[gentoo-user] Re: CIFS mounts started misbehaving
On 2017-03-06, Grant Edwards wrote: > On 2017-03-03, Grant Edwards wrote: > >> For the past 10-15 [years], I've been mounting a handfull of >> directories that reside on a Windows server, and it's always worked >> fine. >> >> About a week ago, they started acting oddly. They all mount fine, >> and work as usual as long as you keep using them. AFAICT, if they >> sit idle for "a while" (tens of minutes, maybe an hour), they >> freeze up. [...] > It's a kernel 4.9 problem. [...] > Rebooting with the 4.4.39 kernel fixes the problem. FWIW, I've been running 4.9.34 since yesterday, and all my CIFS mounts seem stable. -- Grant Edwards grant.b.edwardsYow! TONY RANDALL! Is YOUR at life a PATIO of FUN?? gmail.com
[gentoo-user] Re: CIFS mounts started misbehaving
On 2017-03-07, Marc Joliet wrote: > On Dienstag, 7. M�rz 2017 15:19:33 CET Grant Edwards wrote: >> No, as a rule I run stable gentoo-sources, and that's at 4.9.6-r1. > > Ah, of course. I'm using ~arch kernels ATM. (As a btrfs user I was tracking > the most recent upstream stable series, but want to switch to LTS kernels > now, > which happen to be the ones Gentoo stabilizes. That is, unless a newer > kernel > has something that I really, *really* want.) If I get bored, I may try latest 4.9, 4.10, 4.11 on my "test" machine and see how it acts there. I did some googling and didn't find anything that looked like this problem being reported anywhere -- which makes me wonder what particular combination of ancient windows server (IIRC, it's 2008) and odd confguration is causing it. I have a hard time believing that the CIFS client in 4.9 kernels can be doing this for very many people... -- Grant Edwards grant.b.edwardsYow! Uh-oh!! I forgot at to submit to COMPULSORY gmail.comURINALYSIS!
Re: [gentoo-user] Re: CIFS mounts started misbehaving
On Dienstag, 7. März 2017 15:19:33 CET Grant Edwards wrote: > No, as a rule I run stable gentoo-sources, and that's at 4.9.6-r1. Ah, of course. I'm using ~arch kernels ATM. (As a btrfs user I was tracking the most recent upstream stable series, but want to switch to LTS kernels now, which happen to be the ones Gentoo stabilizes. That is, unless a newer kernel has something that I really, *really* want.) > However, I'm a bit confused about the table shown at > > https://packages.gentoo.org/packages/sys-kernel/gentoo-sources > > There are two rows for some versions (e.g. 4.9.6-r1), with different > indicators. What does that mean? AFAIK that's a (known?) bug. -- Marc Joliet -- "People who think they know everything really annoy those of us who know we don't" - Bjarne Stroustrup signature.asc Description: This is a digitally signed message part.
[gentoo-user] Re: CIFS mounts started misbehaving
On 2017-03-07, Marc Joliet wrote: >> It's a kernel 4.9 problem. >> >> I had built and installed a gentoo-sources 4.9.6-r1 kernel about a >> month ago, but didn't update the grub configuration and reboot until >> two weeks ago. >> >> Rebooting with the 4.4.39 kernel fixes the problem. [...] >> I guess I'll have to stick with the 4.4 series until this gets fixed. > > I'm glad you found the source of the problem and a workaround. > However, the 4.9 series is now at 4.9.13. Have you tried that, too? No, as a rule I run stable gentoo-sources, and that's at 4.9.6-r1. However, I'm a bit confused about the table shown at https://packages.gentoo.org/packages/sys-kernel/gentoo-sources There are two rows for some versions (e.g. 4.9.6-r1), with different indicators. What does that mean? -- Grant Edwards grant.b.edwardsYow! for ARTIFICIAL at FLAVORING!! gmail.com
[gentoo-user] Re: CIFS mounts started misbehaving
Am Mon, 06 Mar 2017 19:01:57 + schrieb "J. Roeleveld" : > On March 6, 2017 5:14:39 PM GMT+01:00, Grant Edwards > wrote: > >On 2017-03-06, Kai Krakow wrote: > > > [...] > >and > [...] > >> > >> Did something on the Windows side change? > > > >Probaby, but I've learned not to ask questions like that. They never > >get answered, and it just causes problems when it is revealed that > >the client having problems is a Linux machine. > > > >> Maybe force Windows down to a lower SMB version or reduce/disable > >> SMB client side caching? > > Windows sharing is designed as a 'link when used' option. Not as a > permanent mount like Linix treats it. > > Even 'mounting' in Windows doesn't mean the share is actually > accessed. > > A windows CIFS server will not be reliable enough for long term > mounting. With Samba, it does work more reliable. (In my experience) > > For this reason, I use KDE/Dolphin to access CIFS shares. It is > closer to how Windows expects the shares to be treated. Then it may help to use automount with a somewhat low timeout, maybe also setup cachefilesd and mount with fsc option. This is how I use my office shares on a 2012 R2 server via VPN. -- Regards, Kai Replies to list-only preferred.
