normal "wait for something to do" loop: 0.000000 sigsuspend([] <unfinished ...> 0.415977 --- SIGALRM (Alarm Clock) --- 0.000214 <... sigsuspend resumed> ) = -1 EINTR (Interrupted system call) <0.415636> 0.000272 setcontext({uc_sigmask=[ALRM], ...}) = ? <0.000113> 0.000525 alarm(0) = 0 <0.000083> 0.000389 sigprocmask(SIG_UNBLOCK, [ALRM], NULL) = 0 <0.000087> 0.000475 sigaction(SIGALRM, {SIG_DFL}, NULL) = 0 <0.000086> 0.000464 waitid(P_ALL, 0, {si_signo=0, si_code=SI_USER, si_pid=0, si_uid=0, ...}, WNOHANG|WEXITED|WTRAPPED) = 0 <0.000091> 0.000499 alarm(0) = 0 <0.000084> 0.000349 sigaction(SIGALRM, {0xef5b8cfc, [], 0}, {SIG_DFL}) = 0 <0.000094> 0.000570 sigprocmask(SIG_BLOCK, [ALRM], []) = 0 <0.000090> 0.000577 alarm(1) = 0 <0.000086> 0.000351 sigsuspend([] <unfinished ...> 0.995342 --- SIGALRM (Alarm Clock) --- 0.000184 <... sigsuspend resumed> ) = -1 EINTR (Interrupted system call) <0.995330> 0.000248 setcontext({uc_sigmask=[ALRM], ...}) = ? <0.000107> 0.000517 alarm(0) = 0 <0.000083> Looking at the spawned child httpd's we see: 0.000000 fcntl(21, F_SETLKW, {type=F_WRLCK, whence=SEEK_SET, start=0, len=0} And they got there by: ef5b6fd8 fcntl (15, 7, 5cfcc) ef5b6fd8 _libc_fcntl (15, 7, 5cfcc, 0, 0, 0) + 8 ef78aba8 s_fcntl (15, 7, 5cfcc, ef611938, 34, effff944) + 164 00019138 accept_mutex_on (62d08, 1, 0, effffa78, effffa88, 2) + 18 0001a874 child_main (6c170, 5d000, 5d000, 5cc00, 5d000, 5b000) + 1f8 0001ad1c make_child (61880, 2, 61400, 11, 0, 5cfc9) + d4 0001b3a0 standalone_main (effffb78, 5cf48, 5f400, 5cc00, 75, 5f7b0) + 2a4 0001b920 main (3, effffc94, effffca4, 5f3f4, 1, 0) + 2a0 00017b48 _start (0, 0, 0, 0, 0, 0) + 5c Ok, so what are we locked on? Parent process says: Current rlimit: 92 file descriptors 0: S_IFCHR mode:0620 dev:32,0 ino:162863 uid:100 gid:7 rdev:24,2 O_RDWR 1: S_IFCHR mode:0620 dev:32,0 ino:162863 uid:100 gid:7 rdev:24,2 O_RDWR 2: S_IFCHR mode:0620 dev:32,0 ino:162863 uid:100 gid:7 rdev:24,2 O_RDWR 4: 0xd000 mode:0444 dev:164,0 ino:8952 uid:0 gid:0 size:0 O_RDONLY close-on-exec 15: S_IFCHR mode:0000 dev:32,0 ino:26208 uid:0 gid:0 rdev:42,8491 O_RDWR 16: S_IFREG mode:0644 dev:162,1 ino:2924135 uid:0 gid:1 size:13387 O_WRONLY|O_APPEND 17: S_IFCHR mode:0000 dev:32,0 ino:6664 uid:0 gid:0 rdev:42,8911 O_RDWR 18: S_IFREG mode:0644 dev:162,1 ino:4842016 uid:0 gid:1 size:61789 O_WRONLY|O_APPEND 19: S_IFREG mode:0644 dev:162,1 ino:4842022 uid:0 gid:1 size:4551030 O_WRONLY|O_APPEND 20: S_IFREG mode:0644 dev:162,1 ino:4842019 uid:0 gid:1 size:29387 O_WRONLY|O_APPEND 21: S_IFREG mode:0644 dev:162,1 ino:4842021 uid:0 gid:102 size:0 O_WRONLY advisory write lock set by system 0x7FFF process 2135 Inode 4842021 is: 4842021 -rw-r--r-- 1 root devel 0 Feb 3 15:34 logs/.nfsCAC So, for some reason the accept queue mutex mechanism is no longer working. There are no lock manager error messages from the NetApp, but this server which has been running fine for months will suddenly no longer lock. Other servers sitting right next to it with the same configuration lock fine. Would be nice to determine why it stopped working, but the solution is to add the the following line to the Apache httpd.conf file: LockFile /tmp/accept.lock That will keep the lockfile on the local disk. This is probably not a bad idea for anybody running Apache off of a Netapp to do. It may save you some future headaches. -Rasmus