Re: [gentoo-user] Re: CIFS mounts started misbehaving
On Dienstag, 7. März 2017 00:12:06 CET Grant Edwards wrote: > On 2017-03-03, Grant Edwards wrote: > > For the past 10-15 [years], I've been mounting a handfull of > > directories that reside on a Windows server, and it's always worked > > find. > > > > About a week ago, they started acting oddly. They all mount fine, and > > work as usual as long as you keep using them. AFAICT, if they sit > > idle for "a while" (tens of minutes, maybe an hour), they freeze up. > > It finally dawned on me that I had changed something. > > It's a kernel 4.9 problem. > > I had built and installed a gentoo-sources 4.9.6-r1 kernel about a > month ago, but didn't update the grub configuration and reboot until > two weeks ago. > > Rebooting with the 4.4.39 kernel fixes the problem. > > [I also tried just rebooting the 4.9.4 kernel, but that didn't help.] > > The configuration of the 4.9.4 kernel is as close to that of the > 4.4.39 as I can get. > > I guess I'll have to stick with the 4.4 series until this gets fixed. I'm glad you found the source of the problem and a workaround. However, the 4.9 series is now at 4.9.13. Have you tried that, too? HTH -- Marc Joliet -- "People who think they know everything really annoy those of us who know we don't" - Bjarne Stroustrup signature.asc Description: This is a digitally signed message part.
[gentoo-user] Re: CIFS mounts started misbehaving
On 2017-03-03, Grant Edwards wrote: > For the past 10-15 [years], I've been mounting a handfull of > directories that reside on a Windows server, and it's always worked > find. > > About a week ago, they started acting oddly. They all mount fine, and > work as usual as long as you keep using them. AFAICT, if they sit > idle for "a while" (tens of minutes, maybe an hour), they freeze up. It finally dawned on me that I had changed something. It's a kernel 4.9 problem. I had built and installed a gentoo-sources 4.9.6-r1 kernel about a month ago, but didn't update the grub configuration and reboot until two weeks ago. Rebooting with the 4.4.39 kernel fixes the problem. [I also tried just rebooting the 4.9.4 kernel, but that didn't help.] The configuration of the 4.9.4 kernel is as close to that of the 4.4.39 as I can get. I guess I'll have to stick with the 4.4 series until this gets fixed. -- Grant Edwards grant.b.edwardsYow! Now we can become at alcoholics! gmail.com
Re: [gentoo-user] Re: CIFS mounts started misbehaving
On March 6, 2017 8:17:37 PM GMT+01:00, Grant Edwards wrote: >On 2017-03-06, J. Roeleveld wrote: >> On March 6, 2017 5:14:39 PM GMT+01:00, Grant Edwards > wrote: >>>On 2017-03-06, Kai Krakow wrote: >>> > I'm going to try to set up a Wireshark capture in ring-buffer mode >>>and > somehow detect the failure and stop the capture... Did something on the Windows side change? >>> >>>Probaby, but I've learned not to ask questions like that. They never >>>get answered, and it just causes problems when it is revealed that >the >>>client having problems is a Linux machine. >>> Maybe force Windows down to a lower SMB version or reduce/disable SMB client side caching? >> >> Windows sharing is designed as a 'link when used' option. Not as a >> permanent mount like Linix treats it. >> >> Even 'mounting' in Windows doesn't mean the share is actually >> accessed. >> >> A windows CIFS server will not be reliable enough for long term >> mounting. With Samba, it does work more reliable. (In my experience) > >It's worked perfectly fine for 10+ years, and apparently continues to >do so for other Linux users in the office. And trying to troubleshoot it is not simple. Especially as MS Windows event viewer never shows anything remotely useful. (I tried to troubleshoot various issues, never got anything usefull from the windows admins or event viewer) How do the other Linux users access the shares? >> For this reason, I use KDE/Dolphin to access CIFS shares. It is >> closer to how Windows expects the shares to be treated. > >I don't see how things like shell scripts or other applications that >need to access files on the CIFS mounts would use something like that. Did you test if a small script that touches a file on the share every minute resolves the issue? -- Joost -- Sent from my Android device with K-9 Mail. Please excuse my brevity.
[gentoo-user] Re: CIFS mounts started misbehaving
On 2017-03-06, J. Roeleveld wrote: > On March 6, 2017 5:14:39 PM GMT+01:00, Grant Edwards > wrote: >>On 2017-03-06, Kai Krakow wrote: >> I'm going to try to set up a Wireshark capture in ring-buffer mode >>and somehow detect the failure and stop the capture... >>> >>> Did something on the Windows side change? >> >>Probaby, but I've learned not to ask questions like that. They never >>get answered, and it just causes problems when it is revealed that the >>client having problems is a Linux machine. >> >>> Maybe force Windows down to a lower SMB version or reduce/disable >>> SMB client side caching? > > Windows sharing is designed as a 'link when used' option. Not as a > permanent mount like Linix treats it. > > Even 'mounting' in Windows doesn't mean the share is actually > accessed. > > A windows CIFS server will not be reliable enough for long term > mounting. With Samba, it does work more reliable. (In my experience) It's worked perfectly fine for 10+ years, and apparently continues to do so for other Linux users in the office. > For this reason, I use KDE/Dolphin to access CIFS shares. It is > closer to how Windows expects the shares to be treated. I don't see how things like shell scripts or other applications that need to access files on the CIFS mounts would use something like that. -- Grant Edwards grant.b.edwardsYow! I think my career at is ruined! gmail.com
Re: [gentoo-user] Re: CIFS mounts started misbehaving
On March 6, 2017 5:14:39 PM GMT+01:00, Grant Edwards wrote: >On 2017-03-06, Kai Krakow wrote: > >>> I'm going to try to set up a Wireshark capture in ring-buffer mode >and >>> somehow detect the failure and stop the capture... >> >> Did something on the Windows side change? > >Probaby, but I've learned not to ask questions like that. They never >get answered, and it just causes problems when it is revealed that the >client having problems is a Linux machine. > >> Maybe force Windows down to a lower SMB version or reduce/disable >> SMB client side caching? Windows sharing is designed as a 'link when used' option. Not as a permanent mount like Linix treats it. Even 'mounting' in Windows doesn't mean the share is actually accessed. A windows CIFS server will not be reliable enough for long term mounting. With Samba, it does work more reliable. (In my experience) For this reason, I use KDE/Dolphin to access CIFS shares. It is closer to how Windows expects the shares to be treated. -- Joost -- Sent from my Android device with K-9 Mail. Please excuse my brevity.
[gentoo-user] Re: CIFS mounts started misbehaving
On 2017-03-06, Kai Krakow wrote: >> I'm going to try to set up a Wireshark capture in ring-buffer mode and >> somehow detect the failure and stop the capture... > > Did something on the Windows side change? Probaby, but I've learned not to ask questions like that. They never get answered, and it just causes problems when it is revealed that the client having problems is a Linux machine. > Maybe force Windows down to a lower SMB version or reduce/disable > SMB client side caching? -- Grant Edwards grant.b.edwardsYow! Like I always say at -- nothing can beat gmail.comthe BRATWURST here in DUSSELDORF!!