1 0

strange cached information for CIFS
by Tom Yates 28 Jan '98

28 Jan '98

i have two F330s, both at 4.3R4, both running CIFS. apart from the number of shares on each one, and the machine names, the CIFS setups are supposed to be identical. one will authenticate through a PDC, and one won't. we have been through the old "create a domain account for the server" thing *many, many* times. it's not helping. when we try to "cifs setup" to use the PDC, we get the following joyous dialogue: Tue Jan 27 18:07:18 EST [rc]: Connection with DC BIGDADDY terminated Tue Jan 27 18:07:18 EST [rc]: CIFS - Filer retrying authentication with backup password. Tue Jan 27 18:07:28 EST [rc]: Connection with DC BIGDADDY terminated Tue Jan 27 18:07:28 EST [rc]: CIFS - Filer's security information differs from domain controller. and, in many cases, cifs will sit there trying to shut down, locking up the console, until i reboot the filer. it's been suggested to me that it needs to authenticate as part of the shutdown process. so, my questions are: 1) what flaming backup password? 2) what decides whether cifs starts up at boot time, absent a "cifs on" rc entry? 3) is it true that the delayed shutdown is caused by failure to authenticate? 4) has anyone else had trouble going from X to 4.3R4, esp. if X=4.0.1c{,D10}? and, if were allowed a slightly less serious wish list, any chance of an option such as options no.stupid.microsoft.authentication on to get the CIFS support to authenticate based on the user logged into the NT Workstation in question, a la NFS? or do i misunderstand terribly what's going on here? Tom Yates - Unix Chap - The Mathworks, Inc. - +1 (508) 647 7561 MAG#65061 DoD#0135 AMA#461546 1024/CFDFDE39 0C E7 46 60 BB 96 87 05 04 BD FB F8 BB 20 C1 8C

1 0

Faq(?) - Sendmail & Toasters getting along well?
by Don Glascock 23 Jan '98

23 Jan '98

Hi. We're evaluating/purchasing an F230, and are slowly getting all of our i's dotted and t's crossed in setting it up. In our environment, the sendmail.cf files on our Unix boxes hand everything over to the main file/mail server in our group, which is a DEC Alpha running Digital Unix. The workstations mount its "/usr/spool/mail" as "/usr/spool/mail" or "/usr/mail", as you would expect. With this configuration, we've had very few file locking problems, presuming that life is easier if the sendmail daemon that actually manipulates mail files runs on the same box on which the files sit. Now, we're debating about moving everyone's mail folders to the F230, which means that sendmail on our Alpha will be making regular use of NFS file locking. Has anyone tried this? How frequently (if ever) do you run into problems? Thanks & hope this isn't a FAQ... Don Glascock (ps, I'm new to the list -- is there an archive for this mailing list? I didn't see one on ftp.mathworks.com, and Majordomo says there are no "toasters"-related files. Thanks, again! :-) -- Donald S. Glascock Mailstop GU-1016 Special Purpose Processor Development Group Mayo Foundation Rochester, MN 55905 Glascock.Donald(a)Mayo.EDU +1.507.284.1850 +1.507.284.9171 (fax) http://www.mayo.edu/sppdg/sppdg_home_page.html "No matter where you go, there you are."

3 2

F630 with more than 52 disks ?
by Daniel Quinlan 23 Jan '98

23 Jan '98

I'm not planning on it, but I'd like to know if anyone out there has an F630 (or anything) with more than 52 9GB disks on it (~ 200GB). If you do, how long does a disk recontruction take? Our F630 with 52 9GB disks took 25 hours to reconstruct a failed disk (while it was being used, so we had to turn the knob down to 2 and 1 for most of the time because response was so slow). Dan

1 0

Solaris 2.6 drops NFS automounts
by Mike Mueller 23 Jan '98

23 Jan '98

We have heard reports that Solaris 2.6 NFS clients drop mounts of NetApp and other NFS servers but I have not been able to get a detailed explanation of the problem. Although it is apparently a Solaris 2.6 bug, I can't seem to find information about it from Sun and I thought that some other NetApp user might be familiar with the problem. We are planning a Solaris upgrade in the next few weeks and I'd like to feel confident that we aren't going to run into problems. Can someone explain what the specific bug is and whether there is a definate work-around or patch available? Is anyone successfully using Solaris 2.6 clients with a NetApp and if so are you using any special settings on either the client and/or the server? Thanks. Mike Mueller

2 1

Re: Anybody besides me having quotas go bonkers?
by Jan-Pieter Cornet 23 Jan '98

23 Jan '98

In article <199801202125.NAA26037(a)tooting.netapp.com>, Rajesh Sundaram <rajeshs(a)netapp.com> wrote: > attribution missing, but someone(a)cdsnet.net wrote: >> I just created some new users, and the filer (running 4.3R4D1) >> is reporting that they've used: >> >> user 506 9840 41943040 1292 10485760 >> user 1001 4294966368 41943040 22 10485760 >> user@web 1001 4294966408 25600 18 1000 >> user 1003 4294967260 41943040 1 10485760 >> user@web 1003 4294967260 25600 0 1000 >> user 1004 4294967260 41943040 1 10485760 >> user@web 1004 4294967260 25600 0 1000 >> >> which is flatly impossible, the users were just created and assigned. yet >> it thinks the disk space usage is all used up. We have been seeing the same thing last week. We fixed it by putting a (root-owner) file in the users' homedirectory, then chowning that file to the user, making their disk usage positive again. >Are default quotas being used here? If that's the case, this is likely to >be the bug with operations running as root that others in the list >have pointed out. > >Default quotas trigger the creation of quotas when a user first >uses an inode or a data block (by creating a file, for example). >This should happen even when the file is created as root and then >"chown"ed to the user. This currently fails to happen. > >Here is the description from the bug report which should soon get >added to the bugs online list on NOW. > >TITLE: >Derived quotas do not get created for operations >running as root, causing usage decrepancies. We upgraded to 4.3R4D2, which apparently fixed it. At least, I haven't seen "negative" disk usages in almost a day, whereas our previous version (4.3R4D1) would have had some by now. >> What's even weirder is that I can create directories all day, and no >> errors, only files cause problems. yet the dirs are owned by user 1001, >> etc. Toasters do not count directory blocks as disk usage, so directories can't be affected by this bug. -- #! ##### Jan-Pieter Cornet ##### <johnpc(a)xs4all.net> ##### perl ++$_;$!=$_+++$_;($:,$,,$/,$*)=$!=~/.(.)...(.)(.).(.)/;$!=$_+$_; ($(a),$\,$~)=$!=~/(.)(.).(.)/; $_="$,$/$:"; $@++; $~="$~$_";($_)= \$$=~/\((.)/;$|=++$_;$_++;$|++;$~="$~ $@$:";`$~$/$\$*$, $|>&$_`

1 0

4.3R4 and statd
by Michael Salmon 22 Jan '98

22 Jan '98

We have just installed 4.3R4 in a 520 with storage shelves 2 and statd has begun complaining about the shelves. Shelfchk has also stopped working so we are forced to conclude that the connection between the filer and the shelves has stopped working. Any comments or suggestions? /Michael -- /--------------------------------\ ___ | Michael Salmon | / \ | Ericsson Utveckling AB | / | | ÄS/UAB/Z/IPE | / _| | Box 1505 | / | | 125 25 Älvsjö | / / | Sweden | / / | | | | +--------------------------------+ | \ | Michael.Salmon(a)uab.ericsson.se | / *| | | / _/ | Tel: +46 8 727 30 74 | | | | FAX: +46 8 647 82 76 | | / # | GSM: +46 705 902 842 | | __// \--------------------------------/ |_/

1 0

Archived mail
by Dawn Hamel 22 Jan '98

22 Jan '98

Is the mail from this list archived anywhere. I believe there was some discussion about problems upgrading to Solaris 2.6.1 and I want to make sure a new Netapp customer is aware of any problems BEFORE they upgrade to the 2.6.1 release Thanks, Dawn

8 9

Anybody besides me having quotas go bonkers?
by Jaye Mathisen 20 Jan '98

20 Jan '98

I have a few quota trees set up on a 540 with 100GB. I just created some new users, and the filer (running 4.3R4D1) is reporting that they've used: user 506 9840 41943040 1292 10485760 user 1001 4294966368 41943040 22 10485760 user@web 1001 4294966408 25600 18 1000 user 1003 4294967260 41943040 1 10485760 user@web 1003 4294967260 25600 0 1000 user 1004 4294967260 41943040 1 10485760 user@web 1004 4294967260 25600 0 1000 which is flatly impossible, the users were just created and assigned. yet it thinks the disk space usage is all used up. What's even weirder is that I can create directories all day, and no errors, only files cause problems. yet the dirs are owned by user 1001, etc. If anybody has any ideas, I'd love to hear it, occasionally working quotas is worse than no quotas, since it prevents mail from being deliverd (using qmail), yet seems to run OK. My quota file is trivial, just a few entries. CC'd to support@netapp, because I have a separate open ticket where df will show -8kb used, -8kb free at odd times. Any tips appreciated, I could almost strangle this thing.

2 1

Re: Anybody besides me having quotas go bonkers?
by Daniel Quinlan 19 Jan '98

19 Jan '98

Jaye Mathisen <mrcpu(a)cdsnet.net> writes: > which is flatly impossible, the users were just created and assigned. yet > it thinks the disk space usage is all used up. We had a similar problem with some 4.2 versions. Basically, files would get created as root and then chowned to a user, but without any quota usage attribution. Later, the user deletes some files. 0 minus some number is a really big number because of integer overflow. Quota exceeded. It looks like you may have a similar problem. Generally, quotas are unreliable for us. I always "quota off; quota on" whenever I add a new directory directly under one of our quota trees. The biggest problem for us is that we don't really want to use quotas to limit disk usage, we just want to track it. It should be really easy to have a "nice quotas" option that causes the filer to never send an EDQUOT (quota exceeded) error. That one feature would make most of our quota problems non-problems. - Dan

2 2

None
by Kelly Wyatt 16 Jan '98

16 Jan '98

subscribe toasters

1 0

Shortcomings of current NetApps?
by Marc Nicholas 15 Jan '98

15 Jan '98

I'm interested to hear what others think the shortcomings of the current crop of NetApps are. From my experience: * Single logical drive * Weedy SCSI performance * Lack of other RAID implementations * Not quite enough redundancy for truly mission critical applications Don't get me wrong, I love 'em :-) -marc --- Marc Nicholas - Hippocampus OSD, Inc. - Eastern Office 416 979 9000 - fax: 416 979 8223 - http://www.hippocampus.net 125 John St. - Suite #100 - Toronto - Ontario - M5V 2E2 - CANADA "Inter/Intra/Extra[net] consulting, corporate access, hardware and software sales"

14 27

Re: Shortcomings of current NetApps?
by Daniel Quinlan 15 Jan '98

15 Jan '98

rasmus(a)lerdorf.on.ca (Rasmus Lerdorf) writes: > > > 1. Out-of-date SNMP implemention - (no CIFS or disk system statistics, > > for example) > > What do you mean by disk system stats here? The information displayed by the "sysstat" command. You can get the NFS and CPU load via SNMP, but that's it. CPU NFS CIFS HTTP Net kB/s Disk kB/s Tape kB/s Cache in out read write read write age --- --- --- ---- Only the items with dashes under them are available via SNMP. Dan

1 0

Re: Shortcomings of current NetApps?
by mds46523＠ggr.co.uk 15 Jan '98

15 Jan '98

On Jan 15, 8:46, Brian Tao wrote: > Subject: Re: Shortcomings of current NetApps? > I have a really simple request: a local "rm" command that can > take arguments either on the command line or from stdin. Anyone > running a traditional news spool on a Netapp will probably appreciate > what I'm getting at. ;-) A local "du" command would be rather spiffy too. A special du keyed to UID, ie the space used by a person in a given directory tree would also be kinda cool. I think that there are certain operations where having the filer do the whole thing would save the network, your client machine and the filer itself a lot of time and probably get the job done quicker. The operation could probably be made as low priority as you like and it would still have advantages. Maybe if there were a JAVA VM on the filer that could manipulate the filesystem consistently we could have a bash ourselves. Clearly this could be open to abuse of the Appliance principle, but that's where discipline comes in. > -- > Brian Tao (BT300, taob(a)netcom.ca) > "Though this be madness, yet there is method in't" >-- End of excerpt from Brian Tao -- -Mark, TSG Unix admin and support, int 782 4412, ext +44 1438 76 4412. ... an Englishman in London ...

2 1

Re: Shortcomings of current NetApps?
by Daniel Quinlan 15 Jan '98

15 Jan '98

Marc Nicholas <marc(a)hippocampus.net> writes: > I'm interested to hear what others think the shortcomings of the > current crop of NetApps are. In addition to what other people have said: 1. Out-of-date SNMP implemention - (no CIFS or disk system statistics, for example) 2. ndmpcopy - incrementals don't work, rsh as UID 0 doesn't work, requires password to be passed on command line 3. NetApps are reboot-prone. Often will reboot when it should have just failed an obviously bad disk. 4. No command line Unix permissions tool for CIFS, the graphical permissions tool uses an undocumented proprietary extension to CIFS (that is, it won't work with Samba). 5. Disk reconstructions take far too long. We had one on a 200GB filesystem take 24 hours (dial set to 2 because any faster was causing severe problems for users, and this is on a decked-out F630). Dan

2 1

Re: Shortcomings of current NetApps?
by mds46523＠ggr.co.uk 15 Jan '98

15 Jan '98

On Jan 15, 11:14, James R Grinter wrote: > Subject: Re: Shortcomings of current NetApps? > On Thu 15 Jan, 1998, Alexei Rodriguez <alexei(a)cimedia.com> wrote: > >I understand the benefits. I just don't care for having 5 4GB > >slices which you have to manage (ala auspex). > > I thought the idea of multiple RAID sets is that you layer one > filesystem atop them all (if you want to. Of course, others want > to have multiple filesystems.) > My thought was that for once we'd have separate fs's in one NetApp box. Which would allow us to have separate snapshot schedules in the same box. Which would be nice and useful. > (Which is something you can, and people do, do with multiple > RAID controllers and Unix systems, etc) > > -- jrg. >-- End of excerpt from James R Grinter -- -Mark, TSG Unix admin and support, int 782 4412, ext +44 1438 76 4412. ... an Englishman in London ...

1 0

Re: Shortcomings of current NetApps?
by mds46523＠ggr.co.uk 15 Jan '98

15 Jan '98

I've got two things to say about these points: (a) Testing. (b) Testing. (insert standard joke about 'it's so important I mention it twice') Seriously though, we've been bitten a once or twice by component failures on our NetApp's but compared to the DOA's we get from Sun, SGI and the rest, I'm quite happy for NetApp to spend more on their factory space, power, personnel and so forth, AND charge a little more. Maybe amortising the RnD costs of NetApp's over the whole range of components isn't as off-putting to users as dumping it all in the base price of the strictly NetApp-designed-and-built hardware and software. Plus it means that NetApp engineers can reasonably expect problems with hardware and software to be NetApp-owned because they were NetApp supplied. NetApp seem to want to make owning their boxes as painless and fruitful as they can. How often do you see Rolls-Royce's in the shop being fitted with parts that aren't approved by RR? Likewise Ferrari? I betcha parts for econoboxes would be cheaper, and might even come from the same basic supplier, but the added scrutiny means added reassurance, which is worth money. Maybe NetApp could introduce a poor-man's NetApp (called something more tactful and market-savvy of course)... But I think that would be more trouble than it's worth no matter which side you're on. On Jan 14, 22:33, Jonah Yokubaitis wrote: > Subject: Re: Shortcomings of current NetApps? > On Wed, 14 Jan 1998, Chris Caputo wrote: > > |>I'm interested to hear what others think the shortcomings of the current > |>crop of NetApps are. From my experience: > | > |My biggest beef is that the upgradability of these things is low. It > |seems that when you buy a unit you are buying a capacity and improvements > |in disk technology don't raise it. For example, our unit uses 4 gig > |drives. Its max capacity is 14 4 gig drives. We'd love to put 9 or 18 > |gig drives in it, but the software doesn't support it. I realize 9 or 18 > |gig drives would raise rebuild time, but that risk should be up to us. > > Amen to that. > > A paste from a sysconfig -r on an F630 > > 1: SEAGATE ST15150W 9107 Size=3.9GB (8388315 blocks) > > The ST15150W is an *OLD* drive that Seagate doesn't even list as a > current product..... > > Some newer drives please? 18gig drives please? It *can't* be that hard > to allow 18gig drives...if it is your programmers didn't do their job > correctly. > > |I also consider the high price per meg to be a shortcoming and because of > |it we are researching and developing alternatives. > > Just negotiate hard with Netapp. I still feel screwed everytime I want > an upgrade from netapp. The above drive model is $355 through > disribution...netapp quoted me $1700/drive. > > Their NVRAM pricing is LUDICRIOUS considering its 3rd party NVRAM > bought from Dallas semiconductor. Ditto for DRAM. > > Netapp: > > Make your money on your wonderful software, and stop trying to price > gouge your customers on *off the shelf* items like Seagate Hard > DRives, Samsung DRAM, and Dallas Semiconductor NVRAM. > > We really aren't that stupid I promise. > > Jonah > > Jonah Barron Yokubaitis | Austin|San Antonio|Houston > President | Dallas|Fort Worth|Boerne > Texas.Net | Georgetown|Dripping Springs > http://www.texas.net | Making 56k affordable > >-- End of excerpt from Jonah Yokubaitis -- -Mark, TSG Unix admin and support, int 782 4412, ext +44 1438 76 4412. ... an Englishman in London ...

1 0

subscribe
by tonylai 15 Jan '98

15 Jan '98

1 0

Re: Shortcomings of current NetApps?
by Richard Fairthorne 15 Jan '98

15 Jan '98

I would like to see the NetApp re-share CIFS according to cifsconfig.txt after a power-loss, not just a clean shut-down. Richard Fairthorne ---------- > From: Brett Rabe <brett(a)uswest.net> > To: Marc Nicholas <marc(a)hippocampus.net> > Cc: toasters(a)mathworks.com > Subject: Re: Shortcomings of current NetApps? > Date: Wednesday, January 14, 1998 1:18 PM > > > On Wed, 14 Jan 1998 12:40:47 EST, Marc Nicholas wrote: > >I'm interested to hear what others think the shortcomings of the current > >crop of NetApps are. From my experience: > > > > [snip] > >* Not quite enough redundancy for truly mission critical applications > > What other types of redundancy would you like to see? > > Brett > > --- > Brett Rabe Email : brett(a)uswest.net > Systems Administrator - U S West Phone : 612.664.3078 > 600 Stinson Blvd. Pager : 612.613.2549 > Minneapolis, MN USA 55413 Fax : 612.664.4770 > > If you aren't the lead dog, the view is always the same.

2 1

Re: NetApp and amanda
by Alan Judge 14 Jan '98

14 Jan '98

> In article <199801131829.MAA18731(a)vortex.more.net>, > David Drum <david(a)more.net> wrote: >> I was wondering if anyone uses amanda to backup their NetApp(s)? Jan-Pieter> Yup, we do. At least, we used to, until yesterday :) The Jan-Pieter> load on the netapp during phase I of the dump is Jan-Pieter> prohibitively high, and consequently makes our webserver Jan-Pieter> very slow... we're currently rethinking things. We're doing something similar, though in our case we modified amanda itself to understand devices like /@filer1 and understand the output of NetApp dump, rather than replacing dump. Works great, though I have the same concerns as you about the cost of amanda's dump estimates. (Though the cost of creating a snapshot doesn't seem significant for us.) I'd be interested in hearing what solution you come up with. -- Alan

1 0

Re: NetApp and amanda
by Jan-Pieter Cornet 13 Jan '98

13 Jan '98

In article <199801131829.MAA18731(a)vortex.more.net>, David Drum <david(a)more.net> wrote: >I was wondering if anyone uses amanda to backup their NetApp(s)? Yup, we do. At least, we used to, until yesterday :) The load on the netapp during phase I of the dump is prohibitively high, and consequently makes our webserver very slow... we're currently rethinking things. There are a few problems to backing up a toaster via amanda, and here are the solutions, roughly: - the toaster doesn't run amandad :) And you can't easily make it do so. Our solution is to have amanda backup /dev/netapp1/path on another machine. To keep the network happy, that "other machine" is also the one that the tape is connected to, and the one that amanda (the master program, or the client, whatever you want to call it) runs on. "dump" on this machine is actually a script that intercepts requests to dump devices of the form /dev/netapp*. If it isn't, it calls the regular dump. If it is, it calls a separate program, which translates some options, and then starts dump on the toaster, via rsh. (Note that the devices that amanda thinks it's backing up need to exist. Just create them and make them symlinks or hardlinks to /dev/null or /dev/zero. Doesn't really matter). Next problem: - amanda does not grok toaster "dump" output. Perl to the rescue again (the dump wrapper is perl, too :). Since we had to start a separate script anyway, we're piping STDERR through the script, and have it do on-the-fly translations so the output of "dump" as produced by the toaster is massaged into something that amanda understands. Additionally, we trap "dump W" and read the /etc/dumpdates file on the toaster too, and generate extra output for it. A few considerations you have to make: - amanda starts by getting estimates for all filesystems it backs up, that is, it makes at least 2 dumps for each filesystem, and terminates the dump once the "estimated" line is in. If the "dump" command on the toaster creates its own snapshot each time you do this, you waste a lot of CPU. We settled on generating a snapshot once each day, and running all backups from that. You have to take extra special care though not to refresh the snapshot while a backup is still running. - amanda doesn't know anything about "oh don't try to backup this so often". At least I can't configure it to do so (without extreme hacks). It knows about priorities, but it will always at least try to make a backup, and if it fits on tape, it will. This generally means that amanda does make a backup of all data on the toaster, at least incrementally. Not strictly necessary with the RAID and all that. If there is interest, I can post the sources we use to the list, or make them available on the web. Note that the sources are very alpha, and most definately not directly usable on other sites, some tweaking will be necessary. Best regards, -- #! ##### Jan-Pieter Cornet ##### <johnpc(a)xs4all.net> ##### perl ++$_;$!=$_+++$_;($:,$,,$/,$*)=$!=~/.(.)...(.)(.).(.)/;$!=$_+$_; ($(a),$\,$~)=$!=~/(.)(.).(.)/; $_="$,$/$:"; $@++; $~="$~$_";($_)= \$$=~/\((.)/;$|=++$_;$_++;$|++;$~="$~ $@$:";`$~$/$\$*$, $|>&$_`

1 0

NetApp and amanda
by David Drum 13 Jan '98

13 Jan '98

Hello, I was wondering if anyone uses amanda to backup their NetApp(s)? Regards, David K. Drum david(a)more.net -- It's hard to be bored when you're as stupid as a line. [1] Reality has a tendency to be so uncomfortably real. [2] You can only measure the size of your head from the inside. [3] When you proceed deliberately, mistakes don't cascade, they instruct. [4] [1] Vernor Vinge [2] Neil Peart [3] Larry Wall [4] Stewart Brand

1 0

Re: Usenet: Crosspost - File Exists error
by Ronan Mullally 13 Jan '98

13 Jan '98

> We've just installed a 540 to provide back-end storage for our news system. > Since going live the INN error log has been full of the following messages: > > crosspost cant link alt/jobs/16941 misc/jobs/offered/24616 : File exists > crosspost cant link talk/politics/misc/2349 alt/feminism/944 : File exists ... > Has anybody come across this before? What is to blame - INN, Linux > or the filer? The answer is "None of the above", I'd lost the '-L' flag for innd in the upgrade. Thanks to Ed Klein of BBN Planet for his speedy diagnosis :) -- Ronan Mullally 'In the Beginning there was Network Manager nothing, which exploded.' Ireland On-Line -- Terry Pratchett.

1 0

Usenet: Crosspost - File Exists error
by Ronan Mullally 13 Jan '98

13 Jan '98

We've just installed a 540 to provide back-end storage for our news system. Since going live the INN error log has been full of the following messages: crosspost cant link alt/jobs/16941 misc/jobs/offered/24616 : File exists crosspost cant link talk/politics/misc/2349 alt/feminism/944 : File exists It appears that the crosspost daemon is generating this error while trying to create the necessary hard links for *every* crossposted article (I have yet to verify this) it encounters. The link *is* successfully made, and the article appears in the directory but the error log contains tens of thousands of error messages per day. Before switching to the filer we encountered no such problems. The INN feeder system is a Pentium running Linux 2.0.33. The filer is a 540 running 4.3R4. I encountered the same problem on a 210, also running 4.3R4. I've tracked the error down to the following code fragment (backends/crosspost.c line 172): /* 2nd try to link */ lnval = link(names[0], names[i]) ; if (lnval < 0 && errno == EXDEV) { #if defined(DONT_HAVE_SYMLINK) ... symlink stuff removed ... #endif /* defined(DONT_HAVE_SYMLINK) */ } else if (lnval < 0) { (void)fprintf(stderr, "crosspost cant link %s % names[0], names[i]); perror(" "); } So it appears that link() is returning an EEXISTS error. The link() man page includes the following: BUGS On NFS file systems, the return code may be wrong in case the NFS server performs the link creation and dies before it can say so. Use stat(2) to find out if the link got created. Has anybody come across this before? What is to blame - INN, Linux or the filer? -- Ronan Mullally 'In the Beginning there was Network Manager nothing, which exploded.' Ireland On-Line -- Terry Pratchett.

1 0

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

toasters