[gentoo-user] Re: CIFS mounts started misbehaving
Am Sat, 4 Mar 2017 16:42:07 + (UTC) schrieb Grant Edwards : > On 2017-03-04, Kai Krakow wrote: > > Am Sat, 04 Mar 2017 08:02:11 + schrieb "J. Roeleveld" > > : > >> > [...] > [...] > [...] > [...] > [...] > [...] > [...] > [...] > [...] > [...] > [...] > >> > [...] > >> > >> Are other hosts linux or windows? > > Other Linux and Windows clients don't seem to be having this problem. > > >> Maybe a dodgy switch forgetting the correct path? > > I don't think so. I can ping the host while the CIFS subsystem says > "host is down". If the switch is forgetting the path, who's sending > back the SYN/ACK and the RST > > > Or an MTU problem... Is there a router in the path? > > Nope. The MTU idea was dumb anyways as you wrote that the problem occurs after some idle time... Which could still be a router problem - but as you wrote: no router. :-) > I'm going to try to set up a Wireshark capture in ring-buffer mode and > somehow detect the failure and stop the capture... Did something on the Windows side change? Maybe force Windows down to a lower SMB version or reduce/disable SMB client side caching? -- Regards, Kai Replies to list-only preferred.
[gentoo-user] Re: CIFS mounts started misbehaving
On 2017-03-04, Kai Krakow wrote: > Am Sat, 04 Mar 2017 08:02:11 + schrieb "J. Roeleveld" > : > >> >> >Normally, when things are working but idle, the TCP connection to 445 >> >shows an SMB echo request/rseponse transaction once per minute. When >> >it fails, the TCP connection evidently got dropped, and the Windows >> >machine repeatedly shuts down new ones: >> > >> >The failure mode looks like this in wireshark: >> > >> > GentooWindows >> > >> > -> SYN -> 445 >> > <-SYN/ACK <- 445 >> > -> ACK -> 445 >> > -> SMB[echo req]-> 445 >> > <- RST <- 445 >> > >> >[that repeats 800 times per second for long periods of time] >> > >> >Then at some point, it starts to work: >> > >> > ->SYN -> 445 >> > <- SYN/ACK <- 445 >> > ->ACK -> 445 >> > -> SMB[proto neg req] -> 445 >> > <- SMB[proto neg rsp] <- 445 >> > -> SMB[ses setup req] -> 445 >> > <- SMB[ses setup rsp] <- 445 >> > ... >> >> >Sometimes the umount times out and "fails" because the "host is >> >down", and when that happens, it seems like it immediately starts to >> >work again. :/ >> >> Are other hosts linux or windows? Other Linux and Windows clients don't seem to be having this problem. >> Maybe a dodgy switch forgetting the correct path? I don't think so. I can ping the host while the CIFS subsystem says "host is down". If the switch is forgetting the path, who's sending back the SYN/ACK and the RST > Or an MTU problem... Is there a router in the path? Nope. I'm going to try to set up a Wireshark capture in ring-buffer mode and somehow detect the failure and stop the capture... -- Grant
[gentoo-user] Re: CIFS mounts started misbehaving
Am Sat, 04 Mar 2017 08:02:11 + schrieb "J. Roeleveld" : > On March 4, 2017 12:41:05 AM GMT+01:00, Grant Edwards > wrote: > >On 2017-03-03, J. Roeleveld wrote: > > > >> On March 3, 2017 7:49:27 PM GMT+01:00, Grant Edwards > > wrote: > > > [...] > >and > [...] > > > >[...] > > > >> My guess would be some timeout setting on the server killing the > >> login. > > > >That doesn't seem to be the problem. I've asked around, and others > >aren't seeing this problem. > > > >I've also noticed that sometimes the mounts will start working again > >without a umount/mount, but I can't figure out what causes it... > > > >Normally, when things are working but idle, the TCP connection to 445 > >shows an SMB echo request/rseponse transaction once per minute. When > >it fails, the TCP connection evidently got dropped, and the Windows > >machine repeatedly shuts down new ones: > > > >The failure mode looks like this in wireshark: > > > > GentooWindows > > > > -> SYN -> 445 > > <-SYN/ACK <- 445 > > -> ACK -> 445 > > -> SMB[echo req]-> 445 > > <- RST <- 445 > > > >[that repeats 800 times per second for long periods of time] > > > >Then at some point, it starts to work: > > > > ->SYN -> 445 > > <- SYN/ACK <- 445 > > ->ACK -> 445 > > -> SMB[proto neg req] -> 445 > > <- SMB[proto neg rsp] <- 445 > > -> SMB[ses setup req] -> 445 > > <- SMB[ses setup rsp] <- 445 > > ... > > > >Sometimes the umount times out and "fails" because the "host is > >down", and when that happens, it seems like it immediately starts to > >work again. :/ > > Are other hosts linux or windows? > > Maybe a dodgy switch forgetting the correct path? Or an MTU problem... Is there a router in the path? -- Regards, Kai Replies to list-only preferred.
Re: [gentoo-user] Re: CIFS mounts started misbehaving
On March 4, 2017 12:41:05 AM GMT+01:00, Grant Edwards wrote: >On 2017-03-03, J. Roeleveld wrote: > >> On March 3, 2017 7:49:27 PM GMT+01:00, Grant Edwards > wrote: > >>>About a week ago, they started acting oddly. They all mount fine, >and >>>work as usual as long as you keep using them. AFAICT, if they sit >>>idle for "a while" (tens of minutes, maybe an hour), they freeze up. > >[...] > >> My guess would be some timeout setting on the server killing the >> login. > >That doesn't seem to be the problem. I've asked around, and others >aren't seeing this problem. > >I've also noticed that sometimes the mounts will start working again >without a umount/mount, but I can't figure out what causes it... > >Normally, when things are working but idle, the TCP connection to 445 >shows an SMB echo request/rseponse transaction once per minute. When >it fails, the TCP connection evidently got dropped, and the Windows >machine repeatedly shuts down new ones: > >The failure mode looks like this in wireshark: > > GentooWindows > > -> SYN -> 445 > <-SYN/ACK <- 445 > -> ACK -> 445 > -> SMB[echo req]-> 445 > <- RST <- 445 > >[that repeats 800 times per second for long periods of time] > >Then at some point, it starts to work: > > ->SYN -> 445 > <- SYN/ACK <- 445 > ->ACK -> 445 > -> SMB[proto neg req] -> 445 > <- SMB[proto neg rsp] <- 445 > -> SMB[ses setup req] -> 445 > <- SMB[ses setup rsp] <- 445 > ... > >Sometimes the umount times out and "fails" because the "host is down", >and when that happens, it seems like it immediately starts to work >again. :/ Are other hosts linux or windows? Maybe a dodgy switch forgetting the correct path? -- Joost -- Sent from my Android device with K-9 Mail. Please excuse my brevity.
[gentoo-user] Re: CIFS mounts started misbehaving
On 2017-03-03, J. Roeleveld wrote: > On March 3, 2017 7:49:27 PM GMT+01:00, Grant Edwards > wrote: >>About a week ago, they started acting oddly. They all mount fine, and >>work as usual as long as you keep using them. AFAICT, if they sit >>idle for "a while" (tens of minutes, maybe an hour), they freeze up. [...] > My guess would be some timeout setting on the server killing the > login. That doesn't seem to be the problem. I've asked around, and others aren't seeing this problem. I've also noticed that sometimes the mounts will start working again without a umount/mount, but I can't figure out what causes it... Normally, when things are working but idle, the TCP connection to 445 shows an SMB echo request/rseponse transaction once per minute. When it fails, the TCP connection evidently got dropped, and the Windows machine repeatedly shuts down new ones: The failure mode looks like this in wireshark: GentooWindows -> SYN -> 445 <-SYN/ACK <- 445 -> ACK -> 445 -> SMB[echo req]-> 445 <- RST <- 445 [that repeats 800 times per second for long periods of time] Then at some point, it starts to work: ->SYN -> 445 <- SYN/ACK <- 445 ->ACK -> 445 -> SMB[proto neg req] -> 445 <- SMB[proto neg rsp] <- 445 -> SMB[ses setup req] -> 445 <- SMB[ses setup rsp] <- 445 ... Sometimes the umount times out and "fails" because the "host is down", and when that happens, it seems like it immediately starts to work again. :/ -- Grant Edwards grant.b.edwardsYow! Wow! Look!! A stray at meatball!! Let's interview gmail.comit